--- /dev/null
+ GNU GENERAL PUBLIC LICENSE
+ Version 2, June 1991
+
+ Copyright (C) 1989, 1991 Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+ Preamble
+
+ The licenses for most software are designed to take away your
+freedom to share and change it. By contrast, the GNU General Public
+License is intended to guarantee your freedom to share and change free
+software--to make sure the software is free for all its users. This
+General Public License applies to most of the Free Software
+Foundation's software and to any other program whose authors commit to
+using it. (Some other Free Software Foundation software is covered by
+the GNU Lesser General Public License instead.) You can apply it to
+your programs, too.
+
+ When we speak of free software, we are referring to freedom, not
+price. Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+this service if you wish), that you receive source code or can get it
+if you want it, that you can change the software or use pieces of it
+in new free programs; and that you know you can do these things.
+
+ To protect your rights, we need to make restrictions that forbid
+anyone to deny you these rights or to ask you to surrender the rights.
+These restrictions translate to certain responsibilities for you if you
+distribute copies of the software, or if you modify it.
+
+ For example, if you distribute copies of such a program, whether
+gratis or for a fee, you must give the recipients all the rights that
+you have. You must make sure that they, too, receive or can get the
+source code. And you must show them these terms so they know their
+rights.
+
+ We protect your rights with two steps: (1) copyright the software, and
+(2) offer you this license which gives you legal permission to copy,
+distribute and/or modify the software.
+
+ Also, for each author's protection and ours, we want to make certain
+that everyone understands that there is no warranty for this free
+software. If the software is modified by someone else and passed on, we
+want its recipients to know that what they have is not the original, so
+that any problems introduced by others will not reflect on the original
+authors' reputations.
+
+ Finally, any free program is threatened constantly by software
+patents. We wish to avoid the danger that redistributors of a free
+program will individually obtain patent licenses, in effect making the
+program proprietary. To prevent this, we have made it clear that any
+patent must be licensed for everyone's free use or not licensed at all.
+
+ The precise terms and conditions for copying, distribution and
+modification follow.
+
+ GNU GENERAL PUBLIC LICENSE
+ TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+ 0. This License applies to any program or other work which contains
+a notice placed by the copyright holder saying it may be distributed
+under the terms of this General Public License. The "Program", below,
+refers to any such program or work, and a "work based on the Program"
+means either the Program or any derivative work under copyright law:
+that is to say, a work containing the Program or a portion of it,
+either verbatim or with modifications and/or translated into another
+language. (Hereinafter, translation is included without limitation in
+the term "modification".) Each licensee is addressed as "you".
+
+Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope. The act of
+running the Program is not restricted, and the output from the Program
+is covered only if its contents constitute a work based on the
+Program (independent of having been made by running the Program).
+Whether that is true depends on what the Program does.
+
+ 1. You may copy and distribute verbatim copies of the Program's
+source code as you receive it, in any medium, provided that you
+conspicuously and appropriately publish on each copy an appropriate
+copyright notice and disclaimer of warranty; keep intact all the
+notices that refer to this License and to the absence of any warranty;
+and give any other recipients of the Program a copy of this License
+along with the Program.
+
+You may charge a fee for the physical act of transferring a copy, and
+you may at your option offer warranty protection in exchange for a fee.
+
+ 2. You may modify your copy or copies of the Program or any portion
+of it, thus forming a work based on the Program, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+ a) You must cause the modified files to carry prominent notices
+ stating that you changed the files and the date of any change.
+
+ b) You must cause any work that you distribute or publish, that in
+ whole or in part contains or is derived from the Program or any
+ part thereof, to be licensed as a whole at no charge to all third
+ parties under the terms of this License.
+
+ c) If the modified program normally reads commands interactively
+ when run, you must cause it, when started running for such
+ interactive use in the most ordinary way, to print or display an
+ announcement including an appropriate copyright notice and a
+ notice that there is no warranty (or else, saying that you provide
+ a warranty) and that users may redistribute the program under
+ these conditions, and telling the user how to view a copy of this
+ License. (Exception: if the Program itself is interactive but
+ does not normally print such an announcement, your work based on
+ the Program is not required to print an announcement.)
+
+These requirements apply to the modified work as a whole. If
+identifiable sections of that work are not derived from the Program,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works. But when you
+distribute the same sections as part of a whole which is a work based
+on the Program, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Program.
+
+In addition, mere aggregation of another work not based on the Program
+with the Program (or with a work based on the Program) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+ 3. You may copy and distribute the Program (or a work based on it,
+under Section 2) in object code or executable form under the terms of
+Sections 1 and 2 above provided that you also do one of the following:
+
+ a) Accompany it with the complete corresponding machine-readable
+ source code, which must be distributed under the terms of Sections
+ 1 and 2 above on a medium customarily used for software interchange; or,
+
+ b) Accompany it with a written offer, valid for at least three
+ years, to give any third party, for a charge no more than your
+ cost of physically performing source distribution, a complete
+ machine-readable copy of the corresponding source code, to be
+ distributed under the terms of Sections 1 and 2 above on a medium
+ customarily used for software interchange; or,
+
+ c) Accompany it with the information you received as to the offer
+ to distribute corresponding source code. (This alternative is
+ allowed only for noncommercial distribution and only if you
+ received the program in object code or executable form with such
+ an offer, in accord with Subsection b above.)
+
+The source code for a work means the preferred form of the work for
+making modifications to it. For an executable work, complete source
+code means all the source code for all modules it contains, plus any
+associated interface definition files, plus the scripts used to
+control compilation and installation of the executable. However, as a
+special exception, the source code distributed need not include
+anything that is normally distributed (in either source or binary
+form) with the major components (compiler, kernel, and so on) of the
+operating system on which the executable runs, unless that component
+itself accompanies the executable.
+
+If distribution of executable or object code is made by offering
+access to copy from a designated place, then offering equivalent
+access to copy the source code from the same place counts as
+distribution of the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+ 4. You may not copy, modify, sublicense, or distribute the Program
+except as expressly provided under this License. Any attempt
+otherwise to copy, modify, sublicense or distribute the Program is
+void, and will automatically terminate your rights under this License.
+However, parties who have received copies, or rights, from you under
+this License will not have their licenses terminated so long as such
+parties remain in full compliance.
+
+ 5. You are not required to accept this License, since you have not
+signed it. However, nothing else grants you permission to modify or
+distribute the Program or its derivative works. These actions are
+prohibited by law if you do not accept this License. Therefore, by
+modifying or distributing the Program (or any work based on the
+Program), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Program or works based on it.
+
+ 6. Each time you redistribute the Program (or any work based on the
+Program), the recipient automatically receives a license from the
+original licensor to copy, distribute or modify the Program subject to
+these terms and conditions. You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties to
+this License.
+
+ 7. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License. If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Program at all. For example, if a patent
+license would not permit royalty-free redistribution of the Program by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Program.
+
+If any portion of this section is held invalid or unenforceable under
+any particular circumstance, the balance of the section is intended to
+apply and the section as a whole is intended to apply in other
+circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system, which is
+implemented by public license practices. Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+ 8. If the distribution and/or use of the Program is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Program under this License
+may add an explicit geographical distribution limitation excluding
+those countries, so that distribution is permitted only in or among
+countries not thus excluded. In such case, this License incorporates
+the limitation as if written in the body of this License.
+
+ 9. The Free Software Foundation may publish revised and/or new versions
+of the General Public License from time to time. Such new versions will
+be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+Each version is given a distinguishing version number. If the Program
+specifies a version number of this License which applies to it and "any
+later version", you have the option of following the terms and conditions
+either of that version or of any later version published by the Free
+Software Foundation. If the Program does not specify a version number of
+this License, you may choose any version ever published by the Free Software
+Foundation.
+
+ 10. If you wish to incorporate parts of the Program into other free
+programs whose distribution conditions are different, write to the author
+to ask for permission. For software which is copyrighted by the Free
+Software Foundation, write to the Free Software Foundation; we sometimes
+make exceptions for this. Our decision will be guided by the two goals
+of preserving the free status of all derivatives of our free software and
+of promoting the sharing and reuse of software generally.
+
+ NO WARRANTY
+
+ 11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
+FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
+OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
+PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
+OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
+MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
+TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
+PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
+REPAIR OR CORRECTION.
+
+ 12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
+REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
+INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
+OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
+TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
+YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
+PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGES.
+
+ END OF TERMS AND CONDITIONS
+
+ How to Apply These Terms to Your New Programs
+
+ If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+ To do so, attach the following notices to the program. It is safest
+to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+ <one line to give the program's name and a brief idea of what it does.>
+ Copyright (C) <year> <name of author>
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program is interactive, make it output a short notice like this
+when it starts in an interactive mode:
+
+ Gnomovision version 69, Copyright (C) year name of author
+ Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+ This is free software, and you are welcome to redistribute it
+ under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate
+parts of the General Public License. Of course, the commands you use may
+be called something other than `show w' and `show c'; they could even be
+mouse-clicks or menu items--whatever suits your program.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the program, if
+necessary. Here is a sample; alter the names:
+
+ Yoyodyne, Inc., hereby disclaims all copyright interest in the program
+ `Gnomovision' (which makes passes at compilers) written by James Hacker.
+
+ <signature of Ty Coon>, 1 April 1989
+ Ty Coon, President of Vice
+
+This General Public License does not permit incorporating your program into
+proprietary programs. If your program is a subroutine library, you may
+consider it more useful to permit linking proprietary applications with the
+library. If this is what you want to do, use the GNU Lesser General
+Public License instead of this License.
--- /dev/null
+ GNU LESSER GENERAL PUBLIC LICENSE
+ Version 2.1, February 1999
+
+ Copyright (C) 1991, 1999 Free Software Foundation, Inc.
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+[This is the first released version of the Lesser GPL. It also counts
+ as the successor of the GNU Library Public License, version 2, hence
+ the version number 2.1.]
+
+ Preamble
+
+ The licenses for most software are designed to take away your
+freedom to share and change it. By contrast, the GNU General Public
+Licenses are intended to guarantee your freedom to share and change
+free software--to make sure the software is free for all its users.
+
+ This license, the Lesser General Public License, applies to some
+specially designated software packages--typically libraries--of the
+Free Software Foundation and other authors who decide to use it. You
+can use it too, but we suggest you first think carefully about whether
+this license or the ordinary General Public License is the better
+strategy to use in any particular case, based on the explanations below.
+
+ When we speak of free software, we are referring to freedom of use,
+not price. Our General Public Licenses are designed to make sure that
+you have the freedom to distribute copies of free software (and charge
+for this service if you wish); that you receive source code or can get
+it if you want it; that you can change the software and use pieces of
+it in new free programs; and that you are informed that you can do
+these things.
+
+ To protect your rights, we need to make restrictions that forbid
+distributors to deny you these rights or to ask you to surrender these
+rights. These restrictions translate to certain responsibilities for
+you if you distribute copies of the library or if you modify it.
+
+ For example, if you distribute copies of the library, whether gratis
+or for a fee, you must give the recipients all the rights that we gave
+you. You must make sure that they, too, receive or can get the source
+code. If you link other code with the library, you must provide
+complete object files to the recipients, so that they can relink them
+with the library after making changes to the library and recompiling
+it. And you must show them these terms so they know their rights.
+
+ We protect your rights with a two-step method: (1) we copyright the
+library, and (2) we offer you this license, which gives you legal
+permission to copy, distribute and/or modify the library.
+
+ To protect each distributor, we want to make it very clear that
+there is no warranty for the free library. Also, if the library is
+modified by someone else and passed on, the recipients should know
+that what they have is not the original version, so that the original
+author's reputation will not be affected by problems that might be
+introduced by others.
+\f
+ Finally, software patents pose a constant threat to the existence of
+any free program. We wish to make sure that a company cannot
+effectively restrict the users of a free program by obtaining a
+restrictive license from a patent holder. Therefore, we insist that
+any patent license obtained for a version of the library must be
+consistent with the full freedom of use specified in this license.
+
+ Most GNU software, including some libraries, is covered by the
+ordinary GNU General Public License. This license, the GNU Lesser
+General Public License, applies to certain designated libraries, and
+is quite different from the ordinary General Public License. We use
+this license for certain libraries in order to permit linking those
+libraries into non-free programs.
+
+ When a program is linked with a library, whether statically or using
+a shared library, the combination of the two is legally speaking a
+combined work, a derivative of the original library. The ordinary
+General Public License therefore permits such linking only if the
+entire combination fits its criteria of freedom. The Lesser General
+Public License permits more lax criteria for linking other code with
+the library.
+
+ We call this license the "Lesser" General Public License because it
+does Less to protect the user's freedom than the ordinary General
+Public License. It also provides other free software developers Less
+of an advantage over competing non-free programs. These disadvantages
+are the reason we use the ordinary General Public License for many
+libraries. However, the Lesser license provides advantages in certain
+special circumstances.
+
+ For example, on rare occasions, there may be a special need to
+encourage the widest possible use of a certain library, so that it becomes
+a de-facto standard. To achieve this, non-free programs must be
+allowed to use the library. A more frequent case is that a free
+library does the same job as widely used non-free libraries. In this
+case, there is little to gain by limiting the free library to free
+software only, so we use the Lesser General Public License.
+
+ In other cases, permission to use a particular library in non-free
+programs enables a greater number of people to use a large body of
+free software. For example, permission to use the GNU C Library in
+non-free programs enables many more people to use the whole GNU
+operating system, as well as its variant, the GNU/Linux operating
+system.
+
+ Although the Lesser General Public License is Less protective of the
+users' freedom, it does ensure that the user of a program that is
+linked with the Library has the freedom and the wherewithal to run
+that program using a modified version of the Library.
+
+ The precise terms and conditions for copying, distribution and
+modification follow. Pay close attention to the difference between a
+"work based on the library" and a "work that uses the library". The
+former contains code derived from the library, whereas the latter must
+be combined with the library in order to run.
+\f
+ GNU LESSER GENERAL PUBLIC LICENSE
+ TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
+
+ 0. This License Agreement applies to any software library or other
+program which contains a notice placed by the copyright holder or
+other authorized party saying it may be distributed under the terms of
+this Lesser General Public License (also called "this License").
+Each licensee is addressed as "you".
+
+ A "library" means a collection of software functions and/or data
+prepared so as to be conveniently linked with application programs
+(which use some of those functions and data) to form executables.
+
+ The "Library", below, refers to any such software library or work
+which has been distributed under these terms. A "work based on the
+Library" means either the Library or any derivative work under
+copyright law: that is to say, a work containing the Library or a
+portion of it, either verbatim or with modifications and/or translated
+straightforwardly into another language. (Hereinafter, translation is
+included without limitation in the term "modification".)
+
+ "Source code" for a work means the preferred form of the work for
+making modifications to it. For a library, complete source code means
+all the source code for all modules it contains, plus any associated
+interface definition files, plus the scripts used to control compilation
+and installation of the library.
+
+ Activities other than copying, distribution and modification are not
+covered by this License; they are outside its scope. The act of
+running a program using the Library is not restricted, and output from
+such a program is covered only if its contents constitute a work based
+on the Library (independent of the use of the Library in a tool for
+writing it). Whether that is true depends on what the Library does
+and what the program that uses the Library does.
+
+ 1. You may copy and distribute verbatim copies of the Library's
+complete source code as you receive it, in any medium, provided that
+you conspicuously and appropriately publish on each copy an
+appropriate copyright notice and disclaimer of warranty; keep intact
+all the notices that refer to this License and to the absence of any
+warranty; and distribute a copy of this License along with the
+Library.
+
+ You may charge a fee for the physical act of transferring a copy,
+and you may at your option offer warranty protection in exchange for a
+fee.
+\f
+ 2. You may modify your copy or copies of the Library or any portion
+of it, thus forming a work based on the Library, and copy and
+distribute such modifications or work under the terms of Section 1
+above, provided that you also meet all of these conditions:
+
+ a) The modified work must itself be a software library.
+
+ b) You must cause the files modified to carry prominent notices
+ stating that you changed the files and the date of any change.
+
+ c) You must cause the whole of the work to be licensed at no
+ charge to all third parties under the terms of this License.
+
+ d) If a facility in the modified Library refers to a function or a
+ table of data to be supplied by an application program that uses
+ the facility, other than as an argument passed when the facility
+ is invoked, then you must make a good faith effort to ensure that,
+ in the event an application does not supply such function or
+ table, the facility still operates, and performs whatever part of
+ its purpose remains meaningful.
+
+ (For example, a function in a library to compute square roots has
+ a purpose that is entirely well-defined independent of the
+ application. Therefore, Subsection 2d requires that any
+ application-supplied function or table used by this function must
+ be optional: if the application does not supply it, the square
+ root function must still compute square roots.)
+
+These requirements apply to the modified work as a whole. If
+identifiable sections of that work are not derived from the Library,
+and can be reasonably considered independent and separate works in
+themselves, then this License, and its terms, do not apply to those
+sections when you distribute them as separate works. But when you
+distribute the same sections as part of a whole which is a work based
+on the Library, the distribution of the whole must be on the terms of
+this License, whose permissions for other licensees extend to the
+entire whole, and thus to each and every part regardless of who wrote
+it.
+
+Thus, it is not the intent of this section to claim rights or contest
+your rights to work written entirely by you; rather, the intent is to
+exercise the right to control the distribution of derivative or
+collective works based on the Library.
+
+In addition, mere aggregation of another work not based on the Library
+with the Library (or with a work based on the Library) on a volume of
+a storage or distribution medium does not bring the other work under
+the scope of this License.
+
+ 3. You may opt to apply the terms of the ordinary GNU General Public
+License instead of this License to a given copy of the Library. To do
+this, you must alter all the notices that refer to this License, so
+that they refer to the ordinary GNU General Public License, version 2,
+instead of to this License. (If a newer version than version 2 of the
+ordinary GNU General Public License has appeared, then you can specify
+that version instead if you wish.) Do not make any other change in
+these notices.
+\f
+ Once this change is made in a given copy, it is irreversible for
+that copy, so the ordinary GNU General Public License applies to all
+subsequent copies and derivative works made from that copy.
+
+ This option is useful when you wish to copy part of the code of
+the Library into a program that is not a library.
+
+ 4. You may copy and distribute the Library (or a portion or
+derivative of it, under Section 2) in object code or executable form
+under the terms of Sections 1 and 2 above provided that you accompany
+it with the complete corresponding machine-readable source code, which
+must be distributed under the terms of Sections 1 and 2 above on a
+medium customarily used for software interchange.
+
+ If distribution of object code is made by offering access to copy
+from a designated place, then offering equivalent access to copy the
+source code from the same place satisfies the requirement to
+distribute the source code, even though third parties are not
+compelled to copy the source along with the object code.
+
+ 5. A program that contains no derivative of any portion of the
+Library, but is designed to work with the Library by being compiled or
+linked with it, is called a "work that uses the Library". Such a
+work, in isolation, is not a derivative work of the Library, and
+therefore falls outside the scope of this License.
+
+ However, linking a "work that uses the Library" with the Library
+creates an executable that is a derivative of the Library (because it
+contains portions of the Library), rather than a "work that uses the
+library". The executable is therefore covered by this License.
+Section 6 states terms for distribution of such executables.
+
+ When a "work that uses the Library" uses material from a header file
+that is part of the Library, the object code for the work may be a
+derivative work of the Library even though the source code is not.
+Whether this is true is especially significant if the work can be
+linked without the Library, or if the work is itself a library. The
+threshold for this to be true is not precisely defined by law.
+
+ If such an object file uses only numerical parameters, data
+structure layouts and accessors, and small macros and small inline
+functions (ten lines or less in length), then the use of the object
+file is unrestricted, regardless of whether it is legally a derivative
+work. (Executables containing this object code plus portions of the
+Library will still fall under Section 6.)
+
+ Otherwise, if the work is a derivative of the Library, you may
+distribute the object code for the work under the terms of Section 6.
+Any executables containing that work also fall under Section 6,
+whether or not they are linked directly with the Library itself.
+\f
+ 6. As an exception to the Sections above, you may also combine or
+link a "work that uses the Library" with the Library to produce a
+work containing portions of the Library, and distribute that work
+under terms of your choice, provided that the terms permit
+modification of the work for the customer's own use and reverse
+engineering for debugging such modifications.
+
+ You must give prominent notice with each copy of the work that the
+Library is used in it and that the Library and its use are covered by
+this License. You must supply a copy of this License. If the work
+during execution displays copyright notices, you must include the
+copyright notice for the Library among them, as well as a reference
+directing the user to the copy of this License. Also, you must do one
+of these things:
+
+ a) Accompany the work with the complete corresponding
+ machine-readable source code for the Library including whatever
+ changes were used in the work (which must be distributed under
+ Sections 1 and 2 above); and, if the work is an executable linked
+ with the Library, with the complete machine-readable "work that
+ uses the Library", as object code and/or source code, so that the
+ user can modify the Library and then relink to produce a modified
+ executable containing the modified Library. (It is understood
+ that the user who changes the contents of definitions files in the
+ Library will not necessarily be able to recompile the application
+ to use the modified definitions.)
+
+ b) Use a suitable shared library mechanism for linking with the
+ Library. A suitable mechanism is one that (1) uses at run time a
+ copy of the library already present on the user's computer system,
+ rather than copying library functions into the executable, and (2)
+ will operate properly with a modified version of the library, if
+ the user installs one, as long as the modified version is
+ interface-compatible with the version that the work was made with.
+
+ c) Accompany the work with a written offer, valid for at
+ least three years, to give the same user the materials
+ specified in Subsection 6a, above, for a charge no more
+ than the cost of performing this distribution.
+
+ d) If distribution of the work is made by offering access to copy
+ from a designated place, offer equivalent access to copy the above
+ specified materials from the same place.
+
+ e) Verify that the user has already received a copy of these
+ materials or that you have already sent this user a copy.
+
+ For an executable, the required form of the "work that uses the
+Library" must include any data and utility programs needed for
+reproducing the executable from it. However, as a special exception,
+the materials to be distributed need not include anything that is
+normally distributed (in either source or binary form) with the major
+components (compiler, kernel, and so on) of the operating system on
+which the executable runs, unless that component itself accompanies
+the executable.
+
+ It may happen that this requirement contradicts the license
+restrictions of other proprietary libraries that do not normally
+accompany the operating system. Such a contradiction means you cannot
+use both them and the Library together in an executable that you
+distribute.
+\f
+ 7. You may place library facilities that are a work based on the
+Library side-by-side in a single library together with other library
+facilities not covered by this License, and distribute such a combined
+library, provided that the separate distribution of the work based on
+the Library and of the other library facilities is otherwise
+permitted, and provided that you do these two things:
+
+ a) Accompany the combined library with a copy of the same work
+ based on the Library, uncombined with any other library
+ facilities. This must be distributed under the terms of the
+ Sections above.
+
+ b) Give prominent notice with the combined library of the fact
+ that part of it is a work based on the Library, and explaining
+ where to find the accompanying uncombined form of the same work.
+
+ 8. You may not copy, modify, sublicense, link with, or distribute
+the Library except as expressly provided under this License. Any
+attempt otherwise to copy, modify, sublicense, link with, or
+distribute the Library is void, and will automatically terminate your
+rights under this License. However, parties who have received copies,
+or rights, from you under this License will not have their licenses
+terminated so long as such parties remain in full compliance.
+
+ 9. You are not required to accept this License, since you have not
+signed it. However, nothing else grants you permission to modify or
+distribute the Library or its derivative works. These actions are
+prohibited by law if you do not accept this License. Therefore, by
+modifying or distributing the Library (or any work based on the
+Library), you indicate your acceptance of this License to do so, and
+all its terms and conditions for copying, distributing or modifying
+the Library or works based on it.
+
+ 10. Each time you redistribute the Library (or any work based on the
+Library), the recipient automatically receives a license from the
+original licensor to copy, distribute, link with or modify the Library
+subject to these terms and conditions. You may not impose any further
+restrictions on the recipients' exercise of the rights granted herein.
+You are not responsible for enforcing compliance by third parties with
+this License.
+\f
+ 11. If, as a consequence of a court judgment or allegation of patent
+infringement or for any other reason (not limited to patent issues),
+conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License. If you cannot
+distribute so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you
+may not distribute the Library at all. For example, if a patent
+license would not permit royalty-free redistribution of the Library by
+all those who receive copies directly or indirectly through you, then
+the only way you could satisfy both it and this License would be to
+refrain entirely from distribution of the Library.
+
+If any portion of this section is held invalid or unenforceable under any
+particular circumstance, the balance of the section is intended to apply,
+and the section as a whole is intended to apply in other circumstances.
+
+It is not the purpose of this section to induce you to infringe any
+patents or other property right claims or to contest validity of any
+such claims; this section has the sole purpose of protecting the
+integrity of the free software distribution system which is
+implemented by public license practices. Many people have made
+generous contributions to the wide range of software distributed
+through that system in reliance on consistent application of that
+system; it is up to the author/donor to decide if he or she is willing
+to distribute software through any other system and a licensee cannot
+impose that choice.
+
+This section is intended to make thoroughly clear what is believed to
+be a consequence of the rest of this License.
+
+ 12. If the distribution and/or use of the Library is restricted in
+certain countries either by patents or by copyrighted interfaces, the
+original copyright holder who places the Library under this License may add
+an explicit geographical distribution limitation excluding those countries,
+so that distribution is permitted only in or among countries not thus
+excluded. In such case, this License incorporates the limitation as if
+written in the body of this License.
+
+ 13. The Free Software Foundation may publish revised and/or new
+versions of the Lesser General Public License from time to time.
+Such new versions will be similar in spirit to the present version,
+but may differ in detail to address new problems or concerns.
+
+Each version is given a distinguishing version number. If the Library
+specifies a version number of this License which applies to it and
+"any later version", you have the option of following the terms and
+conditions either of that version or of any later version published by
+the Free Software Foundation. If the Library does not specify a
+license version number, you may choose any version ever published by
+the Free Software Foundation.
+\f
+ 14. If you wish to incorporate parts of the Library into other free
+programs whose distribution conditions are incompatible with these,
+write to the author to ask for permission. For software which is
+copyrighted by the Free Software Foundation, write to the Free
+Software Foundation; we sometimes make exceptions for this. Our
+decision will be guided by the two goals of preserving the free status
+of all derivatives of our free software and of promoting the sharing
+and reuse of software generally.
+
+ NO WARRANTY
+
+ 15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE, THERE IS NO
+WARRANTY FOR THE LIBRARY, TO THE EXTENT PERMITTED BY APPLICABLE LAW.
+EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR
+OTHER PARTIES PROVIDE THE LIBRARY "AS IS" WITHOUT WARRANTY OF ANY
+KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE
+LIBRARY IS WITH YOU. SHOULD THE LIBRARY PROVE DEFECTIVE, YOU ASSUME
+THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+ 16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN
+WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY
+AND/OR REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE LIABLE TO YOU
+FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR
+CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE
+LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING
+RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A
+FAILURE OF THE LIBRARY TO OPERATE WITH ANY OTHER SOFTWARE), EVEN IF
+SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
+DAMAGES.
+
+ END OF TERMS AND CONDITIONS
+\f
+ How to Apply These Terms to Your New Libraries
+
+ If you develop a new library, and you want it to be of the greatest
+possible use to the public, we recommend making it free software that
+everyone can redistribute and change. You can do so by permitting
+redistribution under these terms (or, alternatively, under the terms of the
+ordinary General Public License).
+
+ To apply these terms, attach the following notices to the library. It is
+safest to attach them to the start of each source file to most effectively
+convey the exclusion of warranty; and each file should have at least the
+"copyright" line and a pointer to where the full notice is found.
+
+ <one line to give the library's name and a brief idea of what it does.>
+ Copyright (C) <year> <name of author>
+
+ This library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ This library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with this library; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+Also add information on how to contact you by electronic and paper mail.
+
+You should also get your employer (if you work as a programmer) or your
+school, if any, to sign a "copyright disclaimer" for the library, if
+necessary. Here is a sample; alter the names:
+
+ Yoyodyne, Inc., hereby disclaims all copyright interest in the
+ library `Frob' (a library for tweaking knobs) written by James Random Hacker.
+
+ <signature of Ty Coon>, 1 April 1990
+ Ty Coon, President of Vice
+
+That's all there is to it!
cgroups/cgroup.c cgroups/cgroup.h \
cgroups/cgroup2_devices.c cgroups/cgroup2_devices.h \
cgroups/cgroup_utils.c cgroups/cgroup_utils.h \
- cpuset.c \
+ cpuset_parse.c cpuset_parse.h \
memory_utils.h \
+ proc_cpuview.c proc_cpuview.h \
+ proc_fuse.c proc_fuse.h \
+ proc_loadavg.c proc_loadavg.h \
sysfs_fuse.c sysfs_fuse.h \
utils.c utils.h
liblxcfs_la_CFLAGS = $(AM_CFLAGS)
cgroups/cgroup.c cgroups/cgroup.h \
cgroups/cgroup2_devices.c cgroups/cgroup2_devices.h \
cgroups/cgroup_utils.c cgroups/cgroup_utils.h \
- cpuset.c \
+ cpuset_parse.c cpuset_parse.h \
memory_utils.h \
+ proc_cpuview.c proc_cpuview.h \
+ proc_fuse.c proc_fuse.h \
+ proc_loadavg.c proc_loadavg.h \
sysfs_fuse.c sysfs_fuse.h \
utils.c utils.h
liblxcfstest_la_CFLAGS = $(AM_CFLAGS) -DRELOADTEST
endif
TEST_READ: tests/test-read.c
- $(CC) -o tests/test-read tests/test-read.c
-TEST_CPUSET: tests/cpusetrange.c cpuset.c
- $(CC) -o tests/cpusetrange tests/cpusetrange.c cpuset.c
+ $(CC) -o tests/test-read \
+ tests/test-read.c
+TEST_CPUSET: tests/cpusetrange.c cpuset_parse.c
+ $(CC) -o tests/cpusetrange \
+ tests/cpusetrange.c \
+ cpuset_parse.c
TEST_SYSCALLS: tests/test_syscalls.c
- $(CC) -o tests/test_syscalls tests/test_syscalls.c
-
+ $(CC) -o tests/test_syscalls \
+ tests/test_syscalls.c
tests: TEST_READ TEST_CPUSET TEST_SYSCALLS
distclean:
* See COPYING file for details.
*/
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
-#define __STDC_FORMAT_MACROS
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include "cgroups/cgroup.h"
#include "cgroups/cgroup_utils.h"
#include "memory_utils.h"
+#include "proc_cpuview.h"
#include "utils.h"
/* Define pivot_root() if missing from the C library */
extern int pivot_root(const char * new_root, const char * put_old);
#endif
-struct cpuacct_usage {
- uint64_t user;
- uint64_t system;
- uint64_t idle;
- bool online;
-};
-
-/* The function of hash table.*/
-#define LOAD_SIZE 100 /*the size of hash_table */
-#define FLUSH_TIME 5 /*the flush rate */
-#define DEPTH_DIR 3 /*the depth of per cgroup */
-/* The function of calculate loadavg .*/
-#define FSHIFT 11 /* nr of bits of precision */
-#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
-#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
-#define EXP_5 2014 /* 1/exp(5sec/5min) */
-#define EXP_15 2037 /* 1/exp(5sec/15min) */
-#define LOAD_INT(x) ((x) >> FSHIFT)
-#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100)
-/*
- * This parameter is used for proc_loadavg_read().
- * 1 means use loadavg, 0 means not use.
- */
-static int loadavg = 0;
-static volatile sig_atomic_t loadavg_stop = 0;
-static int calc_hash(const char *name)
-{
- unsigned int hash = 0;
- unsigned int x = 0;
- /* ELFHash algorithm. */
- while (*name) {
- hash = (hash << 4) + *name++;
- x = hash & 0xf0000000;
- if (x != 0)
- hash ^= (x >> 24);
- hash &= ~x;
- }
- return (hash & 0x7fffffff);
-}
-
-struct load_node {
- char *cg; /*cg */
- unsigned long avenrun[3]; /* Load averages */
- unsigned int run_pid;
- unsigned int total_pid;
- unsigned int last_pid;
- int cfd; /* The file descriptor of the mounted cgroup */
- struct load_node *next;
- struct load_node **pre;
-};
-
-struct load_head {
- /*
- * The lock is about insert load_node and refresh load_node.To the first
- * load_node of each hash bucket, insert and refresh in this hash bucket is
- * mutually exclusive.
- */
- pthread_mutex_t lock;
- /*
- * The rdlock is about read loadavg and delete load_node.To each hash
- * bucket, read and delete is mutually exclusive. But at the same time, we
- * allow paratactic read operation. This rdlock is at list level.
- */
- pthread_rwlock_t rdlock;
- /*
- * The rilock is about read loadavg and insert load_node.To the first
- * load_node of each hash bucket, read and insert is mutually exclusive.
- * But at the same time, we allow paratactic read operation.
- */
- pthread_rwlock_t rilock;
- struct load_node *next;
-};
-
-static struct load_head load_hash[LOAD_SIZE]; /* hash table */
-/*
- * init_load initialize the hash table.
- * Return 0 on success, return -1 on failure.
- */
-static int init_load(void)
-{
- int i;
- int ret;
-
- for (i = 0; i < LOAD_SIZE; i++) {
- load_hash[i].next = NULL;
- ret = pthread_mutex_init(&load_hash[i].lock, NULL);
- if (ret != 0) {
- lxcfs_error("%s\n", "Failed to initialize lock");
- goto out3;
- }
- ret = pthread_rwlock_init(&load_hash[i].rdlock, NULL);
- if (ret != 0) {
- lxcfs_error("%s\n", "Failed to initialize rdlock");
- goto out2;
- }
- ret = pthread_rwlock_init(&load_hash[i].rilock, NULL);
- if (ret != 0) {
- lxcfs_error("%s\n", "Failed to initialize rilock");
- goto out1;
- }
- }
- return 0;
-out1:
- pthread_rwlock_destroy(&load_hash[i].rdlock);
-out2:
- pthread_mutex_destroy(&load_hash[i].lock);
-out3:
- while (i > 0) {
- i--;
- pthread_mutex_destroy(&load_hash[i].lock);
- pthread_rwlock_destroy(&load_hash[i].rdlock);
- pthread_rwlock_destroy(&load_hash[i].rilock);
- }
- return -1;
-}
-
-static void insert_node(struct load_node **n, int locate)
-{
- struct load_node *f;
-
- pthread_mutex_lock(&load_hash[locate].lock);
- pthread_rwlock_wrlock(&load_hash[locate].rilock);
- f = load_hash[locate].next;
- load_hash[locate].next = *n;
-
- (*n)->pre = &(load_hash[locate].next);
- if (f)
- f->pre = &((*n)->next);
- (*n)->next = f;
- pthread_mutex_unlock(&load_hash[locate].lock);
- pthread_rwlock_unlock(&load_hash[locate].rilock);
-}
-/*
- * locate_node() finds special node. Not return NULL means success.
- * It should be noted that rdlock isn't unlocked at the end of code
- * because this function is used to read special node. Delete is not
- * allowed before read has ended.
- * unlock rdlock only in proc_loadavg_read().
- */
-static struct load_node *locate_node(char *cg, int locate)
-{
- struct load_node *f = NULL;
- int i = 0;
-
- pthread_rwlock_rdlock(&load_hash[locate].rilock);
- pthread_rwlock_rdlock(&load_hash[locate].rdlock);
- if (load_hash[locate].next == NULL) {
- pthread_rwlock_unlock(&load_hash[locate].rilock);
- return f;
- }
- f = load_hash[locate].next;
- pthread_rwlock_unlock(&load_hash[locate].rilock);
- while (f && ((i = strcmp(f->cg, cg)) != 0))
- f = f->next;
- return f;
-}
-
-/* Delete the load_node n and return the next node of it. */
-static struct load_node *del_node(struct load_node *n, int locate)
-{
- struct load_node *g;
-
- pthread_rwlock_wrlock(&load_hash[locate].rdlock);
- if (n->next == NULL) {
- *(n->pre) = NULL;
- } else {
- *(n->pre) = n->next;
- n->next->pre = n->pre;
- }
- g = n->next;
- free_disarm(n->cg);
- free_disarm(n);
- pthread_rwlock_unlock(&load_hash[locate].rdlock);
- return g;
-}
-
-static void load_free(void)
-{
- struct load_node *f, *p;
-
- for (int i = 0; i < LOAD_SIZE; i++) {
- pthread_mutex_lock(&load_hash[i].lock);
- pthread_rwlock_wrlock(&load_hash[i].rilock);
- pthread_rwlock_wrlock(&load_hash[i].rdlock);
- if (load_hash[i].next == NULL) {
- pthread_mutex_unlock(&load_hash[i].lock);
- pthread_mutex_destroy(&load_hash[i].lock);
- pthread_rwlock_unlock(&load_hash[i].rilock);
- pthread_rwlock_destroy(&load_hash[i].rilock);
- pthread_rwlock_unlock(&load_hash[i].rdlock);
- pthread_rwlock_destroy(&load_hash[i].rdlock);
- continue;
- }
-
- for (f = load_hash[i].next; f;) {
- free_disarm(f->cg);
- p = f->next;
- free_disarm(f);
- f = p;
- }
-
- pthread_mutex_unlock(&load_hash[i].lock);
- pthread_mutex_destroy(&load_hash[i].lock);
- pthread_rwlock_unlock(&load_hash[i].rilock);
- pthread_rwlock_destroy(&load_hash[i].rilock);
- pthread_rwlock_unlock(&load_hash[i].rdlock);
- pthread_rwlock_destroy(&load_hash[i].rdlock);
- }
-}
-
-/* Data for CPU view */
-struct cg_proc_stat {
- char *cg;
- struct cpuacct_usage *usage; // Real usage as read from the host's /proc/stat
- struct cpuacct_usage *view; // Usage stats reported to the container
- int cpu_count;
- pthread_mutex_t lock; // For node manipulation
- struct cg_proc_stat *next;
-};
-
-struct cg_proc_stat_head {
- struct cg_proc_stat *next;
- time_t lastcheck;
-
- /*
- * For access to the list. Reading can be parallel, pruning is exclusive.
- */
- pthread_rwlock_t lock;
-};
-
-#define CPUVIEW_HASH_SIZE 100
-static struct cg_proc_stat_head *proc_stat_history[CPUVIEW_HASH_SIZE];
-
-static bool cpuview_init_head(struct cg_proc_stat_head **head)
-{
- *head = malloc(sizeof(struct cg_proc_stat_head));
- if (!(*head)) {
- lxcfs_error("%s\n", strerror(errno));
- return false;
- }
-
- (*head)->lastcheck = time(NULL);
- (*head)->next = NULL;
-
- if (pthread_rwlock_init(&(*head)->lock, NULL) != 0) {
- lxcfs_error("%s\n", "Failed to initialize list lock");
- free_disarm(*head);
- return false;
- }
-
- return true;
-}
-
-static bool init_cpuview()
-{
- int i;
-
- for (i = 0; i < CPUVIEW_HASH_SIZE; i++)
- proc_stat_history[i] = NULL;
-
- for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
- if (!cpuview_init_head(&proc_stat_history[i]))
- goto err;
- }
-
- return true;
-
-err:
- for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
- if (proc_stat_history[i])
- free_disarm(proc_stat_history[i]);
- }
-
- return false;
-}
-
-static void free_proc_stat_node(struct cg_proc_stat *node)
-{
- pthread_mutex_destroy(&node->lock);
- free_disarm(node->cg);
- free_disarm(node->usage);
- free_disarm(node->view);
- free_disarm(node);
-}
-
-static void cpuview_free_head(struct cg_proc_stat_head *head)
-{
- struct cg_proc_stat *node, *tmp;
-
- if (head->next) {
- node = head->next;
-
- for (;;) {
- tmp = node;
- node = node->next;
- free_proc_stat_node(tmp);
-
- if (!node)
- break;
- }
- }
-
- pthread_rwlock_destroy(&head->lock);
- free_disarm(head);
-}
-
-static void free_cpuview()
-{
- int i;
-
- for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
- if (proc_stat_history[i])
- cpuview_free_head(proc_stat_history[i]);
- }
-}
-
/*
* A table caching which pid is init for a pid namespace.
* When looking up which pid is init for $qpid, we first
#define SEND_CREDS_OK 0
#define SEND_CREDS_NOTSK 1
#define SEND_CREDS_FAIL 2
-static bool recv_creds(int sock, struct ucred *cred, char *v);
static int wait_for_pid(pid_t pid);
-static int send_creds(int sock, struct ucred *cred, char v, bool pingfirst);
static int send_creds_clone_wrapper(void *arg);
/*
return 0;
}
-char *get_pid_cgroup(pid_t pid, const char *contrl)
-{
- int cfd;
-
- cfd = get_cgroup_fd(contrl);
- if (cfd < 0)
- return false;
-
- if (pure_unified_layout(cgroup_ops))
- return cg_unified_get_current_cgroup(pid);
-
- return cg_legacy_get_current_cgroup(pid, contrl);
-}
-
#define INITSCOPE "/init.scope"
void prune_init_slice(char *cg)
{
}
}
-#define POLLIN_SET ( EPOLLIN | EPOLLHUP | EPOLLRDHUP )
-
-static bool wait_for_sock(int sock, int timeout)
-{
- struct epoll_event ev;
- int epfd, ret, now, starttime, deltatime, saved_errno;
-
- if ((starttime = time(NULL)) < 0)
- return false;
-
- if ((epfd = epoll_create(1)) < 0) {
- lxcfs_error("%s\n", "Failed to create epoll socket: %m.");
- return false;
- }
-
- ev.events = POLLIN_SET;
- ev.data.fd = sock;
- if (epoll_ctl(epfd, EPOLL_CTL_ADD, sock, &ev) < 0) {
- lxcfs_error("%s\n", "Failed adding socket to epoll: %m.");
- close(epfd);
- return false;
- }
-
-again:
- if ((now = time(NULL)) < 0) {
- close(epfd);
- return false;
- }
-
- deltatime = (starttime + timeout) - now;
- if (deltatime < 0) { // timeout
- errno = 0;
- close(epfd);
- return false;
- }
- ret = epoll_wait(epfd, &ev, 1, 1000*deltatime + 1);
- if (ret < 0 && errno == EINTR)
- goto again;
- saved_errno = errno;
- close(epfd);
-
- if (ret <= 0) {
- errno = saved_errno;
- return false;
- }
- return true;
-}
-
-static int msgrecv(int sockfd, void *buf, size_t len)
-{
- if (!wait_for_sock(sockfd, 2))
- return -1;
- return recv(sockfd, buf, len, MSG_DONTWAIT);
-}
-
-static int send_creds(int sock, struct ucred *cred, char v, bool pingfirst)
-{
- struct msghdr msg = { 0 };
- struct iovec iov;
- struct cmsghdr *cmsg;
- char cmsgbuf[CMSG_SPACE(sizeof(*cred))];
- char buf[1];
- buf[0] = 'p';
-
- if (pingfirst) {
- if (msgrecv(sock, buf, 1) != 1) {
- lxcfs_error("%s\n", "Error getting reply from server over socketpair.");
- return SEND_CREDS_FAIL;
- }
- }
-
- msg.msg_control = cmsgbuf;
- msg.msg_controllen = sizeof(cmsgbuf);
-
- cmsg = CMSG_FIRSTHDR(&msg);
- cmsg->cmsg_len = CMSG_LEN(sizeof(struct ucred));
- cmsg->cmsg_level = SOL_SOCKET;
- cmsg->cmsg_type = SCM_CREDENTIALS;
- memcpy(CMSG_DATA(cmsg), cred, sizeof(*cred));
-
- msg.msg_name = NULL;
- msg.msg_namelen = 0;
-
- buf[0] = v;
- iov.iov_base = buf;
- iov.iov_len = sizeof(buf);
- msg.msg_iov = &iov;
- msg.msg_iovlen = 1;
-
- if (sendmsg(sock, &msg, 0) < 0) {
- lxcfs_error("Failed at sendmsg: %s.\n",strerror(errno));
- if (errno == 3)
- return SEND_CREDS_NOTSK;
- return SEND_CREDS_FAIL;
- }
-
- return SEND_CREDS_OK;
-}
-
-static bool recv_creds(int sock, struct ucred *cred, char *v)
-{
- struct msghdr msg = { 0 };
- struct iovec iov;
- struct cmsghdr *cmsg;
- char cmsgbuf[CMSG_SPACE(sizeof(*cred))];
- char buf[1];
- int ret;
- int optval = 1;
-
- *v = '1';
-
- cred->pid = -1;
- cred->uid = -1;
- cred->gid = -1;
-
- if (setsockopt(sock, SOL_SOCKET, SO_PASSCRED, &optval, sizeof(optval)) == -1) {
- lxcfs_error("Failed to set passcred: %s\n", strerror(errno));
- return false;
- }
- buf[0] = '1';
- if (write(sock, buf, 1) != 1) {
- lxcfs_error("Failed to start write on scm fd: %s\n", strerror(errno));
- return false;
- }
-
- msg.msg_name = NULL;
- msg.msg_namelen = 0;
- msg.msg_control = cmsgbuf;
- msg.msg_controllen = sizeof(cmsgbuf);
-
- iov.iov_base = buf;
- iov.iov_len = sizeof(buf);
- msg.msg_iov = &iov;
- msg.msg_iovlen = 1;
-
- if (!wait_for_sock(sock, 2)) {
- lxcfs_error("Timed out waiting for scm_cred: %s\n", strerror(errno));
- return false;
- }
- ret = recvmsg(sock, &msg, MSG_DONTWAIT);
- if (ret < 0) {
- lxcfs_error("Failed to receive scm_cred: %s\n", strerror(errno));
- return false;
- }
-
- cmsg = CMSG_FIRSTHDR(&msg);
-
- if (cmsg && cmsg->cmsg_len == CMSG_LEN(sizeof(struct ucred)) &&
- cmsg->cmsg_level == SOL_SOCKET &&
- cmsg->cmsg_type == SCM_CREDENTIALS) {
- memcpy(cred, CMSG_DATA(cmsg), sizeof(*cred));
- }
- *v = buf[0];
-
- return true;
-}
-
struct pid_ns_clone_args {
int *cpipe;
int sock;
int (*wrapped) (int, pid_t); // pid_from_ns or pid_to_ns
};
-static bool startswith(const char *line, const char *pref)
-{
- if (strncmp(line, pref, strlen(pref)) == 0)
- return true;
- return false;
-}
-
-/* Note that "memory.stat" in cgroup2 is hierarchical by default. */
-static void parse_memstat(int version,
- char *memstat,
- unsigned long *cached,
- unsigned long *active_anon,
- unsigned long *inactive_anon,
- unsigned long *active_file,
- unsigned long *inactive_file,
- unsigned long *unevictable,
- unsigned long *shmem)
-{
- char *eol;
-
- while (*memstat) {
- if (startswith(memstat, is_unified_controller(version)
- ? "cache"
- : "total_cache")) {
- sscanf(memstat + 11, "%lu", cached);
- *cached /= 1024;
- } else if (startswith(memstat, is_unified_controller(version)
- ? "active_anon"
- : "total_active_anon")) {
- sscanf(memstat + 17, "%lu", active_anon);
- *active_anon /= 1024;
- } else if (startswith(memstat, is_unified_controller(version)
- ? "inactive_anon"
- : "total_inactive_anon")) {
- sscanf(memstat + 19, "%lu", inactive_anon);
- *inactive_anon /= 1024;
- } else if (startswith(memstat, is_unified_controller(version)
- ? "active_file"
- : "total_active_file")) {
- sscanf(memstat + 17, "%lu", active_file);
- *active_file /= 1024;
- } else if (startswith(memstat, is_unified_controller(version)
- ? "inactive_file"
- : "total_inactive_file")) {
- sscanf(memstat + 19, "%lu", inactive_file);
- *inactive_file /= 1024;
- } else if (startswith(memstat, is_unified_controller(version)
- ? "unevictable"
- : "total_unevictable")) {
- sscanf(memstat + 17, "%lu", unevictable);
- *unevictable /= 1024;
- } else if (startswith(memstat, is_unified_controller(version)
- ? "shmem"
- : "total_shmem")) {
- sscanf(memstat + 11, "%lu", shmem);
- *shmem /= 1024;
- }
- eol = strchr(memstat, '\n');
- if (!eol)
- return;
- memstat = eol+1;
- }
-}
-
-static void get_blkio_io_value(char *str, unsigned major, unsigned minor, char *iotype, unsigned long *v)
-{
- char *eol;
- char key[32];
-
- memset(key, 0, 32);
- snprintf(key, 32, "%u:%u %s", major, minor, iotype);
-
- size_t len = strlen(key);
- *v = 0;
-
- while (*str) {
- if (startswith(str, key)) {
- sscanf(str + len, "%lu", v);
- return;
- }
- eol = strchr(str, '\n');
- if (!eol)
- return;
- str = eol+1;
- }
-}
-
-int read_file_fuse(const char *path, char *buf, size_t size, struct file_info *d)
-{
- __do_free char *line = NULL;
- __do_fclose FILE *f = NULL;
- size_t linelen = 0, total_len = 0;
- char *cache = d->buf;
- size_t cache_size = d->buflen;
-
- f = fopen(path, "r");
- if (!f)
- return 0;
-
- while (getline(&line, &linelen, f) != -1) {
- ssize_t l = snprintf(cache, cache_size, "%s", line);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
- cache += l;
- cache_size -= l;
- total_len += l;
- }
-
- d->size = total_len;
- if (total_len > size)
- total_len = size;
-
- /* read from off 0 */
- memcpy(buf, d->buf, total_len);
-
- if (d->size > total_len)
- d->cached = d->size - total_len;
- return total_len;
-}
-
-/*
- * FUSE ops for /proc
- */
-
-static unsigned long get_memlimit(const char *cgroup, bool swap)
-{
- int ret;
- __do_free char *memlimit_str = NULL;
- unsigned long memlimit = -1;
-
- if (swap)
- ret = cgroup_ops->get_memory_swap_max(cgroup_ops, cgroup, &memlimit_str);
- else
- ret = cgroup_ops->get_memory_max(cgroup_ops, cgroup, &memlimit_str);
- if (ret > 0)
- memlimit = strtoul(memlimit_str, NULL, 10);
-
- return memlimit;
-}
-
-static unsigned long get_min_memlimit(const char *cgroup, bool swap)
-{
- __do_free char *copy = NULL;
- unsigned long memlimit = 0;
- unsigned long retlimit;
-
- copy = strdup(cgroup);
- retlimit = get_memlimit(copy, swap);
-
- while (strcmp(copy, "/") != 0) {
- char *it = copy;
-
- it = dirname(it);
- memlimit = get_memlimit(it, swap);
- if (memlimit != -1 && memlimit < retlimit)
- retlimit = memlimit;
- };
-
- return retlimit;
-}
-
-static int proc_meminfo_read(char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- __do_free char *cgroup = NULL, *line = NULL,
- *memusage_str = NULL, *memstat_str = NULL,
- *memswlimit_str = NULL, *memswusage_str = NULL;
- __do_fclose FILE *f = NULL;
- struct fuse_context *fc = fuse_get_context();
- struct lxcfs_opts *opts = (struct lxcfs_opts *) fuse_get_context()->private_data;
- struct file_info *d = (struct file_info *)fi->fh;
- unsigned long memlimit = 0, memusage = 0, memswlimit = 0,
- memswusage = 0, cached = 0, hosttotal = 0, active_anon = 0,
- inactive_anon = 0, active_file = 0, inactive_file = 0,
- unevictable = 0, shmem = 0, hostswtotal = 0;
- size_t linelen = 0, total_len = 0;
- char *cache = d->buf;
- size_t cache_size = d->buflen;
- int ret;
-
- if (offset) {
- int left;
-
- if (offset > d->size)
- return -EINVAL;
-
- if (!d->cached)
- return 0;
-
- left = d->size - offset;
- total_len = left > size ? size : left;
- memcpy(buf, cache + offset, total_len);
-
- return total_len;
- }
-
- pid_t initpid = lookup_initpid_in_store(fc->pid);
- if (initpid <= 1 || is_shared_pidns(initpid))
- initpid = fc->pid;
-
- cgroup = get_pid_cgroup(initpid, "memory");
- if (!cgroup)
- return read_file_fuse("/proc/meminfo", buf, size, d);
-
- prune_init_slice(cgroup);
-
- memlimit = get_min_memlimit(cgroup, false);
-
- ret = cgroup_ops->get_memory_current(cgroup_ops, cgroup, &memusage_str);
- if (ret < 0)
- return 0;
-
- ret = cgroup_ops->get_memory_stats(cgroup_ops, cgroup, &memstat_str);
- if (ret < 0)
- return 0;
- parse_memstat(ret, memstat_str, &cached, &active_anon, &inactive_anon,
- &active_file, &inactive_file, &unevictable, &shmem);
-
- /*
- * Following values are allowed to fail, because swapaccount might be
- * turned off for current kernel.
- */
- ret = cgroup_ops->get_memory_swap_max(cgroup_ops, cgroup, &memswlimit_str);
- if (ret >= 0)
- ret = cgroup_ops->get_memory_swap_current(cgroup_ops, cgroup, &memswusage_str);
- if (ret >= 0) {
- memswlimit = get_min_memlimit(cgroup, true);
- memswusage = strtoul(memswusage_str, NULL, 10);
- memswlimit = memswlimit / 1024;
- memswusage = memswusage / 1024;
- }
-
- memusage = strtoul(memusage_str, NULL, 10);
- memlimit /= 1024;
- memusage /= 1024;
-
- f = fopen("/proc/meminfo", "r");
- if (!f)
- return 0;
-
- while (getline(&line, &linelen, f) != -1) {
- ssize_t l;
- char *printme, lbuf[100];
-
- memset(lbuf, 0, 100);
- if (startswith(line, "MemTotal:")) {
- sscanf(line+sizeof("MemTotal:")-1, "%lu", &hosttotal);
- if (hosttotal < memlimit)
- memlimit = hosttotal;
- snprintf(lbuf, 100, "MemTotal: %8lu kB\n", memlimit);
- printme = lbuf;
- } else if (startswith(line, "MemFree:")) {
- snprintf(lbuf, 100, "MemFree: %8lu kB\n", memlimit - memusage);
- printme = lbuf;
- } else if (startswith(line, "MemAvailable:")) {
- snprintf(lbuf, 100, "MemAvailable: %8lu kB\n", memlimit - memusage + cached);
- printme = lbuf;
- } else if (startswith(line, "SwapTotal:") && memswlimit > 0 &&
- opts && opts->swap_off == false) {
- sscanf(line+sizeof("SwapTotal:")-1, "%lu", &hostswtotal);
- if (hostswtotal < memswlimit)
- memswlimit = hostswtotal;
- snprintf(lbuf, 100, "SwapTotal: %8lu kB\n", memswlimit);
- printme = lbuf;
- } else if (startswith(line, "SwapTotal:") && opts && opts->swap_off == true) {
- snprintf(lbuf, 100, "SwapTotal: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "SwapFree:") && memswlimit > 0 &&
- memswusage > 0 && opts && opts->swap_off == false) {
- unsigned long swaptotal = memswlimit,
- swapusage = memusage > memswusage
- ? 0
- : memswusage - memusage,
- swapfree = swapusage < swaptotal
- ? swaptotal - swapusage
- : 0;
- snprintf(lbuf, 100, "SwapFree: %8lu kB\n", swapfree);
- printme = lbuf;
- } else if (startswith(line, "SwapFree:") && opts && opts->swap_off == true) {
- snprintf(lbuf, 100, "SwapFree: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "Slab:")) {
- snprintf(lbuf, 100, "Slab: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "Buffers:")) {
- snprintf(lbuf, 100, "Buffers: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "Cached:")) {
- snprintf(lbuf, 100, "Cached: %8lu kB\n", cached);
- printme = lbuf;
- } else if (startswith(line, "SwapCached:")) {
- snprintf(lbuf, 100, "SwapCached: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "Active:")) {
- snprintf(lbuf, 100, "Active: %8lu kB\n",
- active_anon + active_file);
- printme = lbuf;
- } else if (startswith(line, "Inactive:")) {
- snprintf(lbuf, 100, "Inactive: %8lu kB\n",
- inactive_anon + inactive_file);
- printme = lbuf;
- } else if (startswith(line, "Active(anon)")) {
- snprintf(lbuf, 100, "Active(anon): %8lu kB\n", active_anon);
- printme = lbuf;
- } else if (startswith(line, "Inactive(anon)")) {
- snprintf(lbuf, 100, "Inactive(anon): %8lu kB\n", inactive_anon);
- printme = lbuf;
- } else if (startswith(line, "Active(file)")) {
- snprintf(lbuf, 100, "Active(file): %8lu kB\n", active_file);
- printme = lbuf;
- } else if (startswith(line, "Inactive(file)")) {
- snprintf(lbuf, 100, "Inactive(file): %8lu kB\n", inactive_file);
- printme = lbuf;
- } else if (startswith(line, "Unevictable")) {
- snprintf(lbuf, 100, "Unevictable: %8lu kB\n", unevictable);
- printme = lbuf;
- } else if (startswith(line, "SReclaimable")) {
- snprintf(lbuf, 100, "SReclaimable: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "SUnreclaim")) {
- snprintf(lbuf, 100, "SUnreclaim: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "Shmem:")) {
- snprintf(lbuf, 100, "Shmem: %8lu kB\n", shmem);
- printme = lbuf;
- } else if (startswith(line, "ShmemHugePages")) {
- snprintf(lbuf, 100, "ShmemHugePages: %8lu kB\n", 0UL);
- printme = lbuf;
- } else if (startswith(line, "ShmemPmdMapped")) {
- snprintf(lbuf, 100, "ShmemPmdMapped: %8lu kB\n", 0UL);
- printme = lbuf;
- } else
- printme = line;
-
- l = snprintf(cache, cache_size, "%s", printme);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
-
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
-
- cache += l;
- cache_size -= l;
- total_len += l;
- }
-
- d->cached = 1;
- d->size = total_len;
- if (total_len > size ) total_len = size;
- memcpy(buf, d->buf, total_len);
-
- return total_len;
-}
-
-/*
- * Read the cpuset.cpus for cg
- * Return the answer in a newly allocated string which must be freed
- */
-char *get_cpuset(const char *cg)
-{
- char *value = NULL;
- int ret;
-
- ret = cgroup_ops->get_cpuset_cpus(cgroup_ops, cg, &value);
- if (ret < 0)
- return NULL;
-
- return value;
-}
-
-bool cpu_in_cpuset(int cpu, const char *cpuset);
-
-static bool cpuline_in_cpuset(const char *line, const char *cpuset)
-{
- int cpu;
-
- if (sscanf(line, "processor : %d", &cpu) != 1)
- return false;
- return cpu_in_cpuset(cpu, cpuset);
-}
-
-/*
- * Read cgroup CPU quota parameters from `cpu.cfs_quota_us` or `cpu.cfs_period_us`,
- * depending on `param`. Parameter value is returned throuh `value`.
- */
-static bool read_cpu_cfs_param(const char *cg, const char *param, int64_t *value)
-{
- __do_free char *str = NULL;
- char file[11 + 6 + 1]; /* cpu.cfs__us + quota/period + \0 */
-
- snprintf(file, sizeof(file), "cpu.cfs_%s_us", param);
-
- if (!cgroup_ops->get(cgroup_ops, "cpu", cg, file, &str))
- return false;
-
- if (sscanf(str, "%ld", value) != 1)
- return false;
-
- return true;
-}
-
-/*
- * Return the maximum number of visible CPUs based on CPU quotas.
- * If there is no quota set, zero is returned.
- */
-int max_cpu_count(const char *cg)
-{
- int rv, nprocs;
- int64_t cfs_quota, cfs_period;
- int nr_cpus_in_cpuset = 0;
- char *cpuset = NULL;
-
- if (!read_cpu_cfs_param(cg, "quota", &cfs_quota))
- return 0;
-
- if (!read_cpu_cfs_param(cg, "period", &cfs_period))
- return 0;
-
- cpuset = get_cpuset(cg);
- if (cpuset)
- nr_cpus_in_cpuset = cpu_number_in_cpuset(cpuset);
-
- if (cfs_quota <= 0 || cfs_period <= 0){
- if (nr_cpus_in_cpuset > 0)
- return nr_cpus_in_cpuset;
-
- return 0;
- }
-
- rv = cfs_quota / cfs_period;
-
- /* In case quota/period does not yield a whole number, add one CPU for
- * the remainder.
- */
- if ((cfs_quota % cfs_period) > 0)
- rv += 1;
-
- nprocs = get_nprocs();
-
- if (rv > nprocs)
- rv = nprocs;
-
- /* use min value in cpu quota and cpuset */
- if (nr_cpus_in_cpuset > 0 && nr_cpus_in_cpuset < rv)
- rv = nr_cpus_in_cpuset;
-
- return rv;
-}
-
-/*
- * Return the exact number of visible CPUs based on CPU quotas.
- * If there is no quota set, zero is returned.
- */
-static double exact_cpu_count(const char *cg)
-{
- double rv;
- int nprocs;
- int64_t cfs_quota, cfs_period;
-
- if (!read_cpu_cfs_param(cg, "quota", &cfs_quota))
- return 0;
-
- if (!read_cpu_cfs_param(cg, "period", &cfs_period))
- return 0;
-
- if (cfs_quota <= 0 || cfs_period <= 0)
- return 0;
-
- rv = (double)cfs_quota / (double)cfs_period;
-
- nprocs = get_nprocs();
-
- if (rv > nprocs)
- rv = nprocs;
-
- return rv;
-}
-
-/*
- * check whether this is a '^processor" line in /proc/cpuinfo
- */
-static bool is_processor_line(const char *line)
-{
- int cpu;
-
- if (sscanf(line, "processor : %d", &cpu) == 1)
- return true;
- return false;
-}
-
-static int proc_cpuinfo_read(char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- __do_free char *cg = NULL, *cpuset = NULL, *line = NULL;
- __do_fclose FILE *f = NULL;
- struct fuse_context *fc = fuse_get_context();
- struct file_info *d = (struct file_info *)fi->fh;
- size_t linelen = 0, total_len = 0;
- bool am_printing = false, firstline = true, is_s390x = false;
- int curcpu = -1, cpu, max_cpus = 0;
- bool use_view;
- char *cache = d->buf;
- size_t cache_size = d->buflen;
-
- if (offset){
- int left;
-
- if (offset > d->size)
- return -EINVAL;
-
- if (!d->cached)
- return 0;
-
- left = d->size - offset;
- total_len = left > size ? size: left;
- memcpy(buf, cache + offset, total_len);
-
- return total_len;
- }
-
- pid_t initpid = lookup_initpid_in_store(fc->pid);
- if (initpid <= 1 || is_shared_pidns(initpid))
- initpid = fc->pid;
- cg = get_pid_cgroup(initpid, "cpuset");
- if (!cg)
- return read_file_fuse("proc/cpuinfo", buf, size, d);
- prune_init_slice(cg);
-
- cpuset = get_cpuset(cg);
- if (!cpuset)
- return 0;
-
- use_view = cgroup_ops->can_use_cpuview(cgroup_ops);
- if (use_view)
- max_cpus = max_cpu_count(cg);
-
- f = fopen("/proc/cpuinfo", "r");
- if (!f)
- return 0;
-
- while (getline(&line, &linelen, f) != -1) {
- ssize_t l;
- if (firstline) {
- firstline = false;
- if (strstr(line, "IBM/S390") != NULL) {
- is_s390x = true;
- am_printing = true;
- continue;
- }
- }
- if (strncmp(line, "# processors:", 12) == 0)
- continue;
- if (is_processor_line(line)) {
- if (use_view && max_cpus > 0 && (curcpu+1) == max_cpus)
- break;
- am_printing = cpuline_in_cpuset(line, cpuset);
- if (am_printing) {
- curcpu ++;
- l = snprintf(cache, cache_size, "processor : %d\n", curcpu);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
- cache += l;
- cache_size -= l;
- total_len += l;
- }
- continue;
- } else if (is_s390x && sscanf(line, "processor %d:", &cpu) == 1) {
- char *p;
- if (use_view && max_cpus > 0 && (curcpu+1) == max_cpus)
- break;
- if (!cpu_in_cpuset(cpu, cpuset))
- continue;
- curcpu ++;
- p = strchr(line, ':');
- if (!p || !*p)
- return 0;
- p++;
- l = snprintf(cache, cache_size, "processor %d:%s", curcpu, p);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
- cache += l;
- cache_size -= l;
- total_len += l;
- continue;
-
- }
- if (am_printing) {
- l = snprintf(cache, cache_size, "%s", line);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
- cache += l;
- cache_size -= l;
- total_len += l;
- }
- }
-
- if (is_s390x) {
- __do_free char *origcache = d->buf;
- ssize_t l;
-
- d->buf = malloc(d->buflen);
- if (!d->buf) {
- d->buf = move_ptr(origcache);
- return 0;
- }
-
- cache = d->buf;
- cache_size = d->buflen;
- total_len = 0;
- l = snprintf(cache, cache_size, "vendor_id : IBM/S390\n");
- if (l < 0 || l >= cache_size)
- return 0;
-
- cache_size -= l;
- cache += l;
- total_len += l;
- l = snprintf(cache, cache_size, "# processors : %d\n", curcpu + 1);
- if (l < 0 || l >= cache_size)
- return 0;
-
- cache_size -= l;
- cache += l;
- total_len += l;
- l = snprintf(cache, cache_size, "%s", origcache);
- if (l < 0 || l >= cache_size)
- return 0;
- total_len += l;
- }
-
- d->cached = 1;
- d->size = total_len;
- if (total_len > size ) total_len = size;
-
- /* read from off 0 */
- memcpy(buf, d->buf, total_len);
- return total_len;
-}
-
-static uint64_t get_reaper_start_time(pid_t pid)
-{
- int ret;
- FILE *f;
- uint64_t starttime;
- /* strlen("/proc/") = 6
- * +
- * LXCFS_NUMSTRLEN64
- * +
- * strlen("/stat") = 5
- * +
- * \0 = 1
- * */
-#define __PROC_PID_STAT_LEN (6 + LXCFS_NUMSTRLEN64 + 5 + 1)
- char path[__PROC_PID_STAT_LEN];
- pid_t qpid;
-
- qpid = lookup_initpid_in_store(pid);
- if (qpid <= 0) {
- /* Caller can check for EINVAL on 0. */
- errno = EINVAL;
- return 0;
- }
-
- ret = snprintf(path, __PROC_PID_STAT_LEN, "/proc/%d/stat", qpid);
- if (ret < 0 || ret >= __PROC_PID_STAT_LEN) {
- /* Caller can check for EINVAL on 0. */
- errno = EINVAL;
- return 0;
- }
-
- f = fopen(path, "r");
- if (!f) {
- /* Caller can check for EINVAL on 0. */
- errno = EINVAL;
- return 0;
- }
-
- /* Note that the *scanf() argument supression requires that length
- * modifiers such as "l" are omitted. Otherwise some compilers will yell
- * at us. It's like telling someone you're not married and then asking
- * if you can bring your wife to the party.
- */
- ret = fscanf(f, "%*d " /* (1) pid %d */
- "%*s " /* (2) comm %s */
- "%*c " /* (3) state %c */
- "%*d " /* (4) ppid %d */
- "%*d " /* (5) pgrp %d */
- "%*d " /* (6) session %d */
- "%*d " /* (7) tty_nr %d */
- "%*d " /* (8) tpgid %d */
- "%*u " /* (9) flags %u */
- "%*u " /* (10) minflt %lu */
- "%*u " /* (11) cminflt %lu */
- "%*u " /* (12) majflt %lu */
- "%*u " /* (13) cmajflt %lu */
- "%*u " /* (14) utime %lu */
- "%*u " /* (15) stime %lu */
- "%*d " /* (16) cutime %ld */
- "%*d " /* (17) cstime %ld */
- "%*d " /* (18) priority %ld */
- "%*d " /* (19) nice %ld */
- "%*d " /* (20) num_threads %ld */
- "%*d " /* (21) itrealvalue %ld */
- "%" PRIu64, /* (22) starttime %llu */
- &starttime);
- if (ret != 1) {
- fclose(f);
- /* Caller can check for EINVAL on 0. */
- errno = EINVAL;
- return 0;
- }
-
- fclose(f);
-
- errno = 0;
- return starttime;
-}
-
-static double get_reaper_start_time_in_sec(pid_t pid)
-{
- uint64_t clockticks, ticks_per_sec;
- int64_t ret;
- double res = 0;
-
- clockticks = get_reaper_start_time(pid);
- if (clockticks == 0 && errno == EINVAL) {
- lxcfs_debug("failed to retrieve start time of pid %d\n", pid);
- return 0;
- }
-
- ret = sysconf(_SC_CLK_TCK);
- if (ret < 0 && errno == EINVAL) {
- lxcfs_debug(
- "%s\n",
- "failed to determine number of clock ticks in a second");
- return 0;
- }
-
- ticks_per_sec = (uint64_t)ret;
- res = (double)clockticks / ticks_per_sec;
- return res;
-}
-
-static double get_reaper_age(pid_t pid)
-{
- uint64_t uptime_ms;
- double procstart, procage;
-
- /* We need to substract the time the process has started since system
- * boot minus the time when the system has started to get the actual
- * reaper age.
- */
- procstart = get_reaper_start_time_in_sec(pid);
- procage = procstart;
- if (procstart > 0) {
- int ret;
- struct timespec spec;
-
- ret = clock_gettime(CLOCK_BOOTTIME, &spec);
- if (ret < 0)
- return 0;
-
- /* We could make this more precise here by using the tv_nsec
- * field in the timespec struct and convert it to milliseconds
- * and then create a double for the seconds and milliseconds but
- * that seems more work than it is worth.
- */
- uptime_ms = (spec.tv_sec * 1000) + (spec.tv_nsec * 1e-6);
- procage = (uptime_ms - (procstart * 1000)) / 1000;
- }
-
- return procage;
-}
-
-/*
- * Returns 0 on success.
- * It is the caller's responsibility to free `return_usage`, unless this
- * function returns an error.
- */
-static int read_cpuacct_usage_all(char *cg, char *cpuset, struct cpuacct_usage **return_usage, int *size)
-{
- __do_free char *usage_str = NULL;
- __do_free struct cpuacct_usage *cpu_usage = NULL;
- int cpucount = get_nprocs_conf();
- int read_pos = 0, read_cnt=0;
- int i, j, ret;
- int cg_cpu;
- uint64_t cg_user, cg_system;
- int64_t ticks_per_sec;
-
- ticks_per_sec = sysconf(_SC_CLK_TCK);
-
- if (ticks_per_sec < 0 && errno == EINVAL) {
- lxcfs_v(
- "%s\n",
- "read_cpuacct_usage_all failed to determine number of clock ticks "
- "in a second");
- return -1;
- }
-
- cpu_usage = malloc(sizeof(struct cpuacct_usage) * cpucount);
- if (!cpu_usage)
- return -ENOMEM;
-
- memset(cpu_usage, 0, sizeof(struct cpuacct_usage) * cpucount);
- if (!cgroup_ops->get(cgroup_ops, "cpuacct", cg, "cpuacct.usage_all", &usage_str)) {
- char *data = NULL;
- int i = 0, read_pos = 0, read_cnt=0;
- size_t sz = 0, asz = 0;
-
- /* read cpuacct.usage_percpu instead. */
- lxcfs_v("failed to read cpuacct.usage_all. reading cpuacct.usage_percpu instead\n%s", "");
- if (!cgroup_ops->get(cgroup_ops, "cpuacct", cg, "cpuacct.usage_percpu", &usage_str))
- return -1;
- lxcfs_v("usage_str: %s\n", usage_str);
-
- /* convert cpuacct.usage_percpu into cpuacct.usage_all. */
- lxcfs_v("converting cpuacct.usage_percpu into cpuacct.usage_all\n%s", "");
-
- must_strcat(&data, &sz, &asz, "cpu user system\n");
-
- while (sscanf(usage_str + read_pos, "%lu %n", &cg_user, &read_cnt) > 0) {
- lxcfs_debug("i: %d, cg_user: %lu, read_pos: %d, read_cnt: %d\n", i, cg_user, read_pos, read_cnt);
- must_strcat(&data, &sz, &asz, "%d %lu 0\n", i, cg_user);
- i++;
- read_pos += read_cnt;
- }
-
- usage_str = data;
-
- lxcfs_v("usage_str: %s\n", usage_str);
- }
-
- if (sscanf(usage_str, "cpu user system\n%n", &read_cnt) != 0) {
- lxcfs_error("read_cpuacct_usage_all reading first line from "
- "%s/cpuacct.usage_all failed.\n", cg);
- return -1;
- }
-
- read_pos += read_cnt;
-
- for (i = 0, j = 0; i < cpucount; i++) {
- ret = sscanf(usage_str + read_pos, "%d %lu %lu\n%n", &cg_cpu, &cg_user,
- &cg_system, &read_cnt);
-
- if (ret == EOF)
- break;
-
- if (ret != 3) {
- lxcfs_error("read_cpuacct_usage_all reading from %s/cpuacct.usage_all "
- "failed.\n", cg);
- return -1;
- }
-
- read_pos += read_cnt;
-
- /* Convert the time from nanoseconds to USER_HZ */
- cpu_usage[j].user = cg_user / 1000.0 / 1000 / 1000 * ticks_per_sec;
- cpu_usage[j].system = cg_system / 1000.0 / 1000 / 1000 * ticks_per_sec;
- j++;
- }
-
- *return_usage = move_ptr(cpu_usage);
- *size = cpucount;
- return 0;
-}
-
-static unsigned long diff_cpu_usage(struct cpuacct_usage *older, struct cpuacct_usage *newer, struct cpuacct_usage *diff, int cpu_count)
-{
- int i;
- unsigned long sum = 0;
-
- for (i = 0; i < cpu_count; i++) {
- if (!newer[i].online)
- continue;
-
- /* When cpuset is changed on the fly, the CPUs might get reordered.
- * We could either reset all counters, or check that the substractions
- * below will return expected results.
- */
- if (newer[i].user > older[i].user)
- diff[i].user = newer[i].user - older[i].user;
- else
- diff[i].user = 0;
-
- if (newer[i].system > older[i].system)
- diff[i].system = newer[i].system - older[i].system;
- else
- diff[i].system = 0;
-
- if (newer[i].idle > older[i].idle)
- diff[i].idle = newer[i].idle - older[i].idle;
- else
- diff[i].idle = 0;
-
- sum += diff[i].user;
- sum += diff[i].system;
- sum += diff[i].idle;
- }
-
- return sum;
-}
-
-static void add_cpu_usage(unsigned long *surplus, struct cpuacct_usage *usage, unsigned long *counter, unsigned long threshold)
-{
- unsigned long free_space, to_add;
-
- free_space = threshold - usage->user - usage->system;
-
- if (free_space > usage->idle)
- free_space = usage->idle;
-
- to_add = free_space > *surplus ? *surplus : free_space;
-
- *counter += to_add;
- usage->idle -= to_add;
- *surplus -= to_add;
-}
-
-static struct cg_proc_stat *prune_proc_stat_list(struct cg_proc_stat *node)
-{
- struct cg_proc_stat *first = NULL, *prev, *tmp;
-
- for (prev = NULL; node; ) {
- if (!cgfs_param_exist("cpu", node->cg, "cpu.shares")) {
- tmp = node;
- lxcfs_debug("Removing stat node for %s\n", node->cg);
-
- if (prev)
- prev->next = node->next;
- else
- first = node->next;
-
- node = node->next;
- free_proc_stat_node(tmp);
- } else {
- if (!first)
- first = node;
- prev = node;
- node = node->next;
- }
- }
-
- return first;
-}
-
-#define PROC_STAT_PRUNE_INTERVAL 10
-static void prune_proc_stat_history(void)
-{
- int i;
- time_t now = time(NULL);
-
- for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
- pthread_rwlock_wrlock(&proc_stat_history[i]->lock);
-
- if ((proc_stat_history[i]->lastcheck + PROC_STAT_PRUNE_INTERVAL) > now) {
- pthread_rwlock_unlock(&proc_stat_history[i]->lock);
- return;
- }
-
- if (proc_stat_history[i]->next) {
- proc_stat_history[i]->next = prune_proc_stat_list(proc_stat_history[i]->next);
- proc_stat_history[i]->lastcheck = now;
- }
-
- pthread_rwlock_unlock(&proc_stat_history[i]->lock);
- }
-}
-
-static struct cg_proc_stat *find_proc_stat_node(struct cg_proc_stat_head *head, const char *cg)
-{
- struct cg_proc_stat *node;
-
- pthread_rwlock_rdlock(&head->lock);
-
- if (!head->next) {
- pthread_rwlock_unlock(&head->lock);
- return NULL;
- }
-
- node = head->next;
-
- do {
- if (strcmp(cg, node->cg) == 0)
- goto out;
- } while ((node = node->next));
-
- node = NULL;
-
-out:
- pthread_rwlock_unlock(&head->lock);
- prune_proc_stat_history();
- return node;
-}
-
-static struct cg_proc_stat *new_proc_stat_node(struct cpuacct_usage *usage, int cpu_count, const char *cg)
-{
- struct cg_proc_stat *node;
- int i;
-
- node = malloc(sizeof(struct cg_proc_stat));
- if (!node)
- goto err;
-
- node->cg = NULL;
- node->usage = NULL;
- node->view = NULL;
-
- node->cg = malloc(strlen(cg) + 1);
- if (!node->cg)
- goto err;
-
- strcpy(node->cg, cg);
-
- node->usage = malloc(sizeof(struct cpuacct_usage) * cpu_count);
- if (!node->usage)
- goto err;
-
- memcpy(node->usage, usage, sizeof(struct cpuacct_usage) * cpu_count);
-
- node->view = malloc(sizeof(struct cpuacct_usage) * cpu_count);
- if (!node->view)
- goto err;
-
- node->cpu_count = cpu_count;
- node->next = NULL;
-
- if (pthread_mutex_init(&node->lock, NULL) != 0) {
- lxcfs_error("%s\n", "Failed to initialize node lock");
- goto err;
- }
-
- for (i = 0; i < cpu_count; i++) {
- node->view[i].user = 0;
- node->view[i].system = 0;
- node->view[i].idle = 0;
- }
-
- return node;
-
-err:
- if (node && node->cg)
- free(node->cg);
- if (node && node->usage)
- free(node->usage);
- if (node && node->view)
- free(node->view);
- if (node)
- free(node);
-
- return NULL;
-}
-
-static struct cg_proc_stat *add_proc_stat_node(struct cg_proc_stat *new_node)
-{
- int hash = calc_hash(new_node->cg) % CPUVIEW_HASH_SIZE;
- struct cg_proc_stat_head *head = proc_stat_history[hash];
- struct cg_proc_stat *node, *rv = new_node;
-
- pthread_rwlock_wrlock(&head->lock);
-
- if (!head->next) {
- head->next = new_node;
- goto out;
- }
-
- node = head->next;
-
- for (;;) {
- if (strcmp(node->cg, new_node->cg) == 0) {
- /* The node is already present, return it */
- free_proc_stat_node(new_node);
- rv = node;
- goto out;
- }
-
- if (node->next) {
- node = node->next;
- continue;
- }
-
- node->next = new_node;
- goto out;
- }
-
-out:
- pthread_rwlock_unlock(&head->lock);
- return rv;
-}
-
-static bool expand_proc_stat_node(struct cg_proc_stat *node, int cpu_count)
-{
- __do_free struct cpuacct_usage *new_usage = NULL, *new_view = NULL;
-
- /* Allocate new memory */
- new_usage = malloc(sizeof(struct cpuacct_usage) * cpu_count);
- if (!new_usage)
- return false;
-
- new_view = malloc(sizeof(struct cpuacct_usage) * cpu_count);
- if (!new_view)
- return false;
-
- /* Copy existing data & initialize new elements */
- for (int i = 0; i < cpu_count; i++) {
- if (i < node->cpu_count) {
- new_usage[i].user = node->usage[i].user;
- new_usage[i].system = node->usage[i].system;
- new_usage[i].idle = node->usage[i].idle;
-
- new_view[i].user = node->view[i].user;
- new_view[i].system = node->view[i].system;
- new_view[i].idle = node->view[i].idle;
- } else {
- new_usage[i].user = 0;
- new_usage[i].system = 0;
- new_usage[i].idle = 0;
-
- new_view[i].user = 0;
- new_view[i].system = 0;
- new_view[i].idle = 0;
- }
- }
-
- free(node->usage);
- node->usage = move_ptr(new_usage);
-
- free(node->view);
- node->view = move_ptr(new_view);
- node->cpu_count = cpu_count;
-
- return true;
-}
-
-static struct cg_proc_stat *find_or_create_proc_stat_node(struct cpuacct_usage *usage, int cpu_count, const char *cg)
-{
- int hash = calc_hash(cg) % CPUVIEW_HASH_SIZE;
- struct cg_proc_stat_head *head = proc_stat_history[hash];
- struct cg_proc_stat *node;
-
- node = find_proc_stat_node(head, cg);
-
- if (!node) {
- node = new_proc_stat_node(usage, cpu_count, cg);
- if (!node)
- return NULL;
-
- node = add_proc_stat_node(node);
- lxcfs_debug("New stat node (%d) for %s\n", cpu_count, cg);
- }
-
- pthread_mutex_lock(&node->lock);
-
- /* If additional CPUs on the host have been enabled, CPU usage counter
- * arrays have to be expanded */
- if (node->cpu_count < cpu_count) {
- lxcfs_debug("Expanding stat node %d->%d for %s\n",
- node->cpu_count, cpu_count, cg);
-
- if (!expand_proc_stat_node(node, cpu_count)) {
- pthread_mutex_unlock(&node->lock);
- lxcfs_debug("Unable to expand stat node %d->%d for %s\n",
- node->cpu_count, cpu_count, cg);
- return NULL;
- }
- }
-
- return node;
-}
-
-static void reset_proc_stat_node(struct cg_proc_stat *node, struct cpuacct_usage *usage, int cpu_count)
-{
- int i;
-
- lxcfs_debug("Resetting stat node for %s\n", node->cg);
- memcpy(node->usage, usage, sizeof(struct cpuacct_usage) * cpu_count);
-
- for (i = 0; i < cpu_count; i++) {
- node->view[i].user = 0;
- node->view[i].system = 0;
- node->view[i].idle = 0;
- }
-
- node->cpu_count = cpu_count;
-}
-
-static int cpuview_proc_stat(const char *cg, const char *cpuset,
- struct cpuacct_usage *cg_cpu_usage,
- int cg_cpu_usage_size, FILE *f, char *buf,
- size_t buf_size)
-{
- __do_free char *line = NULL;
- __do_free struct cpuacct_usage *diff = NULL;
- size_t linelen = 0, total_len = 0, l;
- int curcpu = -1; /* cpu numbering starts at 0 */
- int physcpu, i;
- int max_cpus = max_cpu_count(cg), cpu_cnt = 0;
- unsigned long user = 0, nice = 0, system = 0, idle = 0, iowait = 0,
- irq = 0, softirq = 0, steal = 0, guest = 0, guest_nice = 0;
- unsigned long user_sum = 0, system_sum = 0, idle_sum = 0;
- unsigned long user_surplus = 0, system_surplus = 0;
- unsigned long total_sum, threshold;
- struct cg_proc_stat *stat_node;
- int nprocs = get_nprocs_conf();
-
- if (cg_cpu_usage_size < nprocs)
- nprocs = cg_cpu_usage_size;
-
- /* Read all CPU stats and stop when we've encountered other lines */
- while (getline(&line, &linelen, f) != -1) {
- int ret;
- char cpu_char[10]; /* That's a lot of cores */
- uint64_t all_used, cg_used;
-
- if (strlen(line) == 0)
- continue;
-
- /* not a ^cpuN line containing a number N */
- if (sscanf(line, "cpu%9[^ ]", cpu_char) != 1)
- break;
-
- if (sscanf(cpu_char, "%d", &physcpu) != 1)
- continue;
-
- if (physcpu >= cg_cpu_usage_size)
- continue;
-
- curcpu ++;
- cpu_cnt ++;
-
- if (!cpu_in_cpuset(physcpu, cpuset)) {
- for (i = curcpu; i <= physcpu; i++)
- cg_cpu_usage[i].online = false;
- continue;
- }
-
- if (curcpu < physcpu) {
- /* Some CPUs may be disabled */
- for (i = curcpu; i < physcpu; i++)
- cg_cpu_usage[i].online = false;
-
- curcpu = physcpu;
- }
-
- cg_cpu_usage[curcpu].online = true;
-
- ret = sscanf(line, "%*s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
- &user,
- &nice,
- &system,
- &idle,
- &iowait,
- &irq,
- &softirq,
- &steal,
- &guest,
- &guest_nice);
-
- if (ret != 10)
- continue;
-
- all_used = user + nice + system + iowait + irq + softirq + steal + guest + guest_nice;
- cg_used = cg_cpu_usage[curcpu].user + cg_cpu_usage[curcpu].system;
-
- if (all_used >= cg_used) {
- cg_cpu_usage[curcpu].idle = idle + (all_used - cg_used);
-
- } else {
- lxcfs_error("cpu%d from %s has unexpected cpu time: %lu in /proc/stat, "
- "%lu in cpuacct.usage_all; unable to determine idle time\n",
- curcpu, cg, all_used, cg_used);
- cg_cpu_usage[curcpu].idle = idle;
- }
- }
-
- /* Cannot use more CPUs than is available due to cpuset */
- if (max_cpus > cpu_cnt)
- max_cpus = cpu_cnt;
-
- stat_node = find_or_create_proc_stat_node(cg_cpu_usage, nprocs, cg);
-
- if (!stat_node) {
- lxcfs_error("unable to find/create stat node for %s\n", cg);
- return 0;
- }
-
- diff = malloc(sizeof(struct cpuacct_usage) * nprocs);
- if (!diff) {
- return 0;
- }
-
- /*
- * If the new values are LOWER than values stored in memory, it means
- * the cgroup has been reset/recreated and we should reset too.
- */
- for (curcpu = 0; curcpu < nprocs; curcpu++) {
- if (!cg_cpu_usage[curcpu].online)
- continue;
-
- if (cg_cpu_usage[curcpu].user < stat_node->usage[curcpu].user)
- reset_proc_stat_node(stat_node, cg_cpu_usage, nprocs);
-
- break;
- }
-
- total_sum = diff_cpu_usage(stat_node->usage, cg_cpu_usage, diff, nprocs);
-
- for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
- stat_node->usage[curcpu].online = cg_cpu_usage[curcpu].online;
-
- if (!stat_node->usage[curcpu].online)
- continue;
-
- i++;
-
- stat_node->usage[curcpu].user += diff[curcpu].user;
- stat_node->usage[curcpu].system += diff[curcpu].system;
- stat_node->usage[curcpu].idle += diff[curcpu].idle;
-
- if (max_cpus > 0 && i >= max_cpus) {
- user_surplus += diff[curcpu].user;
- system_surplus += diff[curcpu].system;
- }
- }
-
- /* Calculate usage counters of visible CPUs */
- if (max_cpus > 0) {
- unsigned long diff_user = 0;
- unsigned long diff_system = 0;
- unsigned long diff_idle = 0;
- unsigned long max_diff_idle = 0;
- unsigned long max_diff_idle_index = 0;
- double exact_cpus;
-
- /* threshold = maximum usage per cpu, including idle */
- threshold = total_sum / cpu_cnt * max_cpus;
-
- for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
- if (!stat_node->usage[curcpu].online)
- continue;
-
- i++;
-
- if (i == max_cpus)
- break;
-
- if (diff[curcpu].user + diff[curcpu].system >= threshold)
- continue;
-
- /* Add user */
- add_cpu_usage(&user_surplus, &diff[curcpu],
- &diff[curcpu].user, threshold);
-
- if (diff[curcpu].user + diff[curcpu].system >= threshold)
- continue;
-
- /* If there is still room, add system */
- add_cpu_usage(&system_surplus, &diff[curcpu],
- &diff[curcpu].system, threshold);
- }
-
- if (user_surplus > 0)
- lxcfs_debug("leftover user: %lu for %s\n", user_surplus, cg);
- if (system_surplus > 0)
- lxcfs_debug("leftover system: %lu for %s\n", system_surplus, cg);
-
- for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
- if (!stat_node->usage[curcpu].online)
- continue;
-
- i++;
-
- if (i == max_cpus)
- break;
-
- stat_node->view[curcpu].user += diff[curcpu].user;
- stat_node->view[curcpu].system += diff[curcpu].system;
- stat_node->view[curcpu].idle += diff[curcpu].idle;
-
- user_sum += stat_node->view[curcpu].user;
- system_sum += stat_node->view[curcpu].system;
- idle_sum += stat_node->view[curcpu].idle;
-
- diff_user += diff[curcpu].user;
- diff_system += diff[curcpu].system;
- diff_idle += diff[curcpu].idle;
- if (diff[curcpu].idle > max_diff_idle) {
- max_diff_idle = diff[curcpu].idle;
- max_diff_idle_index = curcpu;
- }
-
- lxcfs_v("curcpu: %d, diff_user: %lu, diff_system: %lu, diff_idle: %lu\n", curcpu, diff[curcpu].user, diff[curcpu].system, diff[curcpu].idle);
- }
- lxcfs_v("total. diff_user: %lu, diff_system: %lu, diff_idle: %lu\n", diff_user, diff_system, diff_idle);
-
- /* revise cpu usage view to support partial cpu case. */
- exact_cpus = exact_cpu_count(cg);
- if (exact_cpus < (double)max_cpus){
- unsigned long delta = (unsigned long)((double)(diff_user + diff_system + diff_idle) * (1 - exact_cpus / (double)max_cpus));
-
- lxcfs_v("revising cpu usage view to match the exact cpu count [%f]\n", exact_cpus);
- lxcfs_v("delta: %lu\n", delta);
- lxcfs_v("idle_sum before: %lu\n", idle_sum);
- idle_sum = idle_sum > delta ? idle_sum - delta : 0;
- lxcfs_v("idle_sum after: %lu\n", idle_sum);
-
- curcpu = max_diff_idle_index;
- lxcfs_v("curcpu: %d, idle before: %lu\n", curcpu, stat_node->view[curcpu].idle);
- stat_node->view[curcpu].idle = stat_node->view[curcpu].idle > delta ? stat_node->view[curcpu].idle - delta : 0;
- lxcfs_v("curcpu: %d, idle after: %lu\n", curcpu, stat_node->view[curcpu].idle);
- }
- } else {
- for (curcpu = 0; curcpu < nprocs; curcpu++) {
- if (!stat_node->usage[curcpu].online)
- continue;
-
- stat_node->view[curcpu].user = stat_node->usage[curcpu].user;
- stat_node->view[curcpu].system = stat_node->usage[curcpu].system;
- stat_node->view[curcpu].idle = stat_node->usage[curcpu].idle;
-
- user_sum += stat_node->view[curcpu].user;
- system_sum += stat_node->view[curcpu].system;
- idle_sum += stat_node->view[curcpu].idle;
- }
- }
-
- /* Render the file */
- /* cpu-all */
- l = snprintf(buf, buf_size, "cpu %lu 0 %lu %lu 0 0 0 0 0 0\n",
- user_sum,
- system_sum,
- idle_sum);
- lxcfs_v("cpu-all: %s\n", buf);
-
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
- }
- if (l >= buf_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
-
- buf += l;
- buf_size -= l;
- total_len += l;
-
- /* Render visible CPUs */
- for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
- if (!stat_node->usage[curcpu].online)
- continue;
-
- i++;
-
- if (max_cpus > 0 && i == max_cpus)
- break;
-
- l = snprintf(buf, buf_size, "cpu%d %lu 0 %lu %lu 0 0 0 0 0 0\n",
- i,
- stat_node->view[curcpu].user,
- stat_node->view[curcpu].system,
- stat_node->view[curcpu].idle);
- lxcfs_v("cpu: %s\n", buf);
-
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
-
- }
- if (l >= buf_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
-
- buf += l;
- buf_size -= l;
- total_len += l;
- }
-
- /* Pass the rest of /proc/stat, start with the last line read */
- l = snprintf(buf, buf_size, "%s", line);
-
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
-
- }
- if (l >= buf_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
-
- buf += l;
- buf_size -= l;
- total_len += l;
-
- /* Pass the rest of the host's /proc/stat */
- while (getline(&line, &linelen, f) != -1) {
- l = snprintf(buf, buf_size, "%s", line);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
- }
- if (l >= buf_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
- buf += l;
- buf_size -= l;
- total_len += l;
- }
-
- if (stat_node)
- pthread_mutex_unlock(&stat_node->lock);
- return total_len;
-}
-
-#define CPUALL_MAX_SIZE (BUF_RESERVE_SIZE / 2)
-static int proc_stat_read(char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- __do_free char *cg = NULL, *cpuset = NULL, *line = NULL;
- __do_free struct cpuacct_usage *cg_cpu_usage = NULL;
- __do_fclose FILE *f = NULL;
- struct fuse_context *fc = fuse_get_context();
- struct file_info *d = (struct file_info *)fi->fh;
- size_t linelen = 0, total_len = 0;
- int curcpu = -1; /* cpu numbering starts at 0 */
- int physcpu = 0;
- unsigned long user = 0, nice = 0, system = 0, idle = 0, iowait = 0,
- irq = 0, softirq = 0, steal = 0, guest = 0, guest_nice = 0;
- unsigned long user_sum = 0, nice_sum = 0, system_sum = 0, idle_sum = 0,
- iowait_sum = 0, irq_sum = 0, softirq_sum = 0,
- steal_sum = 0, guest_sum = 0, guest_nice_sum = 0;
- char cpuall[CPUALL_MAX_SIZE];
- /* reserve for cpu all */
- char *cache = d->buf + CPUALL_MAX_SIZE;
- size_t cache_size = d->buflen - CPUALL_MAX_SIZE;
- int cg_cpu_usage_size = 0;
-
- if (offset){
- if (offset > d->size)
- return -EINVAL;
- if (!d->cached)
- return 0;
- int left = d->size - offset;
- total_len = left > size ? size: left;
- memcpy(buf, d->buf + offset, total_len);
- return total_len;
- }
-
- pid_t initpid = lookup_initpid_in_store(fc->pid);
- lxcfs_v("initpid: %d\n", initpid);
- if (initpid <= 0)
- initpid = fc->pid;
-
- /*
- * when container run with host pid namespace initpid == 1, cgroup will "/"
- * we should return host os's /proc contents.
- * in some case cpuacct_usage.all in "/" will larger then /proc/stat
- */
- if (initpid == 1) {
- return read_file_fuse("/proc/stat", buf, size, d);
- }
-
- cg = get_pid_cgroup(initpid, "cpuset");
- lxcfs_v("cg: %s\n", cg);
- if (!cg)
- return read_file_fuse("/proc/stat", buf, size, d);
- prune_init_slice(cg);
-
- cpuset = get_cpuset(cg);
- if (!cpuset)
- return 0;
-
- /*
- * Read cpuacct.usage_all for all CPUs.
- * If the cpuacct cgroup is present, it is used to calculate the container's
- * CPU usage. If not, values from the host's /proc/stat are used.
- */
- if (read_cpuacct_usage_all(cg, cpuset, &cg_cpu_usage, &cg_cpu_usage_size) != 0) {
- lxcfs_v("%s\n", "proc_stat_read failed to read from cpuacct, "
- "falling back to the host's /proc/stat");
- }
-
- f = fopen("/proc/stat", "r");
- if (!f)
- return 0;
-
- //skip first line
- if (getline(&line, &linelen, f) < 0) {
- lxcfs_error("%s\n", "proc_stat_read read first line failed.");
- return 0;
- }
-
- if (cgroup_ops->can_use_cpuview(cgroup_ops) && cg_cpu_usage) {
- total_len = cpuview_proc_stat(cg, cpuset, cg_cpu_usage, cg_cpu_usage_size,
- f, d->buf, d->buflen);
- goto out;
- }
-
- while (getline(&line, &linelen, f) != -1) {
- ssize_t l;
- char cpu_char[10]; /* That's a lot of cores */
- char *c;
- uint64_t all_used, cg_used, new_idle;
- int ret;
-
- if (strlen(line) == 0)
- continue;
- if (sscanf(line, "cpu%9[^ ]", cpu_char) != 1) {
- /* not a ^cpuN line containing a number N, just print it */
- l = snprintf(cache, cache_size, "%s", line);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
- cache += l;
- cache_size -= l;
- total_len += l;
- continue;
- }
-
- if (sscanf(cpu_char, "%d", &physcpu) != 1)
- continue;
- if (!cpu_in_cpuset(physcpu, cpuset))
- continue;
- curcpu ++;
-
- ret = sscanf(line, "%*s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
- &user,
- &nice,
- &system,
- &idle,
- &iowait,
- &irq,
- &softirq,
- &steal,
- &guest,
- &guest_nice);
-
- if (ret != 10 || !cg_cpu_usage) {
- c = strchr(line, ' ');
- if (!c)
- continue;
- l = snprintf(cache, cache_size, "cpu%d%s", curcpu, c);
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
-
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
-
- cache += l;
- cache_size -= l;
- total_len += l;
-
- if (ret != 10)
- continue;
- }
-
- if (cg_cpu_usage) {
- if (physcpu >= cg_cpu_usage_size)
- break;
-
- all_used = user + nice + system + iowait + irq + softirq + steal + guest + guest_nice;
- cg_used = cg_cpu_usage[physcpu].user + cg_cpu_usage[physcpu].system;
-
- if (all_used >= cg_used) {
- new_idle = idle + (all_used - cg_used);
-
- } else {
- lxcfs_error("cpu%d from %s has unexpected cpu time: %lu in /proc/stat, "
- "%lu in cpuacct.usage_all; unable to determine idle time\n",
- curcpu, cg, all_used, cg_used);
- new_idle = idle;
- }
-
- l = snprintf(cache, cache_size, "cpu%d %lu 0 %lu %lu 0 0 0 0 0 0\n",
- curcpu, cg_cpu_usage[physcpu].user, cg_cpu_usage[physcpu].system,
- new_idle);
-
- if (l < 0) {
- perror("Error writing to cache");
- return 0;
-
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
-
- cache += l;
- cache_size -= l;
- total_len += l;
-
- user_sum += cg_cpu_usage[physcpu].user;
- system_sum += cg_cpu_usage[physcpu].system;
- idle_sum += new_idle;
-
- } else {
- user_sum += user;
- nice_sum += nice;
- system_sum += system;
- idle_sum += idle;
- iowait_sum += iowait;
- irq_sum += irq;
- softirq_sum += softirq;
- steal_sum += steal;
- guest_sum += guest;
- guest_nice_sum += guest_nice;
- }
- }
-
- cache = d->buf;
-
- int cpuall_len = snprintf(cpuall, CPUALL_MAX_SIZE, "cpu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu\n",
- user_sum,
- nice_sum,
- system_sum,
- idle_sum,
- iowait_sum,
- irq_sum,
- softirq_sum,
- steal_sum,
- guest_sum,
- guest_nice_sum);
- if (cpuall_len > 0 && cpuall_len < CPUALL_MAX_SIZE) {
- memcpy(cache, cpuall, cpuall_len);
- cache += cpuall_len;
- } else {
- /* shouldn't happen */
- lxcfs_error("proc_stat_read copy cpuall failed, cpuall_len=%d.", cpuall_len);
- cpuall_len = 0;
- }
-
- memmove(cache, d->buf + CPUALL_MAX_SIZE, total_len);
- total_len += cpuall_len;
-
-out:
- d->cached = 1;
- d->size = total_len;
- if (total_len > size)
- total_len = size;
-
- memcpy(buf, d->buf, total_len);
- return total_len;
-}
-
-/* This function retrieves the busy time of a group of tasks by looking at
- * cpuacct.usage. Unfortunately, this only makes sense when the container has
- * been given it's own cpuacct cgroup. If not, this function will take the busy
- * time of all other taks that do not actually belong to the container into
- * account as well. If someone has a clever solution for this please send a
- * patch!
- */
-static double get_reaper_busy(pid_t task)
-{
- __do_free char *cgroup = NULL, *usage_str = NULL;
- unsigned long usage = 0;
- pid_t initpid;
-
- initpid = lookup_initpid_in_store(task);
- if (initpid <= 0)
- return 0;
-
- cgroup = get_pid_cgroup(initpid, "cpuacct");
- if (!cgroup)
- return 0;
- prune_init_slice(cgroup);
- if (!cgroup_ops->get(cgroup_ops, "cpuacct", cgroup, "cpuacct.usage",
- &usage_str))
- return 0;
-
- usage = strtoul(usage_str, NULL, 10);
- return ((double)usage / 1000000000);
-}
-
-#if RELOADTEST
-void iwashere(void)
-{
- int fd;
-
- fd = creat("/tmp/lxcfs-iwashere", 0644);
- if (fd >= 0)
- close(fd);
-}
-#endif
-
-/*
- * We read /proc/uptime and reuse its second field.
- * For the first field, we use the mtime for the reaper for
- * the calling pid as returned by getreaperage
- */
-static int proc_uptime_read(char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- struct fuse_context *fc = fuse_get_context();
- struct file_info *d = (struct file_info *)fi->fh;
- double busytime = get_reaper_busy(fc->pid);
- char *cache = d->buf;
- ssize_t total_len = 0;
- double idletime, reaperage;
-
-#if RELOADTEST
- iwashere();
-#endif
-
- if (offset){
- if (!d->cached)
- return 0;
- if (offset > d->size)
- return -EINVAL;
- int left = d->size - offset;
- total_len = left > size ? size: left;
- memcpy(buf, cache + offset, total_len);
- return total_len;
- }
-
- reaperage = get_reaper_age(fc->pid);
- /* To understand why this is done, please read the comment to the
- * get_reaper_busy() function.
- */
- idletime = reaperage;
- if (reaperage >= busytime)
- idletime = reaperage - busytime;
-
- total_len = snprintf(d->buf, d->buflen, "%.2lf %.2lf\n", reaperage, idletime);
- if (total_len < 0 || total_len >= d->buflen){
- lxcfs_error("%s\n", "failed to write to cache");
- return 0;
- }
-
- d->size = (int)total_len;
- d->cached = 1;
-
- if (total_len > size) total_len = size;
-
- memcpy(buf, d->buf, total_len);
- return total_len;
-}
-
-static int proc_diskstats_read(char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- __do_free char *cg = NULL, *io_serviced_str = NULL,
- *io_merged_str = NULL, *io_service_bytes_str = NULL,
- *io_wait_time_str = NULL, *io_service_time_str = NULL,
- *line = NULL;
- __do_fclose FILE *f = NULL;
- struct fuse_context *fc = fuse_get_context();
- struct file_info *d = (struct file_info *)fi->fh;
- unsigned long read = 0, write = 0;
- unsigned long read_merged = 0, write_merged = 0;
- unsigned long read_sectors = 0, write_sectors = 0;
- unsigned long read_ticks = 0, write_ticks = 0;
- unsigned long ios_pgr = 0, tot_ticks = 0, rq_ticks = 0;
- unsigned long rd_svctm = 0, wr_svctm = 0, rd_wait = 0, wr_wait = 0;
- char *cache = d->buf;
- size_t cache_size = d->buflen;
- size_t linelen = 0, total_len = 0;
- unsigned int major = 0, minor = 0;
- int i = 0;
- int ret;
- char dev_name[72];
-
- if (offset){
- int left;
-
- if (offset > d->size)
- return -EINVAL;
-
- if (!d->cached)
- return 0;
-
- left = d->size - offset;
- total_len = left > size ? size: left;
- memcpy(buf, cache + offset, total_len);
-
- return total_len;
- }
-
- pid_t initpid = lookup_initpid_in_store(fc->pid);
- if (initpid <= 1 || is_shared_pidns(initpid))
- initpid = fc->pid;
- cg = get_pid_cgroup(initpid, "blkio");
- if (!cg)
- return read_file_fuse("/proc/diskstats", buf, size, d);
- prune_init_slice(cg);
-
- ret = cgroup_ops->get_io_serviced(cgroup_ops, cg, &io_serviced_str);
- if (ret < 0) {
- if (ret == -EOPNOTSUPP)
- return read_file_fuse("/proc/diskstats", buf, size, d);
- }
-
- ret = cgroup_ops->get_io_merged(cgroup_ops, cg, &io_merged_str);
- if (ret < 0) {
- if (ret == -EOPNOTSUPP)
- return read_file_fuse("/proc/diskstats", buf, size, d);
- }
-
- ret = cgroup_ops->get_io_service_bytes(cgroup_ops, cg, &io_service_bytes_str);
- if (ret < 0) {
- if (ret == -EOPNOTSUPP)
- return read_file_fuse("/proc/diskstats", buf, size, d);
- }
-
- ret = cgroup_ops->get_io_wait_time(cgroup_ops, cg, &io_wait_time_str);
- if (ret < 0) {
- if (ret == -EOPNOTSUPP)
- return read_file_fuse("/proc/diskstats", buf, size, d);
- }
-
- ret = cgroup_ops->get_io_service_time(cgroup_ops, cg, &io_service_time_str);
- if (ret < 0) {
- if (ret == -EOPNOTSUPP)
- return read_file_fuse("/proc/diskstats", buf, size, d);
- }
-
- f = fopen("/proc/diskstats", "r");
- if (!f)
- return 0;
-
- while (getline(&line, &linelen, f) != -1) {
- ssize_t l;
- char lbuf[256];
-
- i = sscanf(line, "%u %u %71s", &major, &minor, dev_name);
- if (i != 3)
- continue;
-
- get_blkio_io_value(io_serviced_str, major, minor, "Read", &read);
- get_blkio_io_value(io_serviced_str, major, minor, "Write", &write);
- get_blkio_io_value(io_merged_str, major, minor, "Read", &read_merged);
- get_blkio_io_value(io_merged_str, major, minor, "Write", &write_merged);
- get_blkio_io_value(io_service_bytes_str, major, minor, "Read", &read_sectors);
- read_sectors = read_sectors/512;
- get_blkio_io_value(io_service_bytes_str, major, minor, "Write", &write_sectors);
- write_sectors = write_sectors/512;
-
- get_blkio_io_value(io_service_time_str, major, minor, "Read", &rd_svctm);
- rd_svctm = rd_svctm/1000000;
- get_blkio_io_value(io_wait_time_str, major, minor, "Read", &rd_wait);
- rd_wait = rd_wait/1000000;
- read_ticks = rd_svctm + rd_wait;
-
- get_blkio_io_value(io_service_time_str, major, minor, "Write", &wr_svctm);
- wr_svctm = wr_svctm/1000000;
- get_blkio_io_value(io_wait_time_str, major, minor, "Write", &wr_wait);
- wr_wait = wr_wait/1000000;
- write_ticks = wr_svctm + wr_wait;
-
- get_blkio_io_value(io_service_time_str, major, minor, "Total", &tot_ticks);
- tot_ticks = tot_ticks/1000000;
-
- memset(lbuf, 0, 256);
- if (read || write || read_merged || write_merged || read_sectors || write_sectors || read_ticks || write_ticks)
- snprintf(lbuf, 256, "%u %u %s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu\n",
- major, minor, dev_name, read, read_merged, read_sectors, read_ticks,
- write, write_merged, write_sectors, write_ticks, ios_pgr, tot_ticks, rq_ticks);
- else
- continue;
-
- l = snprintf(cache, cache_size, "%s", lbuf);
- if (l < 0) {
- perror("Error writing to fuse buf");
- return 0;
- }
- if (l >= cache_size) {
- lxcfs_error("%s\n", "Internal error: truncated write to cache.");
- return 0;
- }
- cache += l;
- cache_size -= l;
- total_len += l;
- }
-
- d->cached = 1;
- d->size = total_len;
- if (total_len > size ) total_len = size;
- memcpy(buf, d->buf, total_len);
-
- return total_len;
-}
-
-static int proc_swaps_read(char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- __do_free char *cg = NULL, *memswlimit_str = NULL, *memusage_str = NULL,
- *memswusage_str = NULL;
- struct fuse_context *fc = fuse_get_context();
- struct file_info *d = (struct file_info *)fi->fh;
- unsigned long memswlimit = 0, memlimit = 0, memusage = 0,
- memswusage = 0, swap_total = 0, swap_free = 0;
- ssize_t total_len = 0;
- ssize_t l = 0;
- char *cache = d->buf;
- int ret;
-
- if (offset) {
- int left;
-
- if (offset > d->size)
- return -EINVAL;
-
- if (!d->cached)
- return 0;
-
- left = d->size - offset;
- total_len = left > size ? size: left;
- memcpy(buf, cache + offset, total_len);
-
- return total_len;
- }
-
- pid_t initpid = lookup_initpid_in_store(fc->pid);
- if (initpid <= 1 || is_shared_pidns(initpid))
- initpid = fc->pid;
- cg = get_pid_cgroup(initpid, "memory");
- if (!cg)
- return read_file_fuse("/proc/swaps", buf, size, d);
- prune_init_slice(cg);
-
- memlimit = get_min_memlimit(cg, false);
-
- ret = cgroup_ops->get_memory_current(cgroup_ops, cg, &memusage_str);
- if (ret < 0)
- return 0;
-
- memusage = strtoul(memusage_str, NULL, 10);
-
- ret = cgroup_ops->get_memory_swap_max(cgroup_ops, cg, &memswlimit_str);
- if (ret >= 0)
- ret = cgroup_ops->get_memory_swap_current(cgroup_ops, cg, &memswusage_str);
- if (ret >= 0) {
- memswlimit = get_min_memlimit(cg, true);
- memswusage = strtoul(memswusage_str, NULL, 10);
- swap_total = (memswlimit - memlimit) / 1024;
- swap_free = (memswusage - memusage) / 1024;
- }
-
- total_len = snprintf(d->buf, d->size, "Filename\t\t\t\tType\t\tSize\tUsed\tPriority\n");
-
- /* When no mem + swap limit is specified or swapaccount=0*/
- if (!memswlimit) {
- __do_free char *line = NULL;
- __do_fclose FILE *f = NULL;
- size_t linelen = 0;
-
- f = fopen("/proc/meminfo", "r");
- if (!f)
- return 0;
-
- while (getline(&line, &linelen, f) != -1) {
- if (startswith(line, "SwapTotal:"))
- sscanf(line, "SwapTotal: %8lu kB", &swap_total);
- else if (startswith(line, "SwapFree:"))
- sscanf(line, "SwapFree: %8lu kB", &swap_free);
- }
- }
-
- if (swap_total > 0) {
- l = snprintf(d->buf + total_len, d->size - total_len,
- "none%*svirtual\t\t%lu\t%lu\t0\n", 36, " ",
- swap_total, swap_free);
- total_len += l;
- }
-
- if (total_len < 0 || l < 0) {
- perror("Error writing to cache");
- return 0;
- }
-
- d->cached = 1;
- d->size = (int)total_len;
-
- if (total_len > size) total_len = size;
- memcpy(buf, d->buf, total_len);
- return total_len;
-}
-
-/*
- * Find the process pid from cgroup path.
- * eg:from /sys/fs/cgroup/cpu/docker/containerid/cgroup.procs to find the process pid.
- * @pid_buf : put pid to pid_buf.
- * @dpath : the path of cgroup. eg: /docker/containerid or /docker/containerid/child-cgroup ...
- * @depth : the depth of cgroup in container.
- * @sum : return the number of pid.
- * @cfd : the file descriptor of the mounted cgroup. eg: /sys/fs/cgroup/cpu
- */
-static int calc_pid(char ***pid_buf, char *dpath, int depth, int sum, int cfd)
-{
- __do_free char *path = NULL;
- __do_close_prot_errno int fd = -EBADF;
- __do_fclose FILE *f = NULL;
- __do_closedir DIR *dir = NULL;
- struct dirent *file;
- size_t linelen = 0;
- char *line = NULL;
- int pd;
- char **pid;
-
- /* path = dpath + "/cgroup.procs" + /0 */
- path = malloc(strlen(dpath) + 20);
- if (!path)
- return sum;
-
- strcpy(path, dpath);
- fd = openat(cfd, path, O_RDONLY | O_CLOEXEC | O_NOFOLLOW);
- if (fd < 0)
- return sum;
-
- dir = fdopendir(move_fd(fd));
- if (!dir)
- return sum;
-
- while (((file = readdir(dir)) != NULL) && depth > 0) {
- if (strcmp(file->d_name, ".") == 0)
- continue;
-
- if (strcmp(file->d_name, "..") == 0)
- continue;
-
- if (file->d_type == DT_DIR) {
- __do_free char *path_dir = NULL;
-
- /* path + '/' + d_name +/0 */
- path_dir = malloc(strlen(path) + 2 + sizeof(file->d_name));
- if (!path_dir)
- return sum;
-
- strcpy(path_dir, path);
- strcat(path_dir, "/");
- strcat(path_dir, file->d_name);
- pd = depth - 1;
- sum = calc_pid(pid_buf, path_dir, pd, sum, cfd);
- }
- }
-
- strcat(path, "/cgroup.procs");
- fd = openat(cfd, path, O_RDONLY);
- if (fd < 0)
- return sum;
-
- f = fdopen(move_fd(fd), "r");
- if (!f)
- return sum;
-
- while (getline(&line, &linelen, f) != -1) {
- pid = realloc(*pid_buf, sizeof(char *) * (sum + 1));
- if (!pid)
- return sum;
- *pid_buf = pid;
-
- *(*pid_buf + sum) = malloc(strlen(line) + 1);
- if (!*(*pid_buf + sum))
- return sum;
-
- strcpy(*(*pid_buf + sum), line);
- sum++;
- }
-
- return sum;
-}
-
-/*
- * calc_load calculates the load according to the following formula:
- * load1 = load0 * exp + active * (1 - exp)
- *
- * @load1: the new loadavg.
- * @load0: the former loadavg.
- * @active: the total number of running pid at this moment.
- * @exp: the fixed-point defined in the beginning.
- */
-static unsigned long
-calc_load(unsigned long load, unsigned long exp, unsigned long active)
-{
- unsigned long newload;
-
- active = active > 0 ? active * FIXED_1 : 0;
- newload = load * exp + active * (FIXED_1 - exp);
- if (active >= load)
- newload += FIXED_1 - 1;
-
- return newload / FIXED_1;
-}
-
-/*
- * Return 0 means that container p->cg is closed.
- * Return -1 means that error occurred in refresh.
- * Positive num equals the total number of pid.
- */
-static int refresh_load(struct load_node *p, char *path)
-{
- __do_free char *line = NULL;
- char **idbuf;
- char proc_path[256];
- int i, ret, run_pid = 0, total_pid = 0, last_pid = 0;
- size_t linelen = 0;
- int sum, length;
- struct dirent *file;
-
- idbuf = malloc(sizeof(char *));
- if (!idbuf)
- return -1;
-
- sum = calc_pid(&idbuf, path, DEPTH_DIR, 0, p->cfd);
- /* normal exit */
- if (sum == 0)
- goto out;
-
- for (i = 0; i < sum; i++) {
- __do_closedir DIR *dp = NULL;
-
- /*clean up '\n' */
- length = strlen(idbuf[i])-1;
- idbuf[i][length] = '\0';
- ret = snprintf(proc_path, 256, "/proc/%s/task", idbuf[i]);
- if (ret < 0 || ret > 255) {
- lxcfs_error("%s\n", "snprintf() failed in refresh_load.");
- i = sum;
- sum = -1;
- goto err_out;
- }
-
- dp = opendir(proc_path);
- if (!dp) {
- lxcfs_error("%s\n", "Open proc_path failed in refresh_load.");
- continue;
- }
- while ((file = readdir(dp)) != NULL) {
- __do_fclose FILE *f = NULL;
-
- if (strncmp(file->d_name, ".", 1) == 0)
- continue;
- if (strncmp(file->d_name, "..", 1) == 0)
- continue;
- total_pid++;
- /* We make the biggest pid become last_pid.*/
- ret = atof(file->d_name);
- last_pid = (ret > last_pid) ? ret : last_pid;
-
- ret = snprintf(proc_path, 256, "/proc/%s/task/%s/status", idbuf[i], file->d_name);
- if (ret < 0 || ret > 255) {
- lxcfs_error("%s\n", "snprintf() failed in refresh_load.");
- i = sum;
- sum = -1;
- goto err_out;
- }
-
- f = fopen(proc_path, "r");
- if (f != NULL) {
- while (getline(&line, &linelen, f) != -1) {
- /* Find State */
- if ((line[0] == 'S') && (line[1] == 't'))
- break;
- }
-
- if ((line[7] == 'R') || (line[7] == 'D'))
- run_pid++;
- }
- }
- }
- /*Calculate the loadavg.*/
- p->avenrun[0] = calc_load(p->avenrun[0], EXP_1, run_pid);
- p->avenrun[1] = calc_load(p->avenrun[1], EXP_5, run_pid);
- p->avenrun[2] = calc_load(p->avenrun[2], EXP_15, run_pid);
- p->run_pid = run_pid;
- p->total_pid = total_pid;
- p->last_pid = last_pid;
-
-err_out:
- for (; i > 0; i--)
- free(idbuf[i-1]);
-out:
- free(idbuf);
- return sum;
-}
-
-/*
- * Traverse the hash table and update it.
- */
-void *load_begin(void *arg)
-{
-
- int i, sum, length, ret;
- struct load_node *f;
- int first_node;
- clock_t time1, time2;
-
- while (1) {
- if (loadavg_stop == 1)
- return NULL;
-
- time1 = clock();
- for (i = 0; i < LOAD_SIZE; i++) {
- pthread_mutex_lock(&load_hash[i].lock);
- if (load_hash[i].next == NULL) {
- pthread_mutex_unlock(&load_hash[i].lock);
- continue;
- }
- f = load_hash[i].next;
- first_node = 1;
- while (f) {
- __do_free char *path = NULL;
-
- length = strlen(f->cg) + 2;
- /* strlen(f->cg) + '.' or '' + \0 */
- path = malloc(length);
- if (!path)
- goto out;
-
- ret = snprintf(path, length, "%s%s", dot_or_empty(f->cg), f->cg);
- if (ret < 0 || ret > length - 1) {
- /* snprintf failed, ignore the node.*/
- lxcfs_error("Refresh node %s failed for snprintf().\n", f->cg);
- goto out;
- }
-
- sum = refresh_load(f, path);
- if (sum == 0)
- f = del_node(f, i);
- else
-out: f = f->next;
- /* load_hash[i].lock locks only on the first node.*/
- if (first_node == 1) {
- first_node = 0;
- pthread_mutex_unlock(&load_hash[i].lock);
- }
- }
- }
-
- if (loadavg_stop == 1)
- return NULL;
-
- time2 = clock();
- usleep(FLUSH_TIME * 1000000 - (int)((time2 - time1) * 1000000 / CLOCKS_PER_SEC));
- }
-}
-
-static int proc_loadavg_read(char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- struct fuse_context *fc = fuse_get_context();
- struct file_info *d = (struct file_info *)fi->fh;
- pid_t initpid;
- char *cg;
- size_t total_len = 0;
- char *cache = d->buf;
- struct load_node *n;
- int hash;
- int cfd, rv = 0;
- unsigned long a, b, c;
-
- if (offset) {
- if (offset > d->size)
- return -EINVAL;
- if (!d->cached)
- return 0;
- int left = d->size - offset;
- total_len = left > size ? size : left;
- memcpy(buf, cache + offset, total_len);
- return total_len;
- }
- if (!loadavg)
- return read_file_fuse("/proc/loadavg", buf, size, d);
-
- initpid = lookup_initpid_in_store(fc->pid);
- if (initpid <= 1 || is_shared_pidns(initpid))
- initpid = fc->pid;
- cg = get_pid_cgroup(initpid, "cpu");
- if (!cg)
- return read_file_fuse("/proc/loadavg", buf, size, d);
-
- prune_init_slice(cg);
- hash = calc_hash(cg) % LOAD_SIZE;
- n = locate_node(cg, hash);
-
- /* First time */
- if (n == NULL) {
- cfd = get_cgroup_fd("cpu");
- if (cfd >= 0) {
- /*
- * In locate_node() above, pthread_rwlock_unlock() isn't used
- * because delete is not allowed before read has ended.
- */
- pthread_rwlock_unlock(&load_hash[hash].rdlock);
- rv = 0;
- goto err;
- }
- do {
- n = malloc(sizeof(struct load_node));
- } while (!n);
-
- do {
- n->cg = malloc(strlen(cg)+1);
- } while (!n->cg);
- strcpy(n->cg, cg);
- n->avenrun[0] = 0;
- n->avenrun[1] = 0;
- n->avenrun[2] = 0;
- n->run_pid = 0;
- n->total_pid = 1;
- n->last_pid = initpid;
- n->cfd = cfd;
- insert_node(&n, hash);
- }
- a = n->avenrun[0] + (FIXED_1/200);
- b = n->avenrun[1] + (FIXED_1/200);
- c = n->avenrun[2] + (FIXED_1/200);
- total_len = snprintf(d->buf, d->buflen, "%lu.%02lu %lu.%02lu %lu.%02lu %d/%d %d\n",
- LOAD_INT(a), LOAD_FRAC(a),
- LOAD_INT(b), LOAD_FRAC(b),
- LOAD_INT(c), LOAD_FRAC(c),
- n->run_pid, n->total_pid, n->last_pid);
- pthread_rwlock_unlock(&load_hash[hash].rdlock);
- if (total_len < 0 || total_len >= d->buflen) {
- lxcfs_error("%s\n", "Failed to write to cache");
- rv = 0;
- goto err;
- }
- d->size = (int)total_len;
- d->cached = 1;
-
- if (total_len > size)
- total_len = size;
- memcpy(buf, d->buf, total_len);
- rv = total_len;
-
-err:
- free(cg);
- return rv;
-}
-/* Return a positive number on success, return 0 on failure.*/
-pthread_t load_daemon(int load_use)
-{
- int ret;
- pthread_t pid;
-
- ret = init_load();
- if (ret == -1) {
- lxcfs_error("%s\n", "Initialize hash_table fails in load_daemon!");
- return 0;
- }
- ret = pthread_create(&pid, NULL, load_begin, NULL);
- if (ret != 0) {
- lxcfs_error("%s\n", "Create pthread fails in load_daemon!");
- load_free();
- return 0;
- }
- /* use loadavg, here loadavg = 1*/
- loadavg = load_use;
- return pid;
-}
-
-/* Returns 0 on success. */
-int stop_load_daemon(pthread_t pid)
-{
- int s;
-
- /* Signal the thread to gracefully stop */
- loadavg_stop = 1;
-
- s = pthread_join(pid, NULL); /* Make sure sub thread has been canceled. */
- if (s != 0) {
- lxcfs_error("%s\n", "stop_load_daemon error: failed to join");
- return -1;
- }
-
- load_free();
- loadavg_stop = 0;
-
- return 0;
-}
-
-static off_t get_procfile_size(const char *which)
-{
- FILE *f = fopen(which, "r");
- char *line = NULL;
- size_t len = 0;
- ssize_t sz, answer = 0;
- if (!f)
- return 0;
-
- while ((sz = getline(&line, &len, f)) != -1)
- answer += sz;
- fclose (f);
- free(line);
-
- return answer;
-}
-
-int proc_getattr(const char *path, struct stat *sb)
-{
- struct timespec now;
-
- memset(sb, 0, sizeof(struct stat));
- if (clock_gettime(CLOCK_REALTIME, &now) < 0)
- return -EINVAL;
- sb->st_uid = sb->st_gid = 0;
- sb->st_atim = sb->st_mtim = sb->st_ctim = now;
- if (strcmp(path, "/proc") == 0) {
- sb->st_mode = S_IFDIR | 00555;
- sb->st_nlink = 2;
- return 0;
- }
- if (strcmp(path, "/proc/meminfo") == 0 ||
- strcmp(path, "/proc/cpuinfo") == 0 ||
- strcmp(path, "/proc/uptime") == 0 ||
- strcmp(path, "/proc/stat") == 0 ||
- strcmp(path, "/proc/diskstats") == 0 ||
- strcmp(path, "/proc/swaps") == 0 ||
- strcmp(path, "/proc/loadavg") == 0) {
- sb->st_size = 0;
- sb->st_mode = S_IFREG | 00444;
- sb->st_nlink = 1;
- return 0;
- }
-
- return -ENOENT;
-}
-
-int proc_readdir(const char *path, void *buf, fuse_fill_dir_t filler, off_t offset,
- struct fuse_file_info *fi)
-{
- if (filler(buf, ".", NULL, 0) != 0 ||
- filler(buf, "..", NULL, 0) != 0 ||
- filler(buf, "cpuinfo", NULL, 0) != 0 ||
- filler(buf, "meminfo", NULL, 0) != 0 ||
- filler(buf, "stat", NULL, 0) != 0 ||
- filler(buf, "uptime", NULL, 0) != 0 ||
- filler(buf, "diskstats", NULL, 0) != 0 ||
- filler(buf, "swaps", NULL, 0) != 0 ||
- filler(buf, "loadavg", NULL, 0) != 0)
- return -EINVAL;
- return 0;
-}
-
-int proc_open(const char *path, struct fuse_file_info *fi)
-{
- int type = -1;
- struct file_info *info;
-
- if (strcmp(path, "/proc/meminfo") == 0)
- type = LXC_TYPE_PROC_MEMINFO;
- else if (strcmp(path, "/proc/cpuinfo") == 0)
- type = LXC_TYPE_PROC_CPUINFO;
- else if (strcmp(path, "/proc/uptime") == 0)
- type = LXC_TYPE_PROC_UPTIME;
- else if (strcmp(path, "/proc/stat") == 0)
- type = LXC_TYPE_PROC_STAT;
- else if (strcmp(path, "/proc/diskstats") == 0)
- type = LXC_TYPE_PROC_DISKSTATS;
- else if (strcmp(path, "/proc/swaps") == 0)
- type = LXC_TYPE_PROC_SWAPS;
- else if (strcmp(path, "/proc/loadavg") == 0)
- type = LXC_TYPE_PROC_LOADAVG;
- if (type == -1)
- return -ENOENT;
-
- info = malloc(sizeof(*info));
- if (!info)
- return -ENOMEM;
-
- memset(info, 0, sizeof(*info));
- info->type = type;
-
- info->buflen = get_procfile_size(path) + BUF_RESERVE_SIZE;
- do {
- info->buf = malloc(info->buflen);
- } while (!info->buf);
- memset(info->buf, 0, info->buflen);
- /* set actual size to buffer size */
- info->size = info->buflen;
-
- fi->fh = (unsigned long)info;
- return 0;
-}
-
-int proc_access(const char *path, int mask)
-{
- if (strcmp(path, "/proc") == 0 && access(path, R_OK) == 0)
- return 0;
-
- /* these are all read-only */
- if ((mask & ~R_OK) != 0)
- return -EACCES;
- return 0;
-}
-
-int proc_release(const char *path, struct fuse_file_info *fi)
-{
- do_release_file_info(fi);
- return 0;
-}
-
-int proc_read(const char *path, char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi)
-{
- struct file_info *f = (struct file_info *) fi->fh;
-
- switch (f->type) {
- case LXC_TYPE_PROC_MEMINFO:
- return proc_meminfo_read(buf, size, offset, fi);
- case LXC_TYPE_PROC_CPUINFO:
- return proc_cpuinfo_read(buf, size, offset, fi);
- case LXC_TYPE_PROC_UPTIME:
- return proc_uptime_read(buf, size, offset, fi);
- case LXC_TYPE_PROC_STAT:
- return proc_stat_read(buf, size, offset, fi);
- case LXC_TYPE_PROC_DISKSTATS:
- return proc_diskstats_read(buf, size, offset, fi);
- case LXC_TYPE_PROC_SWAPS:
- return proc_swaps_read(buf, size, offset, fi);
- case LXC_TYPE_PROC_LOADAVG:
- return proc_loadavg_read(buf, size, offset, fi);
- default:
- return -EINVAL;
- }
-}
-
/*
* Functions needed to setup cgroups in the __constructor__.
*/
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
#ifndef __LXCFS_BINDINGS_H
#define __LXCFS_BINDINGS_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <fuse.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "config.h"
#include "macro.h"
#include "cgroup_fuse.h"
+#include "proc_cpuview.h"
+#include "proc_fuse.h"
+#include "proc_loadavg.h"
#include "sysfs_fuse.h"
/* directory under which we mount the controllers - /run/lxcfs/controllers */
bool swap_off;
};
-extern int proc_getattr(const char *path, struct stat *sb);
-extern int proc_readdir(const char *path, void *buf, fuse_fill_dir_t filler, off_t offset,
- struct fuse_file_info *fi);
-extern int proc_release(const char *path, struct fuse_file_info *fi);
-extern int proc_open(const char *path, struct fuse_file_info *fi);
-extern int proc_read(const char *path, char *buf, size_t size, off_t offset,
- struct fuse_file_info *fi);
-extern int proc_access(const char *path, int mask);
-extern pthread_t load_daemon(int load_use);
-extern int stop_load_daemon(pthread_t pid);
-
extern pid_t lookup_initpid_in_store(pid_t qpid);
-extern char *get_pid_cgroup(pid_t pid, const char *contrl);
-extern int read_file_fuse(const char *path, char *buf, size_t size,
- struct file_info *d);
extern void prune_init_slice(char *cg);
-extern char *get_cpuset(const char *cg);
-extern int max_cpu_count(const char *cg);
-extern void do_release_file_info(struct fuse_file_info *fi);
-extern int cpu_number_in_cpuset(const char *cpuset);
#endif /* __LXCFS_BINDINGS_H */
/* SPDX-License-Identifier: LGPL-2.1-or-later */
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
#define __STDC_FORMAT_MACROS
#include <dirent.h>
#define POLLIN_SET ( EPOLLIN | EPOLLHUP | EPOLLRDHUP )
-static bool wait_for_sock(int sock, int timeout)
-{
- struct epoll_event ev;
- int epfd, ret, now, starttime, deltatime, saved_errno;
-
- if ((starttime = time(NULL)) < 0)
- return false;
-
- if ((epfd = epoll_create(1)) < 0) {
- lxcfs_error("%s\n", "Failed to create epoll socket: %m.");
- return false;
- }
-
- ev.events = POLLIN_SET;
- ev.data.fd = sock;
- if (epoll_ctl(epfd, EPOLL_CTL_ADD, sock, &ev) < 0) {
- lxcfs_error("%s\n", "Failed adding socket to epoll: %m.");
- close(epfd);
- return false;
- }
-
-again:
- if ((now = time(NULL)) < 0) {
- close(epfd);
- return false;
- }
-
- deltatime = (starttime + timeout) - now;
- if (deltatime < 0) { // timeout
- errno = 0;
- close(epfd);
- return false;
- }
- ret = epoll_wait(epfd, &ev, 1, 1000*deltatime + 1);
- if (ret < 0 && errno == EINTR)
- goto again;
- saved_errno = errno;
- close(epfd);
-
- if (ret <= 0) {
- errno = saved_errno;
- return false;
- }
- return true;
-}
-
-static int msgrecv(int sockfd, void *buf, size_t len)
-{
- if (!wait_for_sock(sockfd, 2))
- return -1;
- return recv(sockfd, buf, len, MSG_DONTWAIT);
-}
-
-#define SEND_CREDS_OK 0
-#define SEND_CREDS_NOTSK 1
-#define SEND_CREDS_FAIL 2
-
-static int send_creds(int sock, struct ucred *cred, char v, bool pingfirst)
-{
- struct msghdr msg = { 0 };
- struct iovec iov;
- struct cmsghdr *cmsg;
- char cmsgbuf[CMSG_SPACE(sizeof(*cred))];
- char buf[1];
- buf[0] = 'p';
-
- if (pingfirst) {
- if (msgrecv(sock, buf, 1) != 1) {
- lxcfs_error("%s\n", "Error getting reply from server over socketpair.");
- return SEND_CREDS_FAIL;
- }
- }
-
- msg.msg_control = cmsgbuf;
- msg.msg_controllen = sizeof(cmsgbuf);
-
- cmsg = CMSG_FIRSTHDR(&msg);
- cmsg->cmsg_len = CMSG_LEN(sizeof(struct ucred));
- cmsg->cmsg_level = SOL_SOCKET;
- cmsg->cmsg_type = SCM_CREDENTIALS;
- memcpy(CMSG_DATA(cmsg), cred, sizeof(*cred));
-
- msg.msg_name = NULL;
- msg.msg_namelen = 0;
-
- buf[0] = v;
- iov.iov_base = buf;
- iov.iov_len = sizeof(buf);
- msg.msg_iov = &iov;
- msg.msg_iovlen = 1;
-
- if (sendmsg(sock, &msg, 0) < 0) {
- lxcfs_error("Failed at sendmsg: %s.\n",strerror(errno));
- if (errno == 3)
- return SEND_CREDS_NOTSK;
- return SEND_CREDS_FAIL;
- }
-
- return SEND_CREDS_OK;
-}
-
static int wait_for_pid(pid_t pid)
{
int status, ret;
return 0;
}
-static bool recv_creds(int sock, struct ucred *cred, char *v)
-{
- struct msghdr msg = { 0 };
- struct iovec iov;
- struct cmsghdr *cmsg;
- char cmsgbuf[CMSG_SPACE(sizeof(*cred))];
- char buf[1];
- int ret;
- int optval = 1;
-
- *v = '1';
-
- cred->pid = -1;
- cred->uid = -1;
- cred->gid = -1;
-
- if (setsockopt(sock, SOL_SOCKET, SO_PASSCRED, &optval, sizeof(optval)) == -1) {
- lxcfs_error("Failed to set passcred: %s\n", strerror(errno));
- return false;
- }
- buf[0] = '1';
- if (write(sock, buf, 1) != 1) {
- lxcfs_error("Failed to start write on scm fd: %s\n", strerror(errno));
- return false;
- }
-
- msg.msg_name = NULL;
- msg.msg_namelen = 0;
- msg.msg_control = cmsgbuf;
- msg.msg_controllen = sizeof(cmsgbuf);
-
- iov.iov_base = buf;
- iov.iov_len = sizeof(buf);
- msg.msg_iov = &iov;
- msg.msg_iovlen = 1;
-
- if (!wait_for_sock(sock, 2)) {
- lxcfs_error("Timed out waiting for scm_cred: %s\n", strerror(errno));
- return false;
- }
- ret = recvmsg(sock, &msg, MSG_DONTWAIT);
- if (ret < 0) {
- lxcfs_error("Failed to receive scm_cred: %s\n", strerror(errno));
- return false;
- }
-
- cmsg = CMSG_FIRSTHDR(&msg);
-
- if (cmsg && cmsg->cmsg_len == CMSG_LEN(sizeof(struct ucred)) &&
- cmsg->cmsg_level == SOL_SOCKET &&
- cmsg->cmsg_type == SCM_CREDENTIALS) {
- memcpy(cred, CMSG_DATA(cmsg), sizeof(*cred));
- }
- *v = buf[0];
-
- return true;
-}
-
/*
* pid_to_ns - reads pids from a ucred over a socket, then writes the
* int value back over the socket. This shifts the pid from the
#ifndef __LXCFS_CGROUP_FUSE_H
#define __LXCFS_CGROUP_FUSE_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <fuse.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "config.h"
+#include "macro.h"
+
extern int cg_getattr(const char *path, struct stat *sb);
extern int cg_mkdir(const char *path, mode_t mode);
extern int cg_rmdir(const char *path);
*/
#ifndef _GNU_SOURCE
-#define _GNU_SOURCE 1
+#define _GNU_SOURCE
#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <ctype.h>
#include <dirent.h>
#include <errno.h>
#include <sys/types.h>
#include <unistd.h>
+#include "../config.h"
+#include "../macro.h"
+#include "../memory_utils.h"
#include "cgroup.h"
#include "cgroup2_devices.h"
#include "cgroup_utils.h"
-#include "macro.h"
-#include "memory_utils.h"
static void free_string_list(char **clist)
{
/* SPDX-License-Identifier: LGPL-2.1+ */
#ifndef _GNU_SOURCE
-#define _GNU_SOURCE 1
+#define _GNU_SOURCE
#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <dirent.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <inttypes.h>
+#include <linux/magic.h>
+#include <linux/sched.h>
+#include <sched.h>
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdio.h>
#include <stdlib.h>
#include <string.h>
+#include <sys/epoll.h>
+#include <sys/mount.h>
+#include <sys/stat.h>
#include <sys/types.h>
+#include <sys/vfs.h>
#include <unistd.h>
+#include "../config.h"
+#include "../macro.h"
+#include "../memory_utils.h"
#include "cgroup.h"
+#include "cgroup_utils.h"
#include "cgroup2_devices.h"
extern struct cgroup_ops *cgfsng_ops_init(void);
*point = '\0';
}
}
+
+char *get_pid_cgroup(pid_t pid, const char *contrl)
+{
+ int cfd;
+
+ cfd = get_cgroup_fd(contrl);
+ if (cfd < 0)
+ return false;
+
+ if (pure_unified_layout(cgroup_ops))
+ return cg_unified_get_current_cgroup(pid);
+
+ return cg_legacy_get_current_cgroup(pid, contrl);
+}
+
+/*
+ * Read the cpuset.cpus for cg
+ * Return the answer in a newly allocated string which must be freed
+ */
+char *get_cpuset(const char *cg)
+{
+ char *value = NULL;
+ int ret;
+
+ ret = cgroup_ops->get_cpuset_cpus(cgroup_ops, cg, &value);
+ if (ret < 0)
+ return NULL;
+
+ return value;
+}
#ifndef __LXC_CGROUP_H
#define __LXC_CGROUP_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <errno.h>
#include <stdbool.h>
#include <stddef.h>
#include <sys/types.h>
-#include "macro.h"
+#include "../config.h"
+#include "../macro.h"
+#include "../memory_utils.h"
#define DEFAULT_CGROUP_MOUNTPOINT "/sys/fs/cgroup"
return h ? h->fd : -EBADF;
}
+extern char *get_pid_cgroup(pid_t pid, const char *contrl);
+
+extern char *get_cpuset(const char *cg);
+
#endif
/* Parts of this taken from systemd's implementation. */
#ifndef _GNU_SOURCE
-#define _GNU_SOURCE 1
+#define _GNU_SOURCE
#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <errno.h>
#include <fcntl.h>
#include <stdbool.h>
#include <sys/types.h>
#include <unistd.h>
+#include "../config.h"
+#include "../macro.h"
+#include "../memory_utils.h"
#include "cgroup2_devices.h"
-#include "macro.h"
-#include "memory_utils.h"
#ifdef HAVE_STRUCT_BPF_CGROUP_DEV_CTX
#include <linux/bpf.h>
#ifndef __LXC_CGROUP2_DEVICES_H
#define __LXC_CGROUP2_DEVICES_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <errno.h>
#include <fcntl.h>
#include <stdbool.h>
/* SPDX-License-Identifier: LGPL-2.1+ */
#ifndef _GNU_SOURCE
-#define _GNU_SOURCE 1
+#define _GNU_SOURCE
#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <fcntl.h>
+#include <stdarg.h>
#include <stdbool.h>
+#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/mount.h>
#include <sys/vfs.h>
#include <unistd.h>
+#include "../config.h"
+#include "../macro.h"
+#include "../memory_utils.h"
#include "cgroup.h"
#include "cgroup_utils.h"
-#include "macro.h"
-#include "memory_utils.h"
int get_cgroup_version(char *line)
{
#ifndef __LXC_CGROUP_UTILS_H
#define __LXC_CGROUP_UTILS_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <fcntl.h>
#include <stdarg.h>
#include <stdbool.h>
#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <sys/vfs.h>
+#include <unistd.h>
+
+#include "../config.h"
+#include "../macro.h"
+#include "../memory_utils.h"
/* Retrieve the cgroup version of a given entry from /proc/<pid>/mountinfo. */
extern int get_cgroup_version(char *line);
+++ /dev/null
-#include <stdio.h>
-#include <string.h>
-#include <stdbool.h>
-#include <stdlib.h>
-
-/*
- * Helper functions for cpuset_in-set
- */
-static char *cpuset_nexttok(const char *c)
-{
- char *r = strchr(c+1, ',');
- if (r)
- return r+1;
- return NULL;
-}
-
-static int cpuset_getrange(const char *c, int *a, int *b)
-{
- int ret;
-
- ret = sscanf(c, "%d-%d", a, b);
- return ret;
-}
-
-/*
- * cpusets are in format "1,2-3,4"
- * iow, comma-delimited ranges
- */
-bool cpu_in_cpuset(int cpu, const char *cpuset)
-{
- const char *c;
-
- for (c = cpuset; c; c = cpuset_nexttok(c)) {
- int a, b, ret;
-
- ret = cpuset_getrange(c, &a, &b);
- if (ret == 1 && cpu == a) // "1" or "1,6"
- return true;
- else if (ret == 2 && cpu >= a && cpu <= b) // range match
- return true;
- }
-
- return false;
-}
-
-/*
- * get cpu number in cpuset
- */
-int cpu_number_in_cpuset(const char *cpuset)
-{
- const char *c;
- int cpu_number = 0;
-
- for (c = cpuset; c; c = cpuset_nexttok(c)) {
- int a, b, ret;
-
- ret = cpuset_getrange(c, &a, &b);
- if (ret == 1)
- cpu_number++;
- else if (ret == 2)
- cpu_number += a > b ? a - b + 1 : b - a + 1;
- }
- return cpu_number;
-}
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#define __STDC_FORMAT_MACROS
+
+#include <stdio.h>
+#include <string.h>
+#include <stdbool.h>
+#include <stdlib.h>
+
+#include "bindings.h"
+#include "cgroups/cgroup.h"
+#include "cgroups/cgroup_utils.h"
+#include "config.h"
+#include "memory_utils.h"
+#include "proc_loadavg.h"
+#include "utils.h"
+
+/*
+ * Helper functions for cpuset_in-set
+ */
+static char *cpuset_nexttok(const char *c)
+{
+ char *r = strchr(c+1, ',');
+ if (r)
+ return r+1;
+ return NULL;
+}
+
+static int cpuset_getrange(const char *c, int *a, int *b)
+{
+ int ret;
+
+ ret = sscanf(c, "%d-%d", a, b);
+ return ret;
+}
+
+/*
+ * cpusets are in format "1,2-3,4"
+ * iow, comma-delimited ranges
+ */
+bool cpu_in_cpuset(int cpu, const char *cpuset)
+{
+ const char *c;
+
+ for (c = cpuset; c; c = cpuset_nexttok(c)) {
+ int a, b, ret;
+
+ ret = cpuset_getrange(c, &a, &b);
+ if (ret == 1 && cpu == a) // "1" or "1,6"
+ return true;
+ else if (ret == 2 && cpu >= a && cpu <= b) // range match
+ return true;
+ }
+
+ return false;
+}
+
+/*
+ * get cpu number in cpuset
+ */
+int cpu_number_in_cpuset(const char *cpuset)
+{
+ const char *c;
+ int cpu_number = 0;
+
+ for (c = cpuset; c; c = cpuset_nexttok(c)) {
+ int a, b, ret;
+
+ ret = cpuset_getrange(c, &a, &b);
+ if (ret == 1)
+ cpu_number++;
+ else if (ret == 2)
+ cpu_number += a > b ? a - b + 1 : b - a + 1;
+ }
+ return cpu_number;
+}
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef __LXCFS_CPUSET_PARSE_H
+#define __LXCFS_CPUSET_PARSE_H
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <fuse.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "config.h"
+#include "macro.h"
+
+extern bool cpu_in_cpuset(int cpu, const char *cpuset);
+extern int cpu_number_in_cpuset(const char *cpuset);
+extern char *get_cpuset(const char *cg);
+
+#endif /* __LXCFS_CPUSET_PARSE_H */
+
+
* See COPYING file for details.
*/
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
#include <alloca.h>
#include <dirent.h>
if (strncmp(path, "/proc", 5) == 0)
return -EPERM;
- if (strncmp(path, "/sys", 4) == 0)
+ if (strncmp(path, "/sys", 4) == 0)
return -EPERM;
return -ENOENT;
}
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
#ifndef __LXCFS_MACRO_H
#define __LXCFS_MACRO_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <stdio.h>
+#include "config.h"
+
#define BATCH_SIZE 50
/* filesystem magic values */
#ifndef __LXCFS_MEMORY_UTILS_H
#define __LXCFS_MEMORY_UTILS_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <dirent.h>
#include <errno.h>
#include <stdio.h>
#include <sys/types.h>
#include <unistd.h>
+#include "config.h"
#include "macro.h"
static inline void __auto_free__(void *p)
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#define __STDC_FORMAT_MACROS
+#include <dirent.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <fuse.h>
+#include <inttypes.h>
+#include <libgen.h>
+#include <pthread.h>
+#include <sched.h>
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include <unistd.h>
+#include <wait.h>
+#include <linux/magic.h>
+#include <linux/sched.h>
+#include <sys/epoll.h>
+#include <sys/mman.h>
+#include <sys/mount.h>
+#include <sys/param.h>
+#include <sys/socket.h>
+#include <sys/syscall.h>
+#include <sys/sysinfo.h>
+#include <sys/vfs.h>
+
+#include "bindings.h"
+#include "config.h"
+#include "cgroup_fuse.h"
+#include "cpuset_parse.h"
+#include "cgroups/cgroup.h"
+#include "cgroups/cgroup_utils.h"
+#include "memory_utils.h"
+#include "utils.h"
+
+struct cpuacct_usage {
+ uint64_t user;
+ uint64_t system;
+ uint64_t idle;
+ bool online;
+};
+
+/* Data for CPU view */
+struct cg_proc_stat {
+ char *cg;
+ struct cpuacct_usage *usage; // Real usage as read from the host's /proc/stat
+ struct cpuacct_usage *view; // Usage stats reported to the container
+ int cpu_count;
+ pthread_mutex_t lock; // For node manipulation
+ struct cg_proc_stat *next;
+};
+
+struct cg_proc_stat_head {
+ struct cg_proc_stat *next;
+ time_t lastcheck;
+
+ /*
+ * For access to the list. Reading can be parallel, pruning is exclusive.
+ */
+ pthread_rwlock_t lock;
+};
+
+#define CPUVIEW_HASH_SIZE 100
+static struct cg_proc_stat_head *proc_stat_history[CPUVIEW_HASH_SIZE];
+
+static void reset_proc_stat_node(struct cg_proc_stat *node, struct cpuacct_usage *usage, int cpu_count)
+{
+ int i;
+
+ lxcfs_debug("Resetting stat node for %s\n", node->cg);
+ memcpy(node->usage, usage, sizeof(struct cpuacct_usage) * cpu_count);
+
+ for (i = 0; i < cpu_count; i++) {
+ node->view[i].user = 0;
+ node->view[i].system = 0;
+ node->view[i].idle = 0;
+ }
+
+ node->cpu_count = cpu_count;
+}
+
+static bool expand_proc_stat_node(struct cg_proc_stat *node, int cpu_count)
+{
+ __do_free struct cpuacct_usage *new_usage = NULL, *new_view = NULL;
+
+ /* Allocate new memory */
+ new_usage = malloc(sizeof(struct cpuacct_usage) * cpu_count);
+ if (!new_usage)
+ return false;
+
+ new_view = malloc(sizeof(struct cpuacct_usage) * cpu_count);
+ if (!new_view)
+ return false;
+
+ /* Copy existing data & initialize new elements */
+ for (int i = 0; i < cpu_count; i++) {
+ if (i < node->cpu_count) {
+ new_usage[i].user = node->usage[i].user;
+ new_usage[i].system = node->usage[i].system;
+ new_usage[i].idle = node->usage[i].idle;
+
+ new_view[i].user = node->view[i].user;
+ new_view[i].system = node->view[i].system;
+ new_view[i].idle = node->view[i].idle;
+ } else {
+ new_usage[i].user = 0;
+ new_usage[i].system = 0;
+ new_usage[i].idle = 0;
+
+ new_view[i].user = 0;
+ new_view[i].system = 0;
+ new_view[i].idle = 0;
+ }
+ }
+
+ free(node->usage);
+ node->usage = move_ptr(new_usage);
+
+ free(node->view);
+ node->view = move_ptr(new_view);
+ node->cpu_count = cpu_count;
+
+ return true;
+}
+
+static struct cg_proc_stat *add_proc_stat_node(struct cg_proc_stat *new_node)
+{
+ int hash = calc_hash(new_node->cg) % CPUVIEW_HASH_SIZE;
+ struct cg_proc_stat_head *head = proc_stat_history[hash];
+ struct cg_proc_stat *node, *rv = new_node;
+
+ pthread_rwlock_wrlock(&head->lock);
+
+ if (!head->next) {
+ head->next = new_node;
+ goto out;
+ }
+
+ node = head->next;
+
+ for (;;) {
+ if (strcmp(node->cg, new_node->cg) == 0) {
+ /* The node is already present, return it */
+ free_proc_stat_node(new_node);
+ rv = node;
+ goto out;
+ }
+
+ if (node->next) {
+ node = node->next;
+ continue;
+ }
+
+ node->next = new_node;
+ goto out;
+ }
+
+out:
+ pthread_rwlock_unlock(&head->lock);
+ return rv;
+}
+
+static struct cg_proc_stat *new_proc_stat_node(struct cpuacct_usage *usage, int cpu_count, const char *cg)
+{
+ struct cg_proc_stat *node;
+ int i;
+
+ node = malloc(sizeof(struct cg_proc_stat));
+ if (!node)
+ goto err;
+
+ node->cg = NULL;
+ node->usage = NULL;
+ node->view = NULL;
+
+ node->cg = malloc(strlen(cg) + 1);
+ if (!node->cg)
+ goto err;
+
+ strcpy(node->cg, cg);
+
+ node->usage = malloc(sizeof(struct cpuacct_usage) * cpu_count);
+ if (!node->usage)
+ goto err;
+
+ memcpy(node->usage, usage, sizeof(struct cpuacct_usage) * cpu_count);
+
+ node->view = malloc(sizeof(struct cpuacct_usage) * cpu_count);
+ if (!node->view)
+ goto err;
+
+ node->cpu_count = cpu_count;
+ node->next = NULL;
+
+ if (pthread_mutex_init(&node->lock, NULL) != 0) {
+ lxcfs_error("%s\n", "Failed to initialize node lock");
+ goto err;
+ }
+
+ for (i = 0; i < cpu_count; i++) {
+ node->view[i].user = 0;
+ node->view[i].system = 0;
+ node->view[i].idle = 0;
+ }
+
+ return node;
+
+err:
+ if (node && node->cg)
+ free(node->cg);
+ if (node && node->usage)
+ free(node->usage);
+ if (node && node->view)
+ free(node->view);
+ if (node)
+ free(node);
+
+ return NULL;
+}
+
+static struct cg_proc_stat *prune_proc_stat_list(struct cg_proc_stat *node)
+{
+ struct cg_proc_stat *first = NULL, *prev, *tmp;
+
+ for (prev = NULL; node; ) {
+ if (!cgfs_param_exist("cpu", node->cg, "cpu.shares")) {
+ tmp = node;
+ lxcfs_debug("Removing stat node for %s\n", node->cg);
+
+ if (prev)
+ prev->next = node->next;
+ else
+ first = node->next;
+
+ node = node->next;
+ free_proc_stat_node(tmp);
+ } else {
+ if (!first)
+ first = node;
+ prev = node;
+ node = node->next;
+ }
+ }
+
+ return first;
+}
+
+#define PROC_STAT_PRUNE_INTERVAL 10
+static void prune_proc_stat_history(void)
+{
+ int i;
+ time_t now = time(NULL);
+
+ for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
+ pthread_rwlock_wrlock(&proc_stat_history[i]->lock);
+
+ if ((proc_stat_history[i]->lastcheck + PROC_STAT_PRUNE_INTERVAL) > now) {
+ pthread_rwlock_unlock(&proc_stat_history[i]->lock);
+ return;
+ }
+
+ if (proc_stat_history[i]->next) {
+ proc_stat_history[i]->next = prune_proc_stat_list(proc_stat_history[i]->next);
+ proc_stat_history[i]->lastcheck = now;
+ }
+
+ pthread_rwlock_unlock(&proc_stat_history[i]->lock);
+ }
+}
+
+static struct cg_proc_stat *find_proc_stat_node(struct cg_proc_stat_head *head,
+ const char *cg)
+{
+ struct cg_proc_stat *node;
+
+ pthread_rwlock_rdlock(&head->lock);
+
+ if (!head->next) {
+ pthread_rwlock_unlock(&head->lock);
+ return NULL;
+ }
+
+ node = head->next;
+
+ do {
+ if (strcmp(cg, node->cg) == 0)
+ goto out;
+ } while ((node = node->next));
+
+ node = NULL;
+
+out:
+ pthread_rwlock_unlock(&head->lock);
+ prune_proc_stat_history();
+ return node;
+}
+
+static struct cg_proc_stat *find_or_create_proc_stat_node(struct cpuacct_usage *usage, int cpu_count, const char *cg)
+{
+ int hash = calc_hash(cg) % CPUVIEW_HASH_SIZE;
+ struct cg_proc_stat_head *head = proc_stat_history[hash];
+ struct cg_proc_stat *node;
+
+ node = find_proc_stat_node(head, cg);
+
+ if (!node) {
+ node = new_proc_stat_node(usage, cpu_count, cg);
+ if (!node)
+ return NULL;
+
+ node = add_proc_stat_node(node);
+ lxcfs_debug("New stat node (%d) for %s\n", cpu_count, cg);
+ }
+
+ pthread_mutex_lock(&node->lock);
+
+ /* If additional CPUs on the host have been enabled, CPU usage counter
+ * arrays have to be expanded */
+ if (node->cpu_count < cpu_count) {
+ lxcfs_debug("Expanding stat node %d->%d for %s\n",
+ node->cpu_count, cpu_count, cg);
+
+ if (!expand_proc_stat_node(node, cpu_count)) {
+ pthread_mutex_unlock(&node->lock);
+ lxcfs_debug("Unable to expand stat node %d->%d for %s\n",
+ node->cpu_count, cpu_count, cg);
+ return NULL;
+ }
+ }
+
+ return node;
+}
+
+static void add_cpu_usage(unsigned long *surplus, struct cpuacct_usage *usage,
+ unsigned long *counter, unsigned long threshold)
+{
+ unsigned long free_space, to_add;
+
+ free_space = threshold - usage->user - usage->system;
+
+ if (free_space > usage->idle)
+ free_space = usage->idle;
+
+ to_add = free_space > *surplus ? *surplus : free_space;
+
+ *counter += to_add;
+ usage->idle -= to_add;
+ *surplus -= to_add;
+}
+
+static unsigned long diff_cpu_usage(struct cpuacct_usage *older,
+ struct cpuacct_usage *newer,
+ struct cpuacct_usage *diff, int cpu_count)
+{
+ int i;
+ unsigned long sum = 0;
+
+ for (i = 0; i < cpu_count; i++) {
+ if (!newer[i].online)
+ continue;
+
+ /* When cpuset is changed on the fly, the CPUs might get reordered.
+ * We could either reset all counters, or check that the substractions
+ * below will return expected results.
+ */
+ if (newer[i].user > older[i].user)
+ diff[i].user = newer[i].user - older[i].user;
+ else
+ diff[i].user = 0;
+
+ if (newer[i].system > older[i].system)
+ diff[i].system = newer[i].system - older[i].system;
+ else
+ diff[i].system = 0;
+
+ if (newer[i].idle > older[i].idle)
+ diff[i].idle = newer[i].idle - older[i].idle;
+ else
+ diff[i].idle = 0;
+
+ sum += diff[i].user;
+ sum += diff[i].system;
+ sum += diff[i].idle;
+ }
+
+ return sum;
+}
+
+/*
+ * Read cgroup CPU quota parameters from `cpu.cfs_quota_us` or `cpu.cfs_period_us`,
+ * depending on `param`. Parameter value is returned throuh `value`.
+ */
+static bool read_cpu_cfs_param(const char *cg, const char *param, int64_t *value)
+{
+ __do_free char *str = NULL;
+ char file[11 + 6 + 1]; /* cpu.cfs__us + quota/period + \0 */
+
+ snprintf(file, sizeof(file), "cpu.cfs_%s_us", param);
+
+ if (!cgroup_ops->get(cgroup_ops, "cpu", cg, file, &str))
+ return false;
+
+ if (sscanf(str, "%ld", value) != 1)
+ return false;
+
+ return true;
+}
+
+/*
+ * Return the exact number of visible CPUs based on CPU quotas.
+ * If there is no quota set, zero is returned.
+ */
+static double exact_cpu_count(const char *cg)
+{
+ double rv;
+ int nprocs;
+ int64_t cfs_quota, cfs_period;
+
+ if (!read_cpu_cfs_param(cg, "quota", &cfs_quota))
+ return 0;
+
+ if (!read_cpu_cfs_param(cg, "period", &cfs_period))
+ return 0;
+
+ if (cfs_quota <= 0 || cfs_period <= 0)
+ return 0;
+
+ rv = (double)cfs_quota / (double)cfs_period;
+
+ nprocs = get_nprocs();
+
+ if (rv > nprocs)
+ rv = nprocs;
+
+ return rv;
+}
+
+/*
+ * Return the maximum number of visible CPUs based on CPU quotas.
+ * If there is no quota set, zero is returned.
+ */
+static int max_cpu_count(const char *cg)
+{
+ int rv, nprocs;
+ int64_t cfs_quota, cfs_period;
+ int nr_cpus_in_cpuset = 0;
+ char *cpuset = NULL;
+
+ if (!read_cpu_cfs_param(cg, "quota", &cfs_quota))
+ return 0;
+
+ if (!read_cpu_cfs_param(cg, "period", &cfs_period))
+ return 0;
+
+ cpuset = get_cpuset(cg);
+ if (cpuset)
+ nr_cpus_in_cpuset = cpu_number_in_cpuset(cpuset);
+
+ if (cfs_quota <= 0 || cfs_period <= 0){
+ if (nr_cpus_in_cpuset > 0)
+ return nr_cpus_in_cpuset;
+
+ return 0;
+ }
+
+ rv = cfs_quota / cfs_period;
+
+ /* In case quota/period does not yield a whole number, add one CPU for
+ * the remainder.
+ */
+ if ((cfs_quota % cfs_period) > 0)
+ rv += 1;
+
+ nprocs = get_nprocs();
+
+ if (rv > nprocs)
+ rv = nprocs;
+
+ /* use min value in cpu quota and cpuset */
+ if (nr_cpus_in_cpuset > 0 && nr_cpus_in_cpuset < rv)
+ rv = nr_cpus_in_cpuset;
+
+ return rv;
+}
+
+int cpuview_proc_stat(const char *cg, const char *cpuset,
+ struct cpuacct_usage *cg_cpu_usage, int cg_cpu_usage_size,
+ FILE *f, char *buf, size_t buf_size)
+{
+ __do_free char *line = NULL;
+ __do_free struct cpuacct_usage *diff = NULL;
+ size_t linelen = 0, total_len = 0, l;
+ int curcpu = -1; /* cpu numbering starts at 0 */
+ int physcpu, i;
+ int max_cpus = max_cpu_count(cg), cpu_cnt = 0;
+ unsigned long user = 0, nice = 0, system = 0, idle = 0, iowait = 0,
+ irq = 0, softirq = 0, steal = 0, guest = 0, guest_nice = 0;
+ unsigned long user_sum = 0, system_sum = 0, idle_sum = 0;
+ unsigned long user_surplus = 0, system_surplus = 0;
+ unsigned long total_sum, threshold;
+ struct cg_proc_stat *stat_node;
+ int nprocs = get_nprocs_conf();
+
+ if (cg_cpu_usage_size < nprocs)
+ nprocs = cg_cpu_usage_size;
+
+ /* Read all CPU stats and stop when we've encountered other lines */
+ while (getline(&line, &linelen, f) != -1) {
+ int ret;
+ char cpu_char[10]; /* That's a lot of cores */
+ uint64_t all_used, cg_used;
+
+ if (strlen(line) == 0)
+ continue;
+
+ /* not a ^cpuN line containing a number N */
+ if (sscanf(line, "cpu%9[^ ]", cpu_char) != 1)
+ break;
+
+ if (sscanf(cpu_char, "%d", &physcpu) != 1)
+ continue;
+
+ if (physcpu >= cg_cpu_usage_size)
+ continue;
+
+ curcpu ++;
+ cpu_cnt ++;
+
+ if (!cpu_in_cpuset(physcpu, cpuset)) {
+ for (i = curcpu; i <= physcpu; i++)
+ cg_cpu_usage[i].online = false;
+ continue;
+ }
+
+ if (curcpu < physcpu) {
+ /* Some CPUs may be disabled */
+ for (i = curcpu; i < physcpu; i++)
+ cg_cpu_usage[i].online = false;
+
+ curcpu = physcpu;
+ }
+
+ cg_cpu_usage[curcpu].online = true;
+
+ ret = sscanf(line, "%*s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
+ &user,
+ &nice,
+ &system,
+ &idle,
+ &iowait,
+ &irq,
+ &softirq,
+ &steal,
+ &guest,
+ &guest_nice);
+
+ if (ret != 10)
+ continue;
+
+ all_used = user + nice + system + iowait + irq + softirq + steal + guest + guest_nice;
+ cg_used = cg_cpu_usage[curcpu].user + cg_cpu_usage[curcpu].system;
+
+ if (all_used >= cg_used) {
+ cg_cpu_usage[curcpu].idle = idle + (all_used - cg_used);
+
+ } else {
+ lxcfs_error("cpu%d from %s has unexpected cpu time: %lu in /proc/stat, "
+ "%lu in cpuacct.usage_all; unable to determine idle time\n",
+ curcpu, cg, all_used, cg_used);
+ cg_cpu_usage[curcpu].idle = idle;
+ }
+ }
+
+ /* Cannot use more CPUs than is available due to cpuset */
+ if (max_cpus > cpu_cnt)
+ max_cpus = cpu_cnt;
+
+ stat_node = find_or_create_proc_stat_node(cg_cpu_usage, nprocs, cg);
+
+ if (!stat_node) {
+ lxcfs_error("unable to find/create stat node for %s\n", cg);
+ return 0;
+ }
+
+ diff = malloc(sizeof(struct cpuacct_usage) * nprocs);
+ if (!diff) {
+ return 0;
+ }
+
+ /*
+ * If the new values are LOWER than values stored in memory, it means
+ * the cgroup has been reset/recreated and we should reset too.
+ */
+ for (curcpu = 0; curcpu < nprocs; curcpu++) {
+ if (!cg_cpu_usage[curcpu].online)
+ continue;
+
+ if (cg_cpu_usage[curcpu].user < stat_node->usage[curcpu].user)
+ reset_proc_stat_node(stat_node, cg_cpu_usage, nprocs);
+
+ break;
+ }
+
+ total_sum = diff_cpu_usage(stat_node->usage, cg_cpu_usage, diff, nprocs);
+
+ for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
+ stat_node->usage[curcpu].online = cg_cpu_usage[curcpu].online;
+
+ if (!stat_node->usage[curcpu].online)
+ continue;
+
+ i++;
+
+ stat_node->usage[curcpu].user += diff[curcpu].user;
+ stat_node->usage[curcpu].system += diff[curcpu].system;
+ stat_node->usage[curcpu].idle += diff[curcpu].idle;
+
+ if (max_cpus > 0 && i >= max_cpus) {
+ user_surplus += diff[curcpu].user;
+ system_surplus += diff[curcpu].system;
+ }
+ }
+
+ /* Calculate usage counters of visible CPUs */
+ if (max_cpus > 0) {
+ unsigned long diff_user = 0;
+ unsigned long diff_system = 0;
+ unsigned long diff_idle = 0;
+ unsigned long max_diff_idle = 0;
+ unsigned long max_diff_idle_index = 0;
+ double exact_cpus;
+
+ /* threshold = maximum usage per cpu, including idle */
+ threshold = total_sum / cpu_cnt * max_cpus;
+
+ for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
+ if (!stat_node->usage[curcpu].online)
+ continue;
+
+ i++;
+
+ if (i == max_cpus)
+ break;
+
+ if (diff[curcpu].user + diff[curcpu].system >= threshold)
+ continue;
+
+ /* Add user */
+ add_cpu_usage(&user_surplus, &diff[curcpu],
+ &diff[curcpu].user, threshold);
+
+ if (diff[curcpu].user + diff[curcpu].system >= threshold)
+ continue;
+
+ /* If there is still room, add system */
+ add_cpu_usage(&system_surplus, &diff[curcpu],
+ &diff[curcpu].system, threshold);
+ }
+
+ if (user_surplus > 0)
+ lxcfs_debug("leftover user: %lu for %s\n", user_surplus, cg);
+ if (system_surplus > 0)
+ lxcfs_debug("leftover system: %lu for %s\n", system_surplus, cg);
+
+ for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
+ if (!stat_node->usage[curcpu].online)
+ continue;
+
+ i++;
+
+ if (i == max_cpus)
+ break;
+
+ stat_node->view[curcpu].user += diff[curcpu].user;
+ stat_node->view[curcpu].system += diff[curcpu].system;
+ stat_node->view[curcpu].idle += diff[curcpu].idle;
+
+ user_sum += stat_node->view[curcpu].user;
+ system_sum += stat_node->view[curcpu].system;
+ idle_sum += stat_node->view[curcpu].idle;
+
+ diff_user += diff[curcpu].user;
+ diff_system += diff[curcpu].system;
+ diff_idle += diff[curcpu].idle;
+ if (diff[curcpu].idle > max_diff_idle) {
+ max_diff_idle = diff[curcpu].idle;
+ max_diff_idle_index = curcpu;
+ }
+
+ lxcfs_v("curcpu: %d, diff_user: %lu, diff_system: %lu, diff_idle: %lu\n", curcpu, diff[curcpu].user, diff[curcpu].system, diff[curcpu].idle);
+ }
+ lxcfs_v("total. diff_user: %lu, diff_system: %lu, diff_idle: %lu\n", diff_user, diff_system, diff_idle);
+
+ /* revise cpu usage view to support partial cpu case. */
+ exact_cpus = exact_cpu_count(cg);
+ if (exact_cpus < (double)max_cpus){
+ unsigned long delta = (unsigned long)((double)(diff_user + diff_system + diff_idle) * (1 - exact_cpus / (double)max_cpus));
+
+ lxcfs_v("revising cpu usage view to match the exact cpu count [%f]\n", exact_cpus);
+ lxcfs_v("delta: %lu\n", delta);
+ lxcfs_v("idle_sum before: %lu\n", idle_sum);
+ idle_sum = idle_sum > delta ? idle_sum - delta : 0;
+ lxcfs_v("idle_sum after: %lu\n", idle_sum);
+
+ curcpu = max_diff_idle_index;
+ lxcfs_v("curcpu: %d, idle before: %lu\n", curcpu, stat_node->view[curcpu].idle);
+ stat_node->view[curcpu].idle = stat_node->view[curcpu].idle > delta ? stat_node->view[curcpu].idle - delta : 0;
+ lxcfs_v("curcpu: %d, idle after: %lu\n", curcpu, stat_node->view[curcpu].idle);
+ }
+ } else {
+ for (curcpu = 0; curcpu < nprocs; curcpu++) {
+ if (!stat_node->usage[curcpu].online)
+ continue;
+
+ stat_node->view[curcpu].user = stat_node->usage[curcpu].user;
+ stat_node->view[curcpu].system = stat_node->usage[curcpu].system;
+ stat_node->view[curcpu].idle = stat_node->usage[curcpu].idle;
+
+ user_sum += stat_node->view[curcpu].user;
+ system_sum += stat_node->view[curcpu].system;
+ idle_sum += stat_node->view[curcpu].idle;
+ }
+ }
+
+ /* Render the file */
+ /* cpu-all */
+ l = snprintf(buf, buf_size, "cpu %lu 0 %lu %lu 0 0 0 0 0 0\n",
+ user_sum,
+ system_sum,
+ idle_sum);
+ lxcfs_v("cpu-all: %s\n", buf);
+
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+ if (l >= buf_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+
+ buf += l;
+ buf_size -= l;
+ total_len += l;
+
+ /* Render visible CPUs */
+ for (curcpu = 0, i = -1; curcpu < nprocs; curcpu++) {
+ if (!stat_node->usage[curcpu].online)
+ continue;
+
+ i++;
+
+ if (max_cpus > 0 && i == max_cpus)
+ break;
+
+ l = snprintf(buf, buf_size, "cpu%d %lu 0 %lu %lu 0 0 0 0 0 0\n",
+ i,
+ stat_node->view[curcpu].user,
+ stat_node->view[curcpu].system,
+ stat_node->view[curcpu].idle);
+ lxcfs_v("cpu: %s\n", buf);
+
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+
+ }
+ if (l >= buf_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+
+ buf += l;
+ buf_size -= l;
+ total_len += l;
+ }
+
+ /* Pass the rest of /proc/stat, start with the last line read */
+ l = snprintf(buf, buf_size, "%s", line);
+
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+
+ }
+ if (l >= buf_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+
+ buf += l;
+ buf_size -= l;
+ total_len += l;
+
+ /* Pass the rest of the host's /proc/stat */
+ while (getline(&line, &linelen, f) != -1) {
+ l = snprintf(buf, buf_size, "%s", line);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+ if (l >= buf_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+ buf += l;
+ buf_size -= l;
+ total_len += l;
+ }
+
+ if (stat_node)
+ pthread_mutex_unlock(&stat_node->lock);
+ return total_len;
+}
+
+/*
+ * check whether this is a '^processor" line in /proc/cpuinfo
+ */
+static bool is_processor_line(const char *line)
+{
+ int cpu;
+
+ if (sscanf(line, "processor : %d", &cpu) == 1)
+ return true;
+ return false;
+}
+
+static bool cpuline_in_cpuset(const char *line, const char *cpuset)
+{
+ int cpu;
+
+ if (sscanf(line, "processor : %d", &cpu) != 1)
+ return false;
+ return cpu_in_cpuset(cpu, cpuset);
+}
+
+int proc_cpuinfo_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ __do_free char *cg = NULL, *cpuset = NULL, *line = NULL;
+ __do_fclose FILE *f = NULL;
+ struct fuse_context *fc = fuse_get_context();
+ struct file_info *d = (struct file_info *)fi->fh;
+ size_t linelen = 0, total_len = 0;
+ bool am_printing = false, firstline = true, is_s390x = false;
+ int curcpu = -1, cpu, max_cpus = 0;
+ bool use_view;
+ char *cache = d->buf;
+ size_t cache_size = d->buflen;
+
+ if (offset){
+ int left;
+
+ if (offset > d->size)
+ return -EINVAL;
+
+ if (!d->cached)
+ return 0;
+
+ left = d->size - offset;
+ total_len = left > size ? size: left;
+ memcpy(buf, cache + offset, total_len);
+
+ return total_len;
+ }
+
+ pid_t initpid = lookup_initpid_in_store(fc->pid);
+ if (initpid <= 1 || is_shared_pidns(initpid))
+ initpid = fc->pid;
+ cg = get_pid_cgroup(initpid, "cpuset");
+ if (!cg)
+ return read_file_fuse("proc/cpuinfo", buf, size, d);
+ prune_init_slice(cg);
+
+ cpuset = get_cpuset(cg);
+ if (!cpuset)
+ return 0;
+
+ use_view = cgroup_ops->can_use_cpuview(cgroup_ops);
+ if (use_view)
+ max_cpus = max_cpu_count(cg);
+
+ f = fopen("/proc/cpuinfo", "r");
+ if (!f)
+ return 0;
+
+ while (getline(&line, &linelen, f) != -1) {
+ ssize_t l;
+ if (firstline) {
+ firstline = false;
+ if (strstr(line, "IBM/S390") != NULL) {
+ is_s390x = true;
+ am_printing = true;
+ continue;
+ }
+ }
+ if (strncmp(line, "# processors:", 12) == 0)
+ continue;
+ if (is_processor_line(line)) {
+ if (use_view && max_cpus > 0 && (curcpu+1) == max_cpus)
+ break;
+ am_printing = cpuline_in_cpuset(line, cpuset);
+ if (am_printing) {
+ curcpu ++;
+ l = snprintf(cache, cache_size, "processor : %d\n", curcpu);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+ }
+ continue;
+ } else if (is_s390x && sscanf(line, "processor %d:", &cpu) == 1) {
+ char *p;
+ if (use_view && max_cpus > 0 && (curcpu+1) == max_cpus)
+ break;
+ if (!cpu_in_cpuset(cpu, cpuset))
+ continue;
+ curcpu ++;
+ p = strchr(line, ':');
+ if (!p || !*p)
+ return 0;
+ p++;
+ l = snprintf(cache, cache_size, "processor %d:%s", curcpu, p);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+ continue;
+
+ }
+ if (am_printing) {
+ l = snprintf(cache, cache_size, "%s", line);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+ }
+ }
+
+ if (is_s390x) {
+ __do_free char *origcache = d->buf;
+ ssize_t l;
+
+ d->buf = malloc(d->buflen);
+ if (!d->buf) {
+ d->buf = move_ptr(origcache);
+ return 0;
+ }
+
+ cache = d->buf;
+ cache_size = d->buflen;
+ total_len = 0;
+ l = snprintf(cache, cache_size, "vendor_id : IBM/S390\n");
+ if (l < 0 || l >= cache_size)
+ return 0;
+
+ cache_size -= l;
+ cache += l;
+ total_len += l;
+ l = snprintf(cache, cache_size, "# processors : %d\n", curcpu + 1);
+ if (l < 0 || l >= cache_size)
+ return 0;
+
+ cache_size -= l;
+ cache += l;
+ total_len += l;
+ l = snprintf(cache, cache_size, "%s", origcache);
+ if (l < 0 || l >= cache_size)
+ return 0;
+ total_len += l;
+ }
+
+ d->cached = 1;
+ d->size = total_len;
+ if (total_len > size ) total_len = size;
+
+ /* read from off 0 */
+ memcpy(buf, d->buf, total_len);
+ return total_len;
+}
+
+/*
+ * Returns 0 on success.
+ * It is the caller's responsibility to free `return_usage`, unless this
+ * function returns an error.
+ */
+int read_cpuacct_usage_all(char *cg, char *cpuset,
+ struct cpuacct_usage **return_usage, int *size)
+{
+ __do_free char *usage_str = NULL;
+ __do_free struct cpuacct_usage *cpu_usage = NULL;
+ int cpucount = get_nprocs_conf();
+ int read_pos = 0, read_cnt=0;
+ int i, j, ret;
+ int cg_cpu;
+ uint64_t cg_user, cg_system;
+ int64_t ticks_per_sec;
+
+ ticks_per_sec = sysconf(_SC_CLK_TCK);
+
+ if (ticks_per_sec < 0 && errno == EINVAL) {
+ lxcfs_v(
+ "%s\n",
+ "read_cpuacct_usage_all failed to determine number of clock ticks "
+ "in a second");
+ return -1;
+ }
+
+ cpu_usage = malloc(sizeof(struct cpuacct_usage) * cpucount);
+ if (!cpu_usage)
+ return -ENOMEM;
+
+ memset(cpu_usage, 0, sizeof(struct cpuacct_usage) * cpucount);
+ if (!cgroup_ops->get(cgroup_ops, "cpuacct", cg, "cpuacct.usage_all", &usage_str)) {
+ char *data = NULL;
+ int i = 0, read_pos = 0, read_cnt=0;
+ size_t sz = 0, asz = 0;
+
+ /* read cpuacct.usage_percpu instead. */
+ lxcfs_v("failed to read cpuacct.usage_all. reading cpuacct.usage_percpu instead\n%s", "");
+ if (!cgroup_ops->get(cgroup_ops, "cpuacct", cg, "cpuacct.usage_percpu", &usage_str))
+ return -1;
+ lxcfs_v("usage_str: %s\n", usage_str);
+
+ /* convert cpuacct.usage_percpu into cpuacct.usage_all. */
+ lxcfs_v("converting cpuacct.usage_percpu into cpuacct.usage_all\n%s", "");
+
+ must_strcat(&data, &sz, &asz, "cpu user system\n");
+
+ while (sscanf(usage_str + read_pos, "%lu %n", &cg_user, &read_cnt) > 0) {
+ lxcfs_debug("i: %d, cg_user: %lu, read_pos: %d, read_cnt: %d\n", i, cg_user, read_pos, read_cnt);
+ must_strcat(&data, &sz, &asz, "%d %lu 0\n", i, cg_user);
+ i++;
+ read_pos += read_cnt;
+ }
+
+ usage_str = data;
+
+ lxcfs_v("usage_str: %s\n", usage_str);
+ }
+
+ if (sscanf(usage_str, "cpu user system\n%n", &read_cnt) != 0) {
+ lxcfs_error("read_cpuacct_usage_all reading first line from "
+ "%s/cpuacct.usage_all failed.\n", cg);
+ return -1;
+ }
+
+ read_pos += read_cnt;
+
+ for (i = 0, j = 0; i < cpucount; i++) {
+ ret = sscanf(usage_str + read_pos, "%d %lu %lu\n%n", &cg_cpu, &cg_user,
+ &cg_system, &read_cnt);
+
+ if (ret == EOF)
+ break;
+
+ if (ret != 3) {
+ lxcfs_error("read_cpuacct_usage_all reading from %s/cpuacct.usage_all "
+ "failed.\n", cg);
+ return -1;
+ }
+
+ read_pos += read_cnt;
+
+ /* Convert the time from nanoseconds to USER_HZ */
+ cpu_usage[j].user = cg_user / 1000.0 / 1000 / 1000 * ticks_per_sec;
+ cpu_usage[j].system = cg_system / 1000.0 / 1000 / 1000 * ticks_per_sec;
+ j++;
+ }
+
+ *return_usage = move_ptr(cpu_usage);
+ *size = cpucount;
+ return 0;
+}
+
+static bool cpuview_init_head(struct cg_proc_stat_head **head)
+{
+ *head = malloc(sizeof(struct cg_proc_stat_head));
+ if (!(*head)) {
+ lxcfs_error("%s\n", strerror(errno));
+ return false;
+ }
+
+ (*head)->lastcheck = time(NULL);
+ (*head)->next = NULL;
+
+ if (pthread_rwlock_init(&(*head)->lock, NULL) != 0) {
+ lxcfs_error("%s\n", "Failed to initialize list lock");
+ free_disarm(*head);
+ return false;
+ }
+
+ return true;
+}
+
+bool init_cpuview()
+{
+ int i;
+
+ for (i = 0; i < CPUVIEW_HASH_SIZE; i++)
+ proc_stat_history[i] = NULL;
+
+ for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
+ if (!cpuview_init_head(&proc_stat_history[i]))
+ goto err;
+ }
+
+ return true;
+
+err:
+ for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
+ if (proc_stat_history[i])
+ free_disarm(proc_stat_history[i]);
+ }
+
+ return false;
+}
+
+static void free_proc_stat_node(struct cg_proc_stat *node)
+{
+ pthread_mutex_destroy(&node->lock);
+ free_disarm(node->cg);
+ free_disarm(node->usage);
+ free_disarm(node->view);
+ free_disarm(node);
+}
+
+static void cpuview_free_head(struct cg_proc_stat_head *head)
+{
+ struct cg_proc_stat *node, *tmp;
+
+ if (head->next) {
+ node = head->next;
+
+ for (;;) {
+ tmp = node;
+ node = node->next;
+ free_proc_stat_node(tmp);
+
+ if (!node)
+ break;
+ }
+ }
+
+ pthread_rwlock_destroy(&head->lock);
+ free_disarm(head);
+}
+
+void free_cpuview()
+{
+ int i;
+
+ for (i = 0; i < CPUVIEW_HASH_SIZE; i++) {
+ if (proc_stat_history[i])
+ cpuview_free_head(proc_stat_history[i]);
+ }
+}
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef __LXCFS_PROC_CPUVIEW_FUSE_H
+#define __LXCFS_PROC_CPUVIEW_FUSE_H
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "config.h"
+#include "macro.h"
+
+struct cpuacct_usage;
+
+extern int cpuview_proc_stat(const char *cg, const char *cpuset,
+ struct cpuacct_usage *cg_cpu_usage,
+ int cg_cpu_usage_size, FILE *f, char *buf,
+ size_t buf_size);
+extern int proc_cpuinfo_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi);
+extern int read_cpuacct_usage_all(char *cg, char *cpuset,
+ struct cpuacct_usage **return_usage, int *size);
+extern bool init_cpuview();
+extern void free_cpuview();
+
+#endif /* __LXCFS_PROC_CPUVIEW_FUSE_H */
+
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#define __STDC_FORMAT_MACROS
+#include <dirent.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <fuse.h>
+#include <inttypes.h>
+#include <libgen.h>
+#include <pthread.h>
+#include <sched.h>
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include <unistd.h>
+#include <wait.h>
+#include <linux/magic.h>
+#include <linux/sched.h>
+#include <sys/epoll.h>
+#include <sys/mman.h>
+#include <sys/mount.h>
+#include <sys/param.h>
+#include <sys/socket.h>
+#include <sys/syscall.h>
+#include <sys/sysinfo.h>
+#include <sys/vfs.h>
+
+#include "bindings.h"
+#include "config.h"
+#include "cgroup_fuse.h"
+#include "cgroups/cgroup.h"
+#include "cgroups/cgroup_utils.h"
+#include "memory_utils.h"
+#include "proc_loadavg.h"
+#include "utils.h"
+
+int proc_getattr(const char *path, struct stat *sb)
+{
+ struct timespec now;
+
+ memset(sb, 0, sizeof(struct stat));
+ if (clock_gettime(CLOCK_REALTIME, &now) < 0)
+ return -EINVAL;
+ sb->st_uid = sb->st_gid = 0;
+ sb->st_atim = sb->st_mtim = sb->st_ctim = now;
+ if (strcmp(path, "/proc") == 0) {
+ sb->st_mode = S_IFDIR | 00555;
+ sb->st_nlink = 2;
+ return 0;
+ }
+ if (strcmp(path, "/proc/meminfo") == 0 ||
+ strcmp(path, "/proc/cpuinfo") == 0 ||
+ strcmp(path, "/proc/uptime") == 0 ||
+ strcmp(path, "/proc/stat") == 0 ||
+ strcmp(path, "/proc/diskstats") == 0 ||
+ strcmp(path, "/proc/swaps") == 0 ||
+ strcmp(path, "/proc/loadavg") == 0) {
+ sb->st_size = 0;
+ sb->st_mode = S_IFREG | 00444;
+ sb->st_nlink = 1;
+ return 0;
+ }
+
+ return -ENOENT;
+}
+
+int proc_readdir(const char *path, void *buf, fuse_fill_dir_t filler,
+ off_t offset, struct fuse_file_info *fi)
+{
+ if (filler(buf, ".", NULL, 0) != 0 ||
+ filler(buf, "..", NULL, 0) != 0 ||
+ filler(buf, "cpuinfo", NULL, 0) != 0 ||
+ filler(buf, "meminfo", NULL, 0) != 0 ||
+ filler(buf, "stat", NULL, 0) != 0 ||
+ filler(buf, "uptime", NULL, 0) != 0 ||
+ filler(buf, "diskstats", NULL, 0) != 0 ||
+ filler(buf, "swaps", NULL, 0) != 0 ||
+ filler(buf, "loadavg", NULL, 0) != 0)
+ return -EINVAL;
+
+ return 0;
+}
+
+static off_t get_procfile_size(const char *which)
+{
+ FILE *f = fopen(which, "r");
+ char *line = NULL;
+ size_t len = 0;
+ ssize_t sz, answer = 0;
+ if (!f)
+ return 0;
+
+ while ((sz = getline(&line, &len, f)) != -1)
+ answer += sz;
+ fclose (f);
+ free(line);
+
+ return answer;
+}
+
+int proc_open(const char *path, struct fuse_file_info *fi)
+{
+ int type = -1;
+ struct file_info *info;
+
+ if (strcmp(path, "/proc/meminfo") == 0)
+ type = LXC_TYPE_PROC_MEMINFO;
+ else if (strcmp(path, "/proc/cpuinfo") == 0)
+ type = LXC_TYPE_PROC_CPUINFO;
+ else if (strcmp(path, "/proc/uptime") == 0)
+ type = LXC_TYPE_PROC_UPTIME;
+ else if (strcmp(path, "/proc/stat") == 0)
+ type = LXC_TYPE_PROC_STAT;
+ else if (strcmp(path, "/proc/diskstats") == 0)
+ type = LXC_TYPE_PROC_DISKSTATS;
+ else if (strcmp(path, "/proc/swaps") == 0)
+ type = LXC_TYPE_PROC_SWAPS;
+ else if (strcmp(path, "/proc/loadavg") == 0)
+ type = LXC_TYPE_PROC_LOADAVG;
+ if (type == -1)
+ return -ENOENT;
+
+ info = malloc(sizeof(*info));
+ if (!info)
+ return -ENOMEM;
+
+ memset(info, 0, sizeof(*info));
+ info->type = type;
+
+ info->buflen = get_procfile_size(path) + BUF_RESERVE_SIZE;
+
+ info->buf = malloc(info->buflen);
+ if (!info->buf)
+ return -ENOMEM;
+
+ memset(info->buf, 0, info->buflen);
+ /* set actual size to buffer size */
+ info->size = info->buflen;
+
+ fi->fh = (unsigned long)info;
+ return 0;
+}
+
+int proc_access(const char *path, int mask)
+{
+ if (strcmp(path, "/proc") == 0 && access(path, R_OK) == 0)
+ return 0;
+
+ /* these are all read-only */
+ if ((mask & ~R_OK) != 0)
+ return -EACCES;
+ return 0;
+}
+
+int proc_release(const char *path, struct fuse_file_info *fi)
+{
+ do_release_file_info(fi);
+ return 0;
+}
+
+static unsigned long get_memlimit(const char *cgroup, bool swap)
+{
+ int ret;
+ __do_free char *memlimit_str = NULL;
+ unsigned long memlimit = -1;
+
+ if (swap)
+ ret = cgroup_ops->get_memory_swap_max(cgroup_ops, cgroup, &memlimit_str);
+ else
+ ret = cgroup_ops->get_memory_max(cgroup_ops, cgroup, &memlimit_str);
+ if (ret > 0)
+ memlimit = strtoul(memlimit_str, NULL, 10);
+
+ return memlimit;
+}
+
+static unsigned long get_min_memlimit(const char *cgroup, bool swap)
+{
+ __do_free char *copy = NULL;
+ unsigned long memlimit = 0;
+ unsigned long retlimit;
+
+ copy = strdup(cgroup);
+ retlimit = get_memlimit(copy, swap);
+
+ while (strcmp(copy, "/") != 0) {
+ char *it = copy;
+
+ it = dirname(it);
+ memlimit = get_memlimit(it, swap);
+ if (memlimit != -1 && memlimit < retlimit)
+ retlimit = memlimit;
+ };
+
+ return retlimit;
+}
+
+static int proc_swaps_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ __do_free char *cg = NULL, *memswlimit_str = NULL, *memusage_str = NULL,
+ *memswusage_str = NULL;
+ struct fuse_context *fc = fuse_get_context();
+ struct file_info *d = (struct file_info *)fi->fh;
+ unsigned long memswlimit = 0, memlimit = 0, memusage = 0,
+ memswusage = 0, swap_total = 0, swap_free = 0;
+ ssize_t total_len = 0;
+ ssize_t l = 0;
+ char *cache = d->buf;
+ int ret;
+
+ if (offset) {
+ int left;
+
+ if (offset > d->size)
+ return -EINVAL;
+
+ if (!d->cached)
+ return 0;
+
+ left = d->size - offset;
+ total_len = left > size ? size: left;
+ memcpy(buf, cache + offset, total_len);
+
+ return total_len;
+ }
+
+ pid_t initpid = lookup_initpid_in_store(fc->pid);
+ if (initpid <= 1 || is_shared_pidns(initpid))
+ initpid = fc->pid;
+ cg = get_pid_cgroup(initpid, "memory");
+ if (!cg)
+ return read_file_fuse("/proc/swaps", buf, size, d);
+ prune_init_slice(cg);
+
+ memlimit = get_min_memlimit(cg, false);
+
+ ret = cgroup_ops->get_memory_current(cgroup_ops, cg, &memusage_str);
+ if (ret < 0)
+ return 0;
+
+ memusage = strtoul(memusage_str, NULL, 10);
+
+ ret = cgroup_ops->get_memory_swap_max(cgroup_ops, cg, &memswlimit_str);
+ if (ret >= 0)
+ ret = cgroup_ops->get_memory_swap_current(cgroup_ops, cg, &memswusage_str);
+ if (ret >= 0) {
+ memswlimit = get_min_memlimit(cg, true);
+ memswusage = strtoul(memswusage_str, NULL, 10);
+ swap_total = (memswlimit - memlimit) / 1024;
+ swap_free = (memswusage - memusage) / 1024;
+ }
+
+ total_len = snprintf(d->buf, d->size, "Filename\t\t\t\tType\t\tSize\tUsed\tPriority\n");
+
+ /* When no mem + swap limit is specified or swapaccount=0*/
+ if (!memswlimit) {
+ __do_free char *line = NULL;
+ __do_fclose FILE *f = NULL;
+ size_t linelen = 0;
+
+ f = fopen("/proc/meminfo", "r");
+ if (!f)
+ return 0;
+
+ while (getline(&line, &linelen, f) != -1) {
+ if (startswith(line, "SwapTotal:"))
+ sscanf(line, "SwapTotal: %8lu kB", &swap_total);
+ else if (startswith(line, "SwapFree:"))
+ sscanf(line, "SwapFree: %8lu kB", &swap_free);
+ }
+ }
+
+ if (swap_total > 0) {
+ l = snprintf(d->buf + total_len, d->size - total_len,
+ "none%*svirtual\t\t%lu\t%lu\t0\n", 36, " ",
+ swap_total, swap_free);
+ total_len += l;
+ }
+
+ if (total_len < 0 || l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+
+ d->cached = 1;
+ d->size = (int)total_len;
+
+ if (total_len > size) total_len = size;
+ memcpy(buf, d->buf, total_len);
+ return total_len;
+}
+
+static void get_blkio_io_value(char *str, unsigned major, unsigned minor,
+ char *iotype, unsigned long *v)
+{
+ char *eol;
+ char key[32];
+
+ memset(key, 0, 32);
+ snprintf(key, 32, "%u:%u %s", major, minor, iotype);
+
+ size_t len = strlen(key);
+ *v = 0;
+
+ while (*str) {
+ if (startswith(str, key)) {
+ sscanf(str + len, "%lu", v);
+ return;
+ }
+ eol = strchr(str, '\n');
+ if (!eol)
+ return;
+ str = eol+1;
+ }
+}
+
+static int proc_diskstats_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ __do_free char *cg = NULL, *io_serviced_str = NULL,
+ *io_merged_str = NULL, *io_service_bytes_str = NULL,
+ *io_wait_time_str = NULL, *io_service_time_str = NULL,
+ *line = NULL;
+ __do_fclose FILE *f = NULL;
+ struct fuse_context *fc = fuse_get_context();
+ struct file_info *d = (struct file_info *)fi->fh;
+ unsigned long read = 0, write = 0;
+ unsigned long read_merged = 0, write_merged = 0;
+ unsigned long read_sectors = 0, write_sectors = 0;
+ unsigned long read_ticks = 0, write_ticks = 0;
+ unsigned long ios_pgr = 0, tot_ticks = 0, rq_ticks = 0;
+ unsigned long rd_svctm = 0, wr_svctm = 0, rd_wait = 0, wr_wait = 0;
+ char *cache = d->buf;
+ size_t cache_size = d->buflen;
+ size_t linelen = 0, total_len = 0;
+ unsigned int major = 0, minor = 0;
+ int i = 0;
+ int ret;
+ char dev_name[72];
+
+ if (offset){
+ int left;
+
+ if (offset > d->size)
+ return -EINVAL;
+
+ if (!d->cached)
+ return 0;
+
+ left = d->size - offset;
+ total_len = left > size ? size: left;
+ memcpy(buf, cache + offset, total_len);
+
+ return total_len;
+ }
+
+ pid_t initpid = lookup_initpid_in_store(fc->pid);
+ if (initpid <= 1 || is_shared_pidns(initpid))
+ initpid = fc->pid;
+ cg = get_pid_cgroup(initpid, "blkio");
+ if (!cg)
+ return read_file_fuse("/proc/diskstats", buf, size, d);
+ prune_init_slice(cg);
+
+ ret = cgroup_ops->get_io_serviced(cgroup_ops, cg, &io_serviced_str);
+ if (ret < 0) {
+ if (ret == -EOPNOTSUPP)
+ return read_file_fuse("/proc/diskstats", buf, size, d);
+ }
+
+ ret = cgroup_ops->get_io_merged(cgroup_ops, cg, &io_merged_str);
+ if (ret < 0) {
+ if (ret == -EOPNOTSUPP)
+ return read_file_fuse("/proc/diskstats", buf, size, d);
+ }
+
+ ret = cgroup_ops->get_io_service_bytes(cgroup_ops, cg, &io_service_bytes_str);
+ if (ret < 0) {
+ if (ret == -EOPNOTSUPP)
+ return read_file_fuse("/proc/diskstats", buf, size, d);
+ }
+
+ ret = cgroup_ops->get_io_wait_time(cgroup_ops, cg, &io_wait_time_str);
+ if (ret < 0) {
+ if (ret == -EOPNOTSUPP)
+ return read_file_fuse("/proc/diskstats", buf, size, d);
+ }
+
+ ret = cgroup_ops->get_io_service_time(cgroup_ops, cg, &io_service_time_str);
+ if (ret < 0) {
+ if (ret == -EOPNOTSUPP)
+ return read_file_fuse("/proc/diskstats", buf, size, d);
+ }
+
+ f = fopen("/proc/diskstats", "r");
+ if (!f)
+ return 0;
+
+ while (getline(&line, &linelen, f) != -1) {
+ ssize_t l;
+ char lbuf[256];
+
+ i = sscanf(line, "%u %u %71s", &major, &minor, dev_name);
+ if (i != 3)
+ continue;
+
+ get_blkio_io_value(io_serviced_str, major, minor, "Read", &read);
+ get_blkio_io_value(io_serviced_str, major, minor, "Write", &write);
+ get_blkio_io_value(io_merged_str, major, minor, "Read", &read_merged);
+ get_blkio_io_value(io_merged_str, major, minor, "Write", &write_merged);
+ get_blkio_io_value(io_service_bytes_str, major, minor, "Read", &read_sectors);
+ read_sectors = read_sectors/512;
+ get_blkio_io_value(io_service_bytes_str, major, minor, "Write", &write_sectors);
+ write_sectors = write_sectors/512;
+
+ get_blkio_io_value(io_service_time_str, major, minor, "Read", &rd_svctm);
+ rd_svctm = rd_svctm/1000000;
+ get_blkio_io_value(io_wait_time_str, major, minor, "Read", &rd_wait);
+ rd_wait = rd_wait/1000000;
+ read_ticks = rd_svctm + rd_wait;
+
+ get_blkio_io_value(io_service_time_str, major, minor, "Write", &wr_svctm);
+ wr_svctm = wr_svctm/1000000;
+ get_blkio_io_value(io_wait_time_str, major, minor, "Write", &wr_wait);
+ wr_wait = wr_wait/1000000;
+ write_ticks = wr_svctm + wr_wait;
+
+ get_blkio_io_value(io_service_time_str, major, minor, "Total", &tot_ticks);
+ tot_ticks = tot_ticks/1000000;
+
+ memset(lbuf, 0, 256);
+ if (read || write || read_merged || write_merged || read_sectors || write_sectors || read_ticks || write_ticks)
+ snprintf(lbuf, 256, "%u %u %s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu\n",
+ major, minor, dev_name, read, read_merged, read_sectors, read_ticks,
+ write, write_merged, write_sectors, write_ticks, ios_pgr, tot_ticks, rq_ticks);
+ else
+ continue;
+
+ l = snprintf(cache, cache_size, "%s", lbuf);
+ if (l < 0) {
+ perror("Error writing to fuse buf");
+ return 0;
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+ }
+
+ d->cached = 1;
+ d->size = total_len;
+ if (total_len > size ) total_len = size;
+ memcpy(buf, d->buf, total_len);
+
+ return total_len;
+}
+
+#if RELOADTEST
+void iwashere(void)
+{
+ int fd;
+
+ fd = creat("/tmp/lxcfs-iwashere", 0644);
+ if (fd >= 0)
+ close(fd);
+}
+#endif
+
+/* This function retrieves the busy time of a group of tasks by looking at
+ * cpuacct.usage. Unfortunately, this only makes sense when the container has
+ * been given it's own cpuacct cgroup. If not, this function will take the busy
+ * time of all other taks that do not actually belong to the container into
+ * account as well. If someone has a clever solution for this please send a
+ * patch!
+ */
+static double get_reaper_busy(pid_t task)
+{
+ __do_free char *cgroup = NULL, *usage_str = NULL;
+ unsigned long usage = 0;
+ pid_t initpid;
+
+ initpid = lookup_initpid_in_store(task);
+ if (initpid <= 0)
+ return 0;
+
+ cgroup = get_pid_cgroup(initpid, "cpuacct");
+ if (!cgroup)
+ return 0;
+ prune_init_slice(cgroup);
+ if (!cgroup_ops->get(cgroup_ops, "cpuacct", cgroup, "cpuacct.usage",
+ &usage_str))
+ return 0;
+
+ usage = strtoul(usage_str, NULL, 10);
+ return ((double)usage / 1000000000);
+}
+
+static uint64_t get_reaper_start_time(pid_t pid)
+{
+ int ret;
+ FILE *f;
+ uint64_t starttime;
+ /* strlen("/proc/") = 6
+ * +
+ * LXCFS_NUMSTRLEN64
+ * +
+ * strlen("/stat") = 5
+ * +
+ * \0 = 1
+ * */
+#define __PROC_PID_STAT_LEN (6 + LXCFS_NUMSTRLEN64 + 5 + 1)
+ char path[__PROC_PID_STAT_LEN];
+ pid_t qpid;
+
+ qpid = lookup_initpid_in_store(pid);
+ if (qpid <= 0) {
+ /* Caller can check for EINVAL on 0. */
+ errno = EINVAL;
+ return 0;
+ }
+
+ ret = snprintf(path, __PROC_PID_STAT_LEN, "/proc/%d/stat", qpid);
+ if (ret < 0 || ret >= __PROC_PID_STAT_LEN) {
+ /* Caller can check for EINVAL on 0. */
+ errno = EINVAL;
+ return 0;
+ }
+
+ f = fopen(path, "r");
+ if (!f) {
+ /* Caller can check for EINVAL on 0. */
+ errno = EINVAL;
+ return 0;
+ }
+
+ /* Note that the *scanf() argument supression requires that length
+ * modifiers such as "l" are omitted. Otherwise some compilers will yell
+ * at us. It's like telling someone you're not married and then asking
+ * if you can bring your wife to the party.
+ */
+ ret = fscanf(f, "%*d " /* (1) pid %d */
+ "%*s " /* (2) comm %s */
+ "%*c " /* (3) state %c */
+ "%*d " /* (4) ppid %d */
+ "%*d " /* (5) pgrp %d */
+ "%*d " /* (6) session %d */
+ "%*d " /* (7) tty_nr %d */
+ "%*d " /* (8) tpgid %d */
+ "%*u " /* (9) flags %u */
+ "%*u " /* (10) minflt %lu */
+ "%*u " /* (11) cminflt %lu */
+ "%*u " /* (12) majflt %lu */
+ "%*u " /* (13) cmajflt %lu */
+ "%*u " /* (14) utime %lu */
+ "%*u " /* (15) stime %lu */
+ "%*d " /* (16) cutime %ld */
+ "%*d " /* (17) cstime %ld */
+ "%*d " /* (18) priority %ld */
+ "%*d " /* (19) nice %ld */
+ "%*d " /* (20) num_threads %ld */
+ "%*d " /* (21) itrealvalue %ld */
+ "%" PRIu64, /* (22) starttime %llu */
+ &starttime);
+ if (ret != 1) {
+ fclose(f);
+ /* Caller can check for EINVAL on 0. */
+ errno = EINVAL;
+ return 0;
+ }
+
+ fclose(f);
+
+ errno = 0;
+ return starttime;
+}
+
+static double get_reaper_start_time_in_sec(pid_t pid)
+{
+ uint64_t clockticks, ticks_per_sec;
+ int64_t ret;
+ double res = 0;
+
+ clockticks = get_reaper_start_time(pid);
+ if (clockticks == 0 && errno == EINVAL) {
+ lxcfs_debug("failed to retrieve start time of pid %d\n", pid);
+ return 0;
+ }
+
+ ret = sysconf(_SC_CLK_TCK);
+ if (ret < 0 && errno == EINVAL) {
+ lxcfs_debug(
+ "%s\n",
+ "failed to determine number of clock ticks in a second");
+ return 0;
+ }
+
+ ticks_per_sec = (uint64_t)ret;
+ res = (double)clockticks / ticks_per_sec;
+ return res;
+}
+
+static double get_reaper_age(pid_t pid)
+{
+ uint64_t uptime_ms;
+ double procstart, procage;
+
+ /* We need to substract the time the process has started since system
+ * boot minus the time when the system has started to get the actual
+ * reaper age.
+ */
+ procstart = get_reaper_start_time_in_sec(pid);
+ procage = procstart;
+ if (procstart > 0) {
+ int ret;
+ struct timespec spec;
+
+ ret = clock_gettime(CLOCK_BOOTTIME, &spec);
+ if (ret < 0)
+ return 0;
+
+ /* We could make this more precise here by using the tv_nsec
+ * field in the timespec struct and convert it to milliseconds
+ * and then create a double for the seconds and milliseconds but
+ * that seems more work than it is worth.
+ */
+ uptime_ms = (spec.tv_sec * 1000) + (spec.tv_nsec * 1e-6);
+ procage = (uptime_ms - (procstart * 1000)) / 1000;
+ }
+
+ return procage;
+}
+
+/*
+ * We read /proc/uptime and reuse its second field.
+ * For the first field, we use the mtime for the reaper for
+ * the calling pid as returned by getreaperage
+ */
+static int proc_uptime_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ struct fuse_context *fc = fuse_get_context();
+ struct file_info *d = (struct file_info *)fi->fh;
+ double busytime = get_reaper_busy(fc->pid);
+ char *cache = d->buf;
+ ssize_t total_len = 0;
+ double idletime, reaperage;
+
+#if RELOADTEST
+ iwashere();
+#endif
+
+ if (offset){
+ if (!d->cached)
+ return 0;
+ if (offset > d->size)
+ return -EINVAL;
+ int left = d->size - offset;
+ total_len = left > size ? size: left;
+ memcpy(buf, cache + offset, total_len);
+ return total_len;
+ }
+
+ reaperage = get_reaper_age(fc->pid);
+ /* To understand why this is done, please read the comment to the
+ * get_reaper_busy() function.
+ */
+ idletime = reaperage;
+ if (reaperage >= busytime)
+ idletime = reaperage - busytime;
+
+ total_len = snprintf(d->buf, d->buflen, "%.2lf %.2lf\n", reaperage, idletime);
+ if (total_len < 0 || total_len >= d->buflen){
+ lxcfs_error("%s\n", "failed to write to cache");
+ return 0;
+ }
+
+ d->size = (int)total_len;
+ d->cached = 1;
+
+ if (total_len > size) total_len = size;
+
+ memcpy(buf, d->buf, total_len);
+ return total_len;
+}
+
+#define CPUALL_MAX_SIZE (BUF_RESERVE_SIZE / 2)
+static int proc_stat_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ __do_free char *cg = NULL, *cpuset = NULL, *line = NULL;
+ __do_free struct cpuacct_usage *cg_cpu_usage = NULL;
+ __do_fclose FILE *f = NULL;
+ struct fuse_context *fc = fuse_get_context();
+ struct file_info *d = (struct file_info *)fi->fh;
+ size_t linelen = 0, total_len = 0;
+ int curcpu = -1; /* cpu numbering starts at 0 */
+ int physcpu = 0;
+ unsigned long user = 0, nice = 0, system = 0, idle = 0, iowait = 0,
+ irq = 0, softirq = 0, steal = 0, guest = 0, guest_nice = 0;
+ unsigned long user_sum = 0, nice_sum = 0, system_sum = 0, idle_sum = 0,
+ iowait_sum = 0, irq_sum = 0, softirq_sum = 0,
+ steal_sum = 0, guest_sum = 0, guest_nice_sum = 0;
+ char cpuall[CPUALL_MAX_SIZE];
+ /* reserve for cpu all */
+ char *cache = d->buf + CPUALL_MAX_SIZE;
+ size_t cache_size = d->buflen - CPUALL_MAX_SIZE;
+ int cg_cpu_usage_size = 0;
+
+ if (offset){
+ if (offset > d->size)
+ return -EINVAL;
+ if (!d->cached)
+ return 0;
+ int left = d->size - offset;
+ total_len = left > size ? size: left;
+ memcpy(buf, d->buf + offset, total_len);
+ return total_len;
+ }
+
+ pid_t initpid = lookup_initpid_in_store(fc->pid);
+ lxcfs_v("initpid: %d\n", initpid);
+ if (initpid <= 0)
+ initpid = fc->pid;
+
+ /*
+ * when container run with host pid namespace initpid == 1, cgroup will "/"
+ * we should return host os's /proc contents.
+ * in some case cpuacct_usage.all in "/" will larger then /proc/stat
+ */
+ if (initpid == 1) {
+ return read_file_fuse("/proc/stat", buf, size, d);
+ }
+
+ cg = get_pid_cgroup(initpid, "cpuset");
+ lxcfs_v("cg: %s\n", cg);
+ if (!cg)
+ return read_file_fuse("/proc/stat", buf, size, d);
+ prune_init_slice(cg);
+
+ cpuset = get_cpuset(cg);
+ if (!cpuset)
+ return 0;
+
+ /*
+ * Read cpuacct.usage_all for all CPUs.
+ * If the cpuacct cgroup is present, it is used to calculate the container's
+ * CPU usage. If not, values from the host's /proc/stat are used.
+ */
+ if (read_cpuacct_usage_all(cg, cpuset, &cg_cpu_usage, &cg_cpu_usage_size) != 0) {
+ lxcfs_v("%s\n", "proc_stat_read failed to read from cpuacct, "
+ "falling back to the host's /proc/stat");
+ }
+
+ f = fopen("/proc/stat", "r");
+ if (!f)
+ return 0;
+
+ //skip first line
+ if (getline(&line, &linelen, f) < 0) {
+ lxcfs_error("%s\n", "proc_stat_read read first line failed.");
+ return 0;
+ }
+
+ if (cgroup_ops->can_use_cpuview(cgroup_ops) && cg_cpu_usage) {
+ total_len = cpuview_proc_stat(cg, cpuset, cg_cpu_usage, cg_cpu_usage_size,
+ f, d->buf, d->buflen);
+ goto out;
+ }
+
+ while (getline(&line, &linelen, f) != -1) {
+ ssize_t l;
+ char cpu_char[10]; /* That's a lot of cores */
+ char *c;
+ uint64_t all_used, cg_used, new_idle;
+ int ret;
+
+ if (strlen(line) == 0)
+ continue;
+ if (sscanf(line, "cpu%9[^ ]", cpu_char) != 1) {
+ /* not a ^cpuN line containing a number N, just print it */
+ l = snprintf(cache, cache_size, "%s", line);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+ continue;
+ }
+
+ if (sscanf(cpu_char, "%d", &physcpu) != 1)
+ continue;
+ if (!cpu_in_cpuset(physcpu, cpuset))
+ continue;
+ curcpu ++;
+
+ ret = sscanf(line, "%*s %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu",
+ &user,
+ &nice,
+ &system,
+ &idle,
+ &iowait,
+ &irq,
+ &softirq,
+ &steal,
+ &guest,
+ &guest_nice);
+
+ if (ret != 10 || !cg_cpu_usage) {
+ c = strchr(line, ' ');
+ if (!c)
+ continue;
+ l = snprintf(cache, cache_size, "cpu%d%s", curcpu, c);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+
+ if (ret != 10)
+ continue;
+ }
+
+ if (cg_cpu_usage) {
+ if (physcpu >= cg_cpu_usage_size)
+ break;
+
+ all_used = user + nice + system + iowait + irq + softirq + steal + guest + guest_nice;
+ cg_used = cg_cpu_usage[physcpu].user + cg_cpu_usage[physcpu].system;
+
+ if (all_used >= cg_used) {
+ new_idle = idle + (all_used - cg_used);
+
+ } else {
+ lxcfs_error("cpu%d from %s has unexpected cpu time: %lu in /proc/stat, "
+ "%lu in cpuacct.usage_all; unable to determine idle time\n",
+ curcpu, cg, all_used, cg_used);
+ new_idle = idle;
+ }
+
+ l = snprintf(cache, cache_size, "cpu%d %lu 0 %lu %lu 0 0 0 0 0 0\n",
+ curcpu, cg_cpu_usage[physcpu].user, cg_cpu_usage[physcpu].system,
+ new_idle);
+
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+
+ user_sum += cg_cpu_usage[physcpu].user;
+ system_sum += cg_cpu_usage[physcpu].system;
+ idle_sum += new_idle;
+
+ } else {
+ user_sum += user;
+ nice_sum += nice;
+ system_sum += system;
+ idle_sum += idle;
+ iowait_sum += iowait;
+ irq_sum += irq;
+ softirq_sum += softirq;
+ steal_sum += steal;
+ guest_sum += guest;
+ guest_nice_sum += guest_nice;
+ }
+ }
+
+ cache = d->buf;
+
+ int cpuall_len = snprintf(cpuall, CPUALL_MAX_SIZE, "cpu %lu %lu %lu %lu %lu %lu %lu %lu %lu %lu\n",
+ user_sum,
+ nice_sum,
+ system_sum,
+ idle_sum,
+ iowait_sum,
+ irq_sum,
+ softirq_sum,
+ steal_sum,
+ guest_sum,
+ guest_nice_sum);
+ if (cpuall_len > 0 && cpuall_len < CPUALL_MAX_SIZE) {
+ memcpy(cache, cpuall, cpuall_len);
+ cache += cpuall_len;
+ } else {
+ /* shouldn't happen */
+ lxcfs_error("proc_stat_read copy cpuall failed, cpuall_len=%d.", cpuall_len);
+ cpuall_len = 0;
+ }
+
+ memmove(cache, d->buf + CPUALL_MAX_SIZE, total_len);
+ total_len += cpuall_len;
+
+out:
+ d->cached = 1;
+ d->size = total_len;
+ if (total_len > size)
+ total_len = size;
+
+ memcpy(buf, d->buf, total_len);
+ return total_len;
+}
+
+/* Note that "memory.stat" in cgroup2 is hierarchical by default. */
+static void parse_memstat(int version,
+ char *memstat,
+ unsigned long *cached,
+ unsigned long *active_anon,
+ unsigned long *inactive_anon,
+ unsigned long *active_file,
+ unsigned long *inactive_file,
+ unsigned long *unevictable,
+ unsigned long *shmem)
+{
+ char *eol;
+
+ while (*memstat) {
+ if (startswith(memstat, is_unified_controller(version)
+ ? "cache"
+ : "total_cache")) {
+ sscanf(memstat + 11, "%lu", cached);
+ *cached /= 1024;
+ } else if (startswith(memstat, is_unified_controller(version)
+ ? "active_anon"
+ : "total_active_anon")) {
+ sscanf(memstat + 17, "%lu", active_anon);
+ *active_anon /= 1024;
+ } else if (startswith(memstat, is_unified_controller(version)
+ ? "inactive_anon"
+ : "total_inactive_anon")) {
+ sscanf(memstat + 19, "%lu", inactive_anon);
+ *inactive_anon /= 1024;
+ } else if (startswith(memstat, is_unified_controller(version)
+ ? "active_file"
+ : "total_active_file")) {
+ sscanf(memstat + 17, "%lu", active_file);
+ *active_file /= 1024;
+ } else if (startswith(memstat, is_unified_controller(version)
+ ? "inactive_file"
+ : "total_inactive_file")) {
+ sscanf(memstat + 19, "%lu", inactive_file);
+ *inactive_file /= 1024;
+ } else if (startswith(memstat, is_unified_controller(version)
+ ? "unevictable"
+ : "total_unevictable")) {
+ sscanf(memstat + 17, "%lu", unevictable);
+ *unevictable /= 1024;
+ } else if (startswith(memstat, is_unified_controller(version)
+ ? "shmem"
+ : "total_shmem")) {
+ sscanf(memstat + 11, "%lu", shmem);
+ *shmem /= 1024;
+ }
+ eol = strchr(memstat, '\n');
+ if (!eol)
+ return;
+ memstat = eol+1;
+ }
+}
+
+static bool startswith(const char *line, const char *pref)
+{
+ if (strncmp(line, pref, strlen(pref)) == 0)
+ return true;
+ return false;
+}
+
+static int proc_meminfo_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ __do_free char *cgroup = NULL, *line = NULL,
+ *memusage_str = NULL, *memstat_str = NULL,
+ *memswlimit_str = NULL, *memswusage_str = NULL;
+ __do_fclose FILE *f = NULL;
+ struct fuse_context *fc = fuse_get_context();
+ struct lxcfs_opts *opts = (struct lxcfs_opts *) fuse_get_context()->private_data;
+ struct file_info *d = (struct file_info *)fi->fh;
+ unsigned long memlimit = 0, memusage = 0, memswlimit = 0,
+ memswusage = 0, cached = 0, hosttotal = 0, active_anon = 0,
+ inactive_anon = 0, active_file = 0, inactive_file = 0,
+ unevictable = 0, shmem = 0, hostswtotal = 0;
+ size_t linelen = 0, total_len = 0;
+ char *cache = d->buf;
+ size_t cache_size = d->buflen;
+ int ret;
+
+ if (offset) {
+ int left;
+
+ if (offset > d->size)
+ return -EINVAL;
+
+ if (!d->cached)
+ return 0;
+
+ left = d->size - offset;
+ total_len = left > size ? size : left;
+ memcpy(buf, cache + offset, total_len);
+
+ return total_len;
+ }
+
+ pid_t initpid = lookup_initpid_in_store(fc->pid);
+ if (initpid <= 1 || is_shared_pidns(initpid))
+ initpid = fc->pid;
+
+ cgroup = get_pid_cgroup(initpid, "memory");
+ if (!cgroup)
+ return read_file_fuse("/proc/meminfo", buf, size, d);
+
+ prune_init_slice(cgroup);
+
+ memlimit = get_min_memlimit(cgroup, false);
+
+ ret = cgroup_ops->get_memory_current(cgroup_ops, cgroup, &memusage_str);
+ if (ret < 0)
+ return 0;
+
+ ret = cgroup_ops->get_memory_stats(cgroup_ops, cgroup, &memstat_str);
+ if (ret < 0)
+ return 0;
+ parse_memstat(ret, memstat_str, &cached, &active_anon, &inactive_anon,
+ &active_file, &inactive_file, &unevictable, &shmem);
+
+ /*
+ * Following values are allowed to fail, because swapaccount might be
+ * turned off for current kernel.
+ */
+ ret = cgroup_ops->get_memory_swap_max(cgroup_ops, cgroup, &memswlimit_str);
+ if (ret >= 0)
+ ret = cgroup_ops->get_memory_swap_current(cgroup_ops, cgroup, &memswusage_str);
+ if (ret >= 0) {
+ memswlimit = get_min_memlimit(cgroup, true);
+ memswusage = strtoul(memswusage_str, NULL, 10);
+ memswlimit = memswlimit / 1024;
+ memswusage = memswusage / 1024;
+ }
+
+ memusage = strtoul(memusage_str, NULL, 10);
+ memlimit /= 1024;
+ memusage /= 1024;
+
+ f = fopen("/proc/meminfo", "r");
+ if (!f)
+ return 0;
+
+ while (getline(&line, &linelen, f) != -1) {
+ ssize_t l;
+ char *printme, lbuf[100];
+
+ memset(lbuf, 0, 100);
+ if (startswith(line, "MemTotal:")) {
+ sscanf(line+sizeof("MemTotal:")-1, "%lu", &hosttotal);
+ if (hosttotal < memlimit)
+ memlimit = hosttotal;
+ snprintf(lbuf, 100, "MemTotal: %8lu kB\n", memlimit);
+ printme = lbuf;
+ } else if (startswith(line, "MemFree:")) {
+ snprintf(lbuf, 100, "MemFree: %8lu kB\n", memlimit - memusage);
+ printme = lbuf;
+ } else if (startswith(line, "MemAvailable:")) {
+ snprintf(lbuf, 100, "MemAvailable: %8lu kB\n", memlimit - memusage + cached);
+ printme = lbuf;
+ } else if (startswith(line, "SwapTotal:") && memswlimit > 0 &&
+ opts && opts->swap_off == false) {
+ sscanf(line+sizeof("SwapTotal:")-1, "%lu", &hostswtotal);
+ if (hostswtotal < memswlimit)
+ memswlimit = hostswtotal;
+ snprintf(lbuf, 100, "SwapTotal: %8lu kB\n", memswlimit);
+ printme = lbuf;
+ } else if (startswith(line, "SwapTotal:") && opts && opts->swap_off == true) {
+ snprintf(lbuf, 100, "SwapTotal: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "SwapFree:") && memswlimit > 0 &&
+ memswusage > 0 && opts && opts->swap_off == false) {
+ unsigned long swaptotal = memswlimit,
+ swapusage = memusage > memswusage
+ ? 0
+ : memswusage - memusage,
+ swapfree = swapusage < swaptotal
+ ? swaptotal - swapusage
+ : 0;
+ snprintf(lbuf, 100, "SwapFree: %8lu kB\n", swapfree);
+ printme = lbuf;
+ } else if (startswith(line, "SwapFree:") && opts && opts->swap_off == true) {
+ snprintf(lbuf, 100, "SwapFree: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "Slab:")) {
+ snprintf(lbuf, 100, "Slab: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "Buffers:")) {
+ snprintf(lbuf, 100, "Buffers: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "Cached:")) {
+ snprintf(lbuf, 100, "Cached: %8lu kB\n", cached);
+ printme = lbuf;
+ } else if (startswith(line, "SwapCached:")) {
+ snprintf(lbuf, 100, "SwapCached: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "Active:")) {
+ snprintf(lbuf, 100, "Active: %8lu kB\n",
+ active_anon + active_file);
+ printme = lbuf;
+ } else if (startswith(line, "Inactive:")) {
+ snprintf(lbuf, 100, "Inactive: %8lu kB\n",
+ inactive_anon + inactive_file);
+ printme = lbuf;
+ } else if (startswith(line, "Active(anon)")) {
+ snprintf(lbuf, 100, "Active(anon): %8lu kB\n", active_anon);
+ printme = lbuf;
+ } else if (startswith(line, "Inactive(anon)")) {
+ snprintf(lbuf, 100, "Inactive(anon): %8lu kB\n", inactive_anon);
+ printme = lbuf;
+ } else if (startswith(line, "Active(file)")) {
+ snprintf(lbuf, 100, "Active(file): %8lu kB\n", active_file);
+ printme = lbuf;
+ } else if (startswith(line, "Inactive(file)")) {
+ snprintf(lbuf, 100, "Inactive(file): %8lu kB\n", inactive_file);
+ printme = lbuf;
+ } else if (startswith(line, "Unevictable")) {
+ snprintf(lbuf, 100, "Unevictable: %8lu kB\n", unevictable);
+ printme = lbuf;
+ } else if (startswith(line, "SReclaimable")) {
+ snprintf(lbuf, 100, "SReclaimable: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "SUnreclaim")) {
+ snprintf(lbuf, 100, "SUnreclaim: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "Shmem:")) {
+ snprintf(lbuf, 100, "Shmem: %8lu kB\n", shmem);
+ printme = lbuf;
+ } else if (startswith(line, "ShmemHugePages")) {
+ snprintf(lbuf, 100, "ShmemHugePages: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else if (startswith(line, "ShmemPmdMapped")) {
+ snprintf(lbuf, 100, "ShmemPmdMapped: %8lu kB\n", 0UL);
+ printme = lbuf;
+ } else
+ printme = line;
+
+ l = snprintf(cache, cache_size, "%s", printme);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+ }
+
+ d->cached = 1;
+ d->size = total_len;
+ if (total_len > size ) total_len = size;
+ memcpy(buf, d->buf, total_len);
+
+ return total_len;
+}
+
+int proc_read(const char *path, char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ struct file_info *f = (struct file_info *) fi->fh;
+
+ switch (f->type) {
+ case LXC_TYPE_PROC_MEMINFO:
+ return proc_meminfo_read(buf, size, offset, fi);
+ case LXC_TYPE_PROC_CPUINFO:
+ return proc_cpuinfo_read(buf, size, offset, fi);
+ case LXC_TYPE_PROC_UPTIME:
+ return proc_uptime_read(buf, size, offset, fi);
+ case LXC_TYPE_PROC_STAT:
+ return proc_stat_read(buf, size, offset, fi);
+ case LXC_TYPE_PROC_DISKSTATS:
+ return proc_diskstats_read(buf, size, offset, fi);
+ case LXC_TYPE_PROC_SWAPS:
+ return proc_swaps_read(buf, size, offset, fi);
+ case LXC_TYPE_PROC_LOADAVG:
+ return proc_loadavg_read(buf, size, offset, fi);
+ default:
+ return -EINVAL;
+ }
+}
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef __LXCFS_PROC_FUSE_H
+#define __LXCFS_PROC_FUSE_H
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <fuse.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "config.h"
+#include "macro.h"
+
+extern int proc_getattr(const char *path, struct stat *sb);
+extern int proc_readdir(const char *path, void *buf, fuse_fill_dir_t filler,
+ off_t offset, struct fuse_file_info *fi);
+extern int proc_open(const char *path, struct fuse_file_info *fi);
+extern int proc_access(const char *path, int mask);
+extern int proc_read(const char *path, char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi);
+extern int proc_release(const char *path, struct fuse_file_info *fi);
+
+#endif /* __LXCFS_PROC_FUSE_H */
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#define __STDC_FORMAT_MACROS
+#include <dirent.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <fuse.h>
+#include <inttypes.h>
+#include <libgen.h>
+#include <pthread.h>
+#include <sched.h>
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <time.h>
+#include <unistd.h>
+#include <wait.h>
+#include <linux/magic.h>
+#include <linux/sched.h>
+#include <sys/epoll.h>
+#include <sys/mman.h>
+#include <sys/mount.h>
+#include <sys/param.h>
+#include <sys/socket.h>
+#include <sys/syscall.h>
+#include <sys/sysinfo.h>
+#include <sys/vfs.h>
+
+#include "bindings.h"
+#include "config.h"
+#include "cgroup_fuse.h"
+#include "cgroups/cgroup.h"
+#include "cgroups/cgroup_utils.h"
+#include "memory_utils.h"
+#include "utils.h"
+
+/*
+ * This parameter is used for proc_loadavg_read().
+ * 1 means use loadavg, 0 means not use.
+ */
+static int loadavg = 0;
+
+/* The function of hash table.*/
+#define LOAD_SIZE 100 /*the size of hash_table */
+#define FLUSH_TIME 5 /*the flush rate */
+#define DEPTH_DIR 3 /*the depth of per cgroup */
+/* The function of calculate loadavg .*/
+#define FSHIFT 11 /* nr of bits of precision */
+#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
+#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
+#define EXP_5 2014 /* 1/exp(5sec/5min) */
+#define EXP_15 2037 /* 1/exp(5sec/15min) */
+#define LOAD_INT(x) ((x) >> FSHIFT)
+#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100)
+static volatile sig_atomic_t loadavg_stop = 0;
+
+struct load_node {
+ char *cg; /*cg */
+ unsigned long avenrun[3]; /* Load averages */
+ unsigned int run_pid;
+ unsigned int total_pid;
+ unsigned int last_pid;
+ int cfd; /* The file descriptor of the mounted cgroup */
+ struct load_node *next;
+ struct load_node **pre;
+};
+
+struct load_head {
+ /*
+ * The lock is about insert load_node and refresh load_node.To the first
+ * load_node of each hash bucket, insert and refresh in this hash bucket is
+ * mutually exclusive.
+ */
+ pthread_mutex_t lock;
+ /*
+ * The rdlock is about read loadavg and delete load_node.To each hash
+ * bucket, read and delete is mutually exclusive. But at the same time, we
+ * allow paratactic read operation. This rdlock is at list level.
+ */
+ pthread_rwlock_t rdlock;
+ /*
+ * The rilock is about read loadavg and insert load_node.To the first
+ * load_node of each hash bucket, read and insert is mutually exclusive.
+ * But at the same time, we allow paratactic read operation.
+ */
+ pthread_rwlock_t rilock;
+ struct load_node *next;
+};
+
+static struct load_head load_hash[LOAD_SIZE]; /* hash table */
+
+/*
+ * locate_node() finds special node. Not return NULL means success.
+ * It should be noted that rdlock isn't unlocked at the end of code
+ * because this function is used to read special node. Delete is not
+ * allowed before read has ended.
+ * unlock rdlock only in proc_loadavg_read().
+ */
+static struct load_node *locate_node(char *cg, int locate)
+{
+ struct load_node *f = NULL;
+ int i = 0;
+
+ pthread_rwlock_rdlock(&load_hash[locate].rilock);
+ pthread_rwlock_rdlock(&load_hash[locate].rdlock);
+ if (load_hash[locate].next == NULL) {
+ pthread_rwlock_unlock(&load_hash[locate].rilock);
+ return f;
+ }
+ f = load_hash[locate].next;
+ pthread_rwlock_unlock(&load_hash[locate].rilock);
+ while (f && ((i = strcmp(f->cg, cg)) != 0))
+ f = f->next;
+ return f;
+}
+
+static void insert_node(struct load_node **n, int locate)
+{
+ struct load_node *f;
+
+ pthread_mutex_lock(&load_hash[locate].lock);
+ pthread_rwlock_wrlock(&load_hash[locate].rilock);
+ f = load_hash[locate].next;
+ load_hash[locate].next = *n;
+
+ (*n)->pre = &(load_hash[locate].next);
+ if (f)
+ f->pre = &((*n)->next);
+ (*n)->next = f;
+ pthread_mutex_unlock(&load_hash[locate].lock);
+ pthread_rwlock_unlock(&load_hash[locate].rilock);
+}
+
+static int calc_hash(const char *name)
+{
+ unsigned int hash = 0;
+ unsigned int x = 0;
+ /* ELFHash algorithm. */
+ while (*name) {
+ hash = (hash << 4) + *name++;
+ x = hash & 0xf0000000;
+ if (x != 0)
+ hash ^= (x >> 24);
+ hash &= ~x;
+ }
+ return (hash & 0x7fffffff);
+}
+
+int proc_loadavg_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi)
+{
+ struct fuse_context *fc = fuse_get_context();
+ struct file_info *d = (struct file_info *)fi->fh;
+ pid_t initpid;
+ char *cg;
+ size_t total_len = 0;
+ char *cache = d->buf;
+ struct load_node *n;
+ int hash;
+ int cfd, rv = 0;
+ unsigned long a, b, c;
+
+ if (offset) {
+ int left;
+
+ if (offset > d->size)
+ return -EINVAL;
+
+ if (!d->cached)
+ return 0;
+
+ left = d->size - offset;
+ total_len = left > size ? size : left;
+ memcpy(buf, cache + offset, total_len);
+
+ return total_len;
+ }
+ if (!loadavg)
+ return read_file_fuse("/proc/loadavg", buf, size, d);
+
+ initpid = lookup_initpid_in_store(fc->pid);
+ if (initpid <= 1 || is_shared_pidns(initpid))
+ initpid = fc->pid;
+
+ cg = get_pid_cgroup(initpid, "cpu");
+ if (!cg)
+ return read_file_fuse("/proc/loadavg", buf, size, d);
+
+ prune_init_slice(cg);
+ hash = calc_hash(cg) % LOAD_SIZE;
+ n = locate_node(cg, hash);
+
+ /* First time */
+ if (n == NULL) {
+ cfd = get_cgroup_fd("cpu");
+ if (cfd >= 0) {
+ /*
+ * In locate_node() above, pthread_rwlock_unlock() isn't used
+ * because delete is not allowed before read has ended.
+ */
+ pthread_rwlock_unlock(&load_hash[hash].rdlock);
+ rv = 0;
+ goto err;
+ }
+ do {
+ n = malloc(sizeof(struct load_node));
+ } while (!n);
+
+ do {
+ n->cg = malloc(strlen(cg)+1);
+ } while (!n->cg);
+ strcpy(n->cg, cg);
+ n->avenrun[0] = 0;
+ n->avenrun[1] = 0;
+ n->avenrun[2] = 0;
+ n->run_pid = 0;
+ n->total_pid = 1;
+ n->last_pid = initpid;
+ n->cfd = cfd;
+ insert_node(&n, hash);
+ }
+ a = n->avenrun[0] + (FIXED_1/200);
+ b = n->avenrun[1] + (FIXED_1/200);
+ c = n->avenrun[2] + (FIXED_1/200);
+ total_len = snprintf(d->buf, d->buflen, "%lu.%02lu %lu.%02lu %lu.%02lu %d/%d %d\n",
+ LOAD_INT(a), LOAD_FRAC(a),
+ LOAD_INT(b), LOAD_FRAC(b),
+ LOAD_INT(c), LOAD_FRAC(c),
+ n->run_pid, n->total_pid, n->last_pid);
+ pthread_rwlock_unlock(&load_hash[hash].rdlock);
+ if (total_len < 0 || total_len >= d->buflen) {
+ lxcfs_error("%s\n", "Failed to write to cache");
+ rv = 0;
+ goto err;
+ }
+ d->size = (int)total_len;
+ d->cached = 1;
+
+ if (total_len > size)
+ total_len = size;
+ memcpy(buf, d->buf, total_len);
+ rv = total_len;
+
+err:
+ free(cg);
+ return rv;
+}
+
+/*
+ * Find the process pid from cgroup path.
+ * eg:from /sys/fs/cgroup/cpu/docker/containerid/cgroup.procs to find the process pid.
+ * @pid_buf : put pid to pid_buf.
+ * @dpath : the path of cgroup. eg: /docker/containerid or /docker/containerid/child-cgroup ...
+ * @depth : the depth of cgroup in container.
+ * @sum : return the number of pid.
+ * @cfd : the file descriptor of the mounted cgroup. eg: /sys/fs/cgroup/cpu
+ */
+static int calc_pid(char ***pid_buf, char *dpath, int depth, int sum, int cfd)
+{
+ __do_free char *path = NULL;
+ __do_close_prot_errno int fd = -EBADF;
+ __do_fclose FILE *f = NULL;
+ __do_closedir DIR *dir = NULL;
+ struct dirent *file;
+ size_t linelen = 0;
+ char *line = NULL;
+ int pd;
+ char **pid;
+
+ /* path = dpath + "/cgroup.procs" + /0 */
+ path = malloc(strlen(dpath) + 20);
+ if (!path)
+ return sum;
+
+ strcpy(path, dpath);
+ fd = openat(cfd, path, O_RDONLY | O_CLOEXEC | O_NOFOLLOW);
+ if (fd < 0)
+ return sum;
+
+ dir = fdopendir(move_fd(fd));
+ if (!dir)
+ return sum;
+
+ while (((file = readdir(dir)) != NULL) && depth > 0) {
+ if (strcmp(file->d_name, ".") == 0)
+ continue;
+
+ if (strcmp(file->d_name, "..") == 0)
+ continue;
+
+ if (file->d_type == DT_DIR) {
+ __do_free char *path_dir = NULL;
+
+ /* path + '/' + d_name +/0 */
+ path_dir = malloc(strlen(path) + 2 + sizeof(file->d_name));
+ if (!path_dir)
+ return sum;
+
+ strcpy(path_dir, path);
+ strcat(path_dir, "/");
+ strcat(path_dir, file->d_name);
+ pd = depth - 1;
+ sum = calc_pid(pid_buf, path_dir, pd, sum, cfd);
+ }
+ }
+
+ strcat(path, "/cgroup.procs");
+ fd = openat(cfd, path, O_RDONLY);
+ if (fd < 0)
+ return sum;
+
+ f = fdopen(move_fd(fd), "r");
+ if (!f)
+ return sum;
+
+ while (getline(&line, &linelen, f) != -1) {
+ pid = realloc(*pid_buf, sizeof(char *) * (sum + 1));
+ if (!pid)
+ return sum;
+ *pid_buf = pid;
+
+ *(*pid_buf + sum) = malloc(strlen(line) + 1);
+ if (!*(*pid_buf + sum))
+ return sum;
+
+ strcpy(*(*pid_buf + sum), line);
+ sum++;
+ }
+
+ return sum;
+}
+
+/*
+ * calc_load calculates the load according to the following formula:
+ * load1 = load0 * exp + active * (1 - exp)
+ *
+ * @load1: the new loadavg.
+ * @load0: the former loadavg.
+ * @active: the total number of running pid at this moment.
+ * @exp: the fixed-point defined in the beginning.
+ */
+static unsigned long calc_load(unsigned long load, unsigned long exp,
+ unsigned long active)
+{
+ unsigned long newload;
+
+ active = active > 0 ? active * FIXED_1 : 0;
+ newload = load * exp + active * (FIXED_1 - exp);
+ if (active >= load)
+ newload += FIXED_1 - 1;
+
+ return newload / FIXED_1;
+}
+
+/*
+ * Return 0 means that container p->cg is closed.
+ * Return -1 means that error occurred in refresh.
+ * Positive num equals the total number of pid.
+ */
+static int refresh_load(struct load_node *p, char *path)
+{
+ __do_free char *line = NULL;
+ char **idbuf;
+ char proc_path[256];
+ int i, ret, run_pid = 0, total_pid = 0, last_pid = 0;
+ size_t linelen = 0;
+ int sum, length;
+ struct dirent *file;
+
+ idbuf = malloc(sizeof(char *));
+ if (!idbuf)
+ return -1;
+
+ sum = calc_pid(&idbuf, path, DEPTH_DIR, 0, p->cfd);
+ /* normal exit */
+ if (sum == 0)
+ goto out;
+
+ for (i = 0; i < sum; i++) {
+ __do_closedir DIR *dp = NULL;
+
+ /*clean up '\n' */
+ length = strlen(idbuf[i])-1;
+ idbuf[i][length] = '\0';
+ ret = snprintf(proc_path, 256, "/proc/%s/task", idbuf[i]);
+ if (ret < 0 || ret > 255) {
+ lxcfs_error("%s\n", "snprintf() failed in refresh_load.");
+ i = sum;
+ sum = -1;
+ goto err_out;
+ }
+
+ dp = opendir(proc_path);
+ if (!dp) {
+ lxcfs_error("%s\n", "Open proc_path failed in refresh_load.");
+ continue;
+ }
+ while ((file = readdir(dp)) != NULL) {
+ __do_fclose FILE *f = NULL;
+
+ if (strncmp(file->d_name, ".", 1) == 0)
+ continue;
+ if (strncmp(file->d_name, "..", 1) == 0)
+ continue;
+ total_pid++;
+ /* We make the biggest pid become last_pid.*/
+ ret = atof(file->d_name);
+ last_pid = (ret > last_pid) ? ret : last_pid;
+
+ ret = snprintf(proc_path, 256, "/proc/%s/task/%s/status", idbuf[i], file->d_name);
+ if (ret < 0 || ret > 255) {
+ lxcfs_error("%s\n", "snprintf() failed in refresh_load.");
+ i = sum;
+ sum = -1;
+ goto err_out;
+ }
+
+ f = fopen(proc_path, "r");
+ if (f != NULL) {
+ while (getline(&line, &linelen, f) != -1) {
+ /* Find State */
+ if ((line[0] == 'S') && (line[1] == 't'))
+ break;
+ }
+
+ if ((line[7] == 'R') || (line[7] == 'D'))
+ run_pid++;
+ }
+ }
+ }
+ /*Calculate the loadavg.*/
+ p->avenrun[0] = calc_load(p->avenrun[0], EXP_1, run_pid);
+ p->avenrun[1] = calc_load(p->avenrun[1], EXP_5, run_pid);
+ p->avenrun[2] = calc_load(p->avenrun[2], EXP_15, run_pid);
+ p->run_pid = run_pid;
+ p->total_pid = total_pid;
+ p->last_pid = last_pid;
+
+err_out:
+ for (; i > 0; i--)
+ free(idbuf[i-1]);
+out:
+ free(idbuf);
+ return sum;
+}
+
+/* Delete the load_node n and return the next node of it. */
+static struct load_node *del_node(struct load_node *n, int locate)
+{
+ struct load_node *g;
+
+ pthread_rwlock_wrlock(&load_hash[locate].rdlock);
+ if (n->next == NULL) {
+ *(n->pre) = NULL;
+ } else {
+ *(n->pre) = n->next;
+ n->next->pre = n->pre;
+ }
+ g = n->next;
+ free_disarm(n->cg);
+ free_disarm(n);
+ pthread_rwlock_unlock(&load_hash[locate].rdlock);
+ return g;
+}
+
+/*
+ * Traverse the hash table and update it.
+ */
+static void *load_begin(void *arg)
+{
+
+ int i, sum, length, ret;
+ struct load_node *f;
+ int first_node;
+ clock_t time1, time2;
+
+ while (1) {
+ if (loadavg_stop == 1)
+ return NULL;
+
+ time1 = clock();
+ for (i = 0; i < LOAD_SIZE; i++) {
+ pthread_mutex_lock(&load_hash[i].lock);
+ if (load_hash[i].next == NULL) {
+ pthread_mutex_unlock(&load_hash[i].lock);
+ continue;
+ }
+ f = load_hash[i].next;
+ first_node = 1;
+ while (f) {
+ __do_free char *path = NULL;
+
+ length = strlen(f->cg) + 2;
+ /* strlen(f->cg) + '.' or '' + \0 */
+ path = malloc(length);
+ if (!path)
+ goto out;
+
+ ret = snprintf(path, length, "%s%s", dot_or_empty(f->cg), f->cg);
+ if (ret < 0 || ret > length - 1) {
+ /* snprintf failed, ignore the node.*/
+ lxcfs_error("Refresh node %s failed for snprintf().\n", f->cg);
+ goto out;
+ }
+
+ sum = refresh_load(f, path);
+ if (sum == 0)
+ f = del_node(f, i);
+ else
+out: f = f->next;
+ /* load_hash[i].lock locks only on the first node.*/
+ if (first_node == 1) {
+ first_node = 0;
+ pthread_mutex_unlock(&load_hash[i].lock);
+ }
+ }
+ }
+
+ if (loadavg_stop == 1)
+ return NULL;
+
+ time2 = clock();
+ usleep(FLUSH_TIME * 1000000 - (int)((time2 - time1) * 1000000 / CLOCKS_PER_SEC));
+ }
+}
+
+/*
+ * init_load initialize the hash table.
+ * Return 0 on success, return -1 on failure.
+ */
+static int init_load(void)
+{
+ int i;
+ int ret;
+
+ for (i = 0; i < LOAD_SIZE; i++) {
+ load_hash[i].next = NULL;
+ ret = pthread_mutex_init(&load_hash[i].lock, NULL);
+ if (ret != 0) {
+ lxcfs_error("%s\n", "Failed to initialize lock");
+ goto out3;
+ }
+ ret = pthread_rwlock_init(&load_hash[i].rdlock, NULL);
+ if (ret != 0) {
+ lxcfs_error("%s\n", "Failed to initialize rdlock");
+ goto out2;
+ }
+ ret = pthread_rwlock_init(&load_hash[i].rilock, NULL);
+ if (ret != 0) {
+ lxcfs_error("%s\n", "Failed to initialize rilock");
+ goto out1;
+ }
+ }
+ return 0;
+out1:
+ pthread_rwlock_destroy(&load_hash[i].rdlock);
+out2:
+ pthread_mutex_destroy(&load_hash[i].lock);
+out3:
+ while (i > 0) {
+ i--;
+ pthread_mutex_destroy(&load_hash[i].lock);
+ pthread_rwlock_destroy(&load_hash[i].rdlock);
+ pthread_rwlock_destroy(&load_hash[i].rilock);
+ }
+ return -1;
+}
+
+static void load_free(void)
+{
+ struct load_node *f, *p;
+
+ for (int i = 0; i < LOAD_SIZE; i++) {
+ pthread_mutex_lock(&load_hash[i].lock);
+ pthread_rwlock_wrlock(&load_hash[i].rilock);
+ pthread_rwlock_wrlock(&load_hash[i].rdlock);
+ if (load_hash[i].next == NULL) {
+ pthread_mutex_unlock(&load_hash[i].lock);
+ pthread_mutex_destroy(&load_hash[i].lock);
+ pthread_rwlock_unlock(&load_hash[i].rilock);
+ pthread_rwlock_destroy(&load_hash[i].rilock);
+ pthread_rwlock_unlock(&load_hash[i].rdlock);
+ pthread_rwlock_destroy(&load_hash[i].rdlock);
+ continue;
+ }
+
+ for (f = load_hash[i].next; f;) {
+ free_disarm(f->cg);
+ p = f->next;
+ free_disarm(f);
+ f = p;
+ }
+
+ pthread_mutex_unlock(&load_hash[i].lock);
+ pthread_mutex_destroy(&load_hash[i].lock);
+ pthread_rwlock_unlock(&load_hash[i].rilock);
+ pthread_rwlock_destroy(&load_hash[i].rilock);
+ pthread_rwlock_unlock(&load_hash[i].rdlock);
+ pthread_rwlock_destroy(&load_hash[i].rdlock);
+ }
+}
+
+/* Return a positive number on success, return 0 on failure.*/
+pthread_t load_daemon(int load_use)
+{
+ int ret;
+ pthread_t pid;
+
+ ret = init_load();
+ if (ret == -1) {
+ lxcfs_error("%s\n", "Initialize hash_table fails in load_daemon!");
+ return 0;
+ }
+ ret = pthread_create(&pid, NULL, load_begin, NULL);
+ if (ret != 0) {
+ lxcfs_error("%s\n", "Create pthread fails in load_daemon!");
+ load_free();
+ return 0;
+ }
+ /* use loadavg, here loadavg = 1*/
+ loadavg = load_use;
+ return pid;
+}
+
+/* Returns 0 on success. */
+int stop_load_daemon(pthread_t pid)
+{
+ int s;
+
+ /* Signal the thread to gracefully stop */
+ loadavg_stop = 1;
+
+ s = pthread_join(pid, NULL); /* Make sure sub thread has been canceled. */
+ if (s != 0) {
+ lxcfs_error("%s\n", "stop_load_daemon error: failed to join");
+ return -1;
+ }
+
+ load_free();
+ loadavg_stop = 0;
+
+ return 0;
+}
--- /dev/null
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef __LXCFS_PROC_LOADAVG_FUSE_H
+#define __LXCFS_PROC_LOADAVG_FUSE_H
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <fuse.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "config.h"
+#include "macro.h"
+
+extern int proc_loadavg_read(char *buf, size_t size, off_t offset,
+ struct fuse_file_info *fi);
+extern pthread_t load_daemon(int load_use);
+extern int stop_load_daemon(pthread_t pid);
+
+#endif /* __LXCFS_PROC_LOADAVG_FUSE_H */
+
/* SPDX-License-Identifier: LGPL-2.1-or-later */
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
#define __STDC_FORMAT_MACROS
#include <dirent.h>
#ifndef __LXCFS_SYSFS_FUSE_H
#define __LXCFS_SYSFS_FUSE_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <fuse.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "config.h"
+#include "macro.h"
+
extern int sys_getattr(const char *path, struct stat *sb);
extern int sys_readdir(const char *path, void *buf, fuse_fill_dir_t filler, off_t offset,
struct fuse_file_info *fi);
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <stdio.h>
#include <string.h>
#include <stdbool.h>
#include <stdlib.h>
-/*
- * cpusets are in format "1,2-3,4"
- * iow, comma-delimited ranges
- */
-extern bool cpu_in_cpuset(int cpu, const char *cpuset);
+#include "../cpuset_parse.h"
-void verify(bool condition) {
+static void verify(bool condition) {
if (condition) {
printf(" PASS\n");
} else {
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <stdio.h>
#include <unistd.h>
#include <sys/types.h>
fd = open(file, O_RDONLY|O_DIRECT);
else
fd = open(file, O_RDONLY);
-
+
while(i++ < read_count){
memset(buf, 0, BUFSIZE);
ret = read(fd, buf, BUFSIZE-1);
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
+#include <errno.h>
+#include <fcntl.h>
+#include <libgen.h>
+#include <stddef.h>
+#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
-#include <unistd.h>
+#include <sys/mount.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
#include <sys/sysmacros.h>
#include <sys/types.h>
-#include <sys/socket.h>
-#include <linux/un.h>
-#include <attr/xattr.h>
+#include <sys/types.h>
+#include <sys/un.h>
+#include <sys/xattr.h>
+#include <unistd.h>
#include <utime.h>
-#include <sys/stat.h>
-#include <sys/mount.h>
-#include <libgen.h>
-#include <fcntl.h>
-#include <errno.h>
-
void test_open(const char *path)
{
-#define __STDC_FORMAT_MACROS
+/* SPDX-License-Identifier: LGPL-2.1-or-later */
+
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
+#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
+
#include <errno.h>
#include <fcntl.h>
+#include <fuse.h>
#include <inttypes.h>
#include <sched.h>
#include <stdarg.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
+#include <sys/epoll.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>
+#include "bindings.h"
#include "config.h"
#include "macro.h"
#include "memory_utils.h"
free_disarm(f->buf);
free_disarm(f);
}
+
+#define POLLIN_SET ( EPOLLIN | EPOLLHUP | EPOLLRDHUP )
+
+bool wait_for_sock(int sock, int timeout)
+{
+ struct epoll_event ev;
+ int epfd, ret, now, starttime, deltatime, saved_errno;
+
+ if ((starttime = time(NULL)) < 0)
+ return false;
+
+ if ((epfd = epoll_create(1)) < 0) {
+ lxcfs_error("%s\n", "Failed to create epoll socket: %m.");
+ return false;
+ }
+
+ ev.events = POLLIN_SET;
+ ev.data.fd = sock;
+ if (epoll_ctl(epfd, EPOLL_CTL_ADD, sock, &ev) < 0) {
+ lxcfs_error("%s\n", "Failed adding socket to epoll: %m.");
+ close(epfd);
+ return false;
+ }
+
+again:
+ if ((now = time(NULL)) < 0) {
+ close(epfd);
+ return false;
+ }
+
+ deltatime = (starttime + timeout) - now;
+ if (deltatime < 0) { // timeout
+ errno = 0;
+ close(epfd);
+ return false;
+ }
+ ret = epoll_wait(epfd, &ev, 1, 1000*deltatime + 1);
+ if (ret < 0 && errno == EINTR)
+ goto again;
+ saved_errno = errno;
+ close(epfd);
+
+ if (ret <= 0) {
+ errno = saved_errno;
+ return false;
+ }
+ return true;
+}
+
+bool recv_creds(int sock, struct ucred *cred, char *v)
+{
+ struct msghdr msg = { 0 };
+ struct iovec iov;
+ struct cmsghdr *cmsg;
+ char cmsgbuf[CMSG_SPACE(sizeof(*cred))];
+ char buf[1];
+ int ret;
+ int optval = 1;
+
+ *v = '1';
+
+ cred->pid = -1;
+ cred->uid = -1;
+ cred->gid = -1;
+
+ if (setsockopt(sock, SOL_SOCKET, SO_PASSCRED, &optval, sizeof(optval)) == -1) {
+ lxcfs_error("Failed to set passcred: %s\n", strerror(errno));
+ return false;
+ }
+ buf[0] = '1';
+ if (write(sock, buf, 1) != 1) {
+ lxcfs_error("Failed to start write on scm fd: %s\n", strerror(errno));
+ return false;
+ }
+
+ msg.msg_name = NULL;
+ msg.msg_namelen = 0;
+ msg.msg_control = cmsgbuf;
+ msg.msg_controllen = sizeof(cmsgbuf);
+
+ iov.iov_base = buf;
+ iov.iov_len = sizeof(buf);
+ msg.msg_iov = &iov;
+ msg.msg_iovlen = 1;
+
+ if (!wait_for_sock(sock, 2)) {
+ lxcfs_error("Timed out waiting for scm_cred: %s\n", strerror(errno));
+ return false;
+ }
+ ret = recvmsg(sock, &msg, MSG_DONTWAIT);
+ if (ret < 0) {
+ lxcfs_error("Failed to receive scm_cred: %s\n", strerror(errno));
+ return false;
+ }
+
+ cmsg = CMSG_FIRSTHDR(&msg);
+
+ if (cmsg && cmsg->cmsg_len == CMSG_LEN(sizeof(struct ucred)) &&
+ cmsg->cmsg_level == SOL_SOCKET &&
+ cmsg->cmsg_type == SCM_CREDENTIALS) {
+ memcpy(cred, CMSG_DATA(cmsg), sizeof(*cred));
+ }
+ *v = buf[0];
+
+ return true;
+}
+
+static int msgrecv(int sockfd, void *buf, size_t len)
+{
+ if (!wait_for_sock(sockfd, 2))
+ return -1;
+ return recv(sockfd, buf, len, MSG_DONTWAIT);
+}
+
+int send_creds(int sock, struct ucred *cred, char v, bool pingfirst)
+{
+ struct msghdr msg = { 0 };
+ struct iovec iov;
+ struct cmsghdr *cmsg;
+ char cmsgbuf[CMSG_SPACE(sizeof(*cred))];
+ char buf[1];
+ buf[0] = 'p';
+
+ if (pingfirst) {
+ if (msgrecv(sock, buf, 1) != 1) {
+ lxcfs_error("%s\n", "Error getting reply from server over socketpair.");
+ return SEND_CREDS_FAIL;
+ }
+ }
+
+ msg.msg_control = cmsgbuf;
+ msg.msg_controllen = sizeof(cmsgbuf);
+
+ cmsg = CMSG_FIRSTHDR(&msg);
+ cmsg->cmsg_len = CMSG_LEN(sizeof(struct ucred));
+ cmsg->cmsg_level = SOL_SOCKET;
+ cmsg->cmsg_type = SCM_CREDENTIALS;
+ memcpy(CMSG_DATA(cmsg), cred, sizeof(*cred));
+
+ msg.msg_name = NULL;
+ msg.msg_namelen = 0;
+
+ buf[0] = v;
+ iov.iov_base = buf;
+ iov.iov_len = sizeof(buf);
+ msg.msg_iov = &iov;
+ msg.msg_iovlen = 1;
+
+ if (sendmsg(sock, &msg, 0) < 0) {
+ lxcfs_error("Failed at sendmsg: %s.\n",strerror(errno));
+ if (errno == 3)
+ return SEND_CREDS_NOTSK;
+ return SEND_CREDS_FAIL;
+ }
+
+ return SEND_CREDS_OK;
+}
+
+int read_file_fuse(const char *path, char *buf, size_t size, struct file_info *d)
+{
+ __do_free char *line = NULL;
+ __do_fclose FILE *f = NULL;
+ size_t linelen = 0, total_len = 0;
+ char *cache = d->buf;
+ size_t cache_size = d->buflen;
+
+ f = fopen(path, "r");
+ if (!f)
+ return 0;
+
+ while (getline(&line, &linelen, f) != -1) {
+ ssize_t l = snprintf(cache, cache_size, "%s", line);
+ if (l < 0) {
+ perror("Error writing to cache");
+ return 0;
+ }
+ if (l >= cache_size) {
+ lxcfs_error("%s\n", "Internal error: truncated write to cache.");
+ return 0;
+ }
+ cache += l;
+ cache_size -= l;
+ total_len += l;
+ }
+
+ d->size = total_len;
+ if (total_len > size)
+ total_len = size;
+
+ /* read from off 0 */
+ memcpy(buf, d->buf, total_len);
+
+ if (d->size > total_len)
+ d->cached = d->size - total_len;
+ return total_len;
+}
#ifndef __LXCFS_UTILS_H
#define __LXCFS_UTILS_H
+#ifndef _GNU_SOURCE
+#define _GNU_SOURCE
+#endif
+
+#ifndef FUSE_USE_VERSION
#define FUSE_USE_VERSION 26
+#endif
+
+#define _FILE_OFFSET_BITS 64
#include <fuse.h>
+#include <stdbool.h>
+#include <sys/socket.h>
+#include <sys/types.h>
+#include <sys/un.h>
+#include <unistd.h>
+
+#include "config.h"
+#include "macro.h"
/* Reserve buffer size to account for file size changes. */
#define BUF_RESERVE_SIZE 512
+#define SEND_CREDS_OK 0
+#define SEND_CREDS_NOTSK 1
+#define SEND_CREDS_FAIL 2
+
+struct file_info;
+
extern void must_strcat(char **src, size_t *sz, size_t *asz, const char *format, ...);
extern bool is_shared_pidns(pid_t pid);
extern int preserve_ns(const int pid, const char *ns);
extern void do_release_file_info(struct fuse_file_info *fi);
+extern bool recv_creds(int sock, struct ucred *cred, char *v);
+extern int send_creds(int sock, struct ucred *cred, char v, bool pingfirst);
+extern bool wait_for_sock(int sock, int timeout);
+extern int read_file_fuse(const char *path, char *buf, size_t size,
+ struct file_info *d);
#endif /* __LXCFS_UTILS_H */