qerror: add proper ErrorClass value for QERR_ macros
This commit replaces the place holder value for the ErrorClass
argument with a proper ErrorClass value for all QERR_ macros.
All current errors are mapped to GenericError, except for errors
CommandNotFound, DeviceEncrypted, DeviceNotActive, DeviceNotFound,
KVMMissingCap and MigrationExpected, which are maintained as they
are today.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
error, qerror: add ErrorClass argument to error functions
The new argument is added to functions qerror_report() and error_set().
It's stored in Error and QError. qerror_report_err() is also updated to
take care of it.
The QERR_ macros are changed to contain a place holder value for the
new argument, so that the value is used on all current calls to
qerror_report() and error_set() (and also to initialize qerror_table[]).
Next commit will update the QERR_ macros with a proper ErrorClass
value.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
Several block/ files are relying on qerror.h being provided by
qapi-types.h. Fix this, as a future commit will change qapi-types.h
not to provide qerror.h.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
Luiz Capitulino [Mon, 6 Aug 2012 19:26:47 +0000 (16:26 -0300)]
migration: don't rely on any QERR_SOCKET_*
Use the in_progress argument for QERR_SOCKET_CONNECT_IN_PROGRESS. The
other errors are handled the same by checking if the error is set and
then calling migrate_fd_error() if it's.
It's also necessary to change inet_connect_opts() not to set
QERR_SOCKET_CONNECT_IN_PROGRESS. This error is only used by
tcp_start_outgoing_migration() and not changing it along with the
usage of in_progress would break migration.
Furthermore this commit fixes a bug. Today, there's a spurious error
report when migration succeeds:
(qemu) migrate tcp:0:4444
migrate: Connection can not be completed immediately
(qemu)
After this commit no spurious error is reported anymore.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
Today, the error message is only constructed when it's used. This commit
changes that to construct the error message when the error object is
built (ie. when the error is reported).
This simplifies the Error object.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
Today, the error message is only constructed when it's used. This commit
changes qerror to construct the error message when the error object is
built (ie. when the error is reported).
This eliminates the need of storing a pointer to qerror_table[], which
will be dropped soon, and also simplifies the code.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
qerror_abort() depends on the 'file', 'func' and 'linenr' members of
QError. However, these members are going to be dropped by the next
commit, so let's drop qerror_abort() in favor of printing an error
message to stderr plus a call to abort().
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
In the old QMP days, this code was used to find out QMP commands that
might be calling monitor_printf() down its call chain.
This is almost impossible to happen today, because the qapi converted
commands don't even have a monitor object. Besides, it's been more than
a year since I used this last time.
Let's just drop it.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
Anthony Liguori [Fri, 10 Aug 2012 03:16:10 +0000 (13:16 +1000)]
qom: Reimplement Interfaces
The current implementation of Interfaces is poorly designed. Each interface
that an object implements ends up being an object that's tracked by the
implementing object. There's all sorts of gymnastics to deal with casting
between these objects.
But an interface shouldn't be associated with an Object. Interfaces are global
to a class. This patch moves all Interface knowledge to ObjectClass eliminating
the relationship between Object and Interfaces.
Interfaces are now abstract (as they should be) but this is okay. Interfaces
essentially act as additional parents for the classes and are treated as such.
With this new implementation, we should fully support derived interfaces
including reimplementing an inherited interface.
PC: Rebased against qom-next merge Jun-2012.
PC: Removed replication of cast logic for interfaces, i.e. there is only
one cast function - object_dynamic_cast() (and object_dynamic_cast_assert())
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com> Signed-off-by: Peter A. G. Crosthwaite <peter.crosthwaite@petalogix.com> Acked-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
Re-implemented the interconnect between the Xilinx AXI ethernet and DMA
controllers. A QOM interface "stream" is created, for the two stream interfaces.
As per Edgars request, this is designed to be more generic than AXI-stream,
so in the future we may see more clients of this interface beyond AXI stream.
This is based primarily on Paolos original refactoring of the interconnect.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Peter A.G. Crosthwaite <peter.crosthwaite@petalogix.com> Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
Anthony Liguori [Sun, 12 Aug 2012 00:49:03 +0000 (19:49 -0500)]
Merge remote-tracking branch 'stefanha/trivial-patches' into staging
* stefanha/trivial-patches:
target-arm: Fix typos in comments
arm: translate: comment typo - s/middel/middle/
vl.c: Exit QEMU early if no machine is found
Anthony Liguori [Sun, 12 Aug 2012 00:48:50 +0000 (19:48 -0500)]
Merge remote-tracking branch 'kwolf/for-anthony' into staging
* kwolf/for-anthony:
qemu-iotests: skip 039 with ./check -nocache
block: add BLOCK_O_CHECK for qemu-img check
qcow2: mark image clean after repair succeeds
qed: mark image clean after repair succeeds
blockdev: flip default cache mode from writethrough to writeback
virtio-blk: disable write cache if not negotiated
virtio-blk: support VIRTIO_BLK_F_CONFIG_WCE
qemu-iotests: Save some sed processes
ahci: Fix sglist memleak in ahci_dma_rw_buf()
ahci: Fix ahci cdrom read corruptions for reads > 128k
virtio-blk: fix use-after-free while handling scsi commands
Anthony Liguori [Sat, 11 Aug 2012 22:11:23 +0000 (17:11 -0500)]
Merge remote-tracking branch 'bonzini/scsi-next' into staging
* bonzini/scsi-next:
scsi-disk: add support for the UNMAP command
scsi-disk: improve out-of-range LBA detection for WRITE SAME
scsi-disk: more assertions and resets for aiocb
virtio-scsi: do not compare 32-bit QEMU tags against 64-bit virtio-scsi tags
iscsi: Pick default initiator-name based on the name of the VM
iscsi: reorganize code for parse_initiator_name
iscsi: do not leak initiator_name
Blue Swirl [Thu, 9 Aug 2012 20:21:25 +0000 (20:21 +0000)]
configure: fix double check tests with Clang
Configuring with Clang compiler with -Werror would not work after
improved checks:
/tmp/qemu-conf--25992-.c:4:32: error: self-comparison always evaluates
to true [-Werror,-Wtautological-compare]
int main(void) { return preadv == preadv; }
/tmp/qemu-conf--25992-.c:13:26: error: self-comparison always
evaluates to true [-Werror,-Wtautological-compare]
return epoll_create1 == epoll_create1;
/tmp/qemu-conf--25992-.c:3:13: error: explicitly assigning a variable
of type 'char **' to itself [-Werror,-Wself-assign]
environ = environ;
Avoid the errors by adjusting the tests.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Peter Maydell [Wed, 18 Jul 2012 14:10:28 +0000 (15:10 +0100)]
configure: Check for -Werror causing failures when compiling tests
Add support for checking whether test case code can compile without
warnings, by recompiling each successful test with -Werror. If the
-Werror version doesn't pass, we bail out. This gives us the same
level of visibility of warnings in test code as --enable-werror
provides for the main compile.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Igor Mitsyanko [Fri, 10 Aug 2012 14:45:11 +0000 (18:45 +0400)]
exec.c: fix dirty bitmap reallocation
For each newly created RAM block, dirty bitmap is reallocated with g_realloc, which doesn't
make any promises on initial content of new extra data in returned buffer. In theory,
we initialize this new data with cpu_physical_memory_set_dirty_range() call. The
problem is, cpu_physical_memory_set_dirty_range() has a side effect of incrementing
ram_list.dirty_pages variable, but only for pages which are not already dirty. And
page "cleanliness" is determined using the same not yet uninitialized dirty bitmap
we've just reallocated. This results in inconsistency between real dirty page number
and value in ram_list.dirty_pages variable, which in turn could (and will) result
in errors during VM migration.
Zero initialize new dirty bitmap bytes to fix this problem.
Signed-off-by: Igor Mitsyanko <i.mitsyanko@samsung.com> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Stefan Hajnoczi [Fri, 10 Aug 2012 13:08:42 +0000 (14:08 +0100)]
Makefile: add qapi.py dependencies
Commit 427a1a2cb1d35b83b6302886f46289f6d617134d ("qapi: avoid reserved
keywords") modifies qapi.py, which is used by qapi-types.py and other
Python scripts. Because Makefile has no dependencies for qapi.py the
qapi code generator will not be rerun and the following build error is
produced:
net/slirp.c: In function ‘net_init_slirp’:
net/slirp.c:721:50: error: ‘NetdevUserOptions’ has no member named ‘q_restrict’
Fix this issue by adding the missing qapi.py dependencies.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Guan Xuetao [Fri, 10 Aug 2012 06:42:39 +0000 (14:42 +0800)]
unicore32-softmmu: Add a minimal curses screen support
This patch adds a minimal curses screen support for unicore32-softmmu.
We assume 80*30 screen size to minimize the implementation.
Two problems are not solved, but they are innocuous.
1. curses windows will be blank when switching to monitor screen and back
2. backspace is not handled yet
Guan Xuetao [Fri, 10 Aug 2012 06:42:31 +0000 (14:42 +0800)]
unicore32-softmmu: Add puv3 pm support
This patch adds puv3 pm (power management) support,
include pm device simulation for kernel booting.
Thank Blue Swirl for pointing out the missing "break".
v1->v2: Add initialization to ret in puv3_pm_read.
Signed-off-by: Guan Xuetao <gxt@mprc.pku.edu.cn> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Guan Xuetao [Fri, 10 Aug 2012 06:42:25 +0000 (14:42 +0800)]
unicore32-softmmu: Implement softmmu specific functions
This patch implements softmmu specific functions, include tlb_fill,
switch_mode, do_interrupt and uc32_cpu_handle_mmu_fault.
So the full exception handlers and page table walking could work now.
Signed-off-by: Guan Xuetao <gxt@mprc.pku.edu.cn> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Guan Xuetao [Fri, 10 Aug 2012 06:42:23 +0000 (14:42 +0800)]
unicore32-softmmu: Make UniCore32 cpuid & exceptions correct and runable
This patch initializes the cpuid to exactly correct value because
linux kernel will check it.
In addition, the exception types are specified in proper situations.
Then it could make exceptions generated correctly and timely.
Signed-off-by: Guan Xuetao <gxt@mprc.pku.edu.cn> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Guan Xuetao [Fri, 10 Aug 2012 06:42:22 +0000 (14:42 +0800)]
unicore32-softmmu: Add coprocessor 0(sysctrl) and 1(ocd) instruction support
Coprocessor 0 is system control coprocessor, and we need get/set its contents.
Also, all cache/tlb ops shoule be implemented here, but just ignored with no harm.
Coprocessor 1 is OCD (on-chip-debugger), which is used for faked console,
so we could output chars to this console without graphic card.
TODO: curses display should be added lator for screen output.
Signed-off-by: Guan Xuetao <gxt@mprc.pku.edu.cn> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Guan Xuetao [Fri, 10 Aug 2012 06:42:21 +0000 (14:42 +0800)]
unicore32-softmmu: Add unicore32-softmmu build support
This patch adds unicore32-softmmu build support, include configure,
makefile, arch_init, and all missing functions needed by softmmu.
Although all missing functions are empty, unicore32-softmmu could
be build successfully.
By 20120804: change QEMU_ARCH_UNICORE32 to 0x4000
Signed-off-by: Guan Xuetao <gxt@mprc.pku.edu.cn> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Peter Maydell [Mon, 6 Aug 2012 16:42:18 +0000 (17:42 +0100)]
target-arm: Fix typos in comments
Fix a variety of typos in comments in target-arm files.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org> Reviewed-by: Peter Crosthwaite <peter.crosthwaite@petalogix.com> Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Peter A. G. Crosthwaite <peter.crosthwaite@petalogix.com> Reviewed-by: Andreas Färber <afaerber@suse.de> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
We check whether the variable machine is NULL or not before accessing
it. If machine is NULL, exit QEMU with an error, this can avoids a
segfault error.
Markus Armbruster <armbru@redhat.com> adds that the segfault can be
reproduced as follows:
$ qemu-system-xtensa -cpu help
Signed-off-by: Dunrong Huang <riegamaths@gmail.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Stefan Hajnoczi [Thu, 9 Aug 2012 12:05:56 +0000 (13:05 +0100)]
block: add BLOCK_O_CHECK for qemu-img check
Image formats with a dirty bit, like qed and qcow2, repair dirty image
files upon open with BDRV_O_RDWR. Performing automatic repair when
qemu-img check runs is not ideal because the bdrv_open() call repairs
the image before the actual bdrv_check() call from qemu-img.c.
Fix this "double repair" since it leads to confusing output from
qemu-img check. Tell the block driver that this image is being opened
just for bdrv_check(). This skips automatic repair and qemu-img.c can
invoke it manually with bdrv_check().
Update the golden output for qemu-iotests 039 to reflect the new
qemu-img check output.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Stefan Hajnoczi [Thu, 9 Aug 2012 12:05:55 +0000 (13:05 +0100)]
qcow2: mark image clean after repair succeeds
The dirty bit is cleared after image repair succeeds in qcow2_open().
Move this into qcow2_check() so that all callers benefit from this
behavior when fix mode is enabled.
This is necessary so qemu-img check can call .bdrv_check() and mark the
image clean.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Stefan Hajnoczi [Thu, 9 Aug 2012 12:05:54 +0000 (13:05 +0100)]
qed: mark image clean after repair succeeds
The dirty bit is cleared after image repair succeeds in qed_open().
Move this into qed_check() so that all callers benefit from this
behavior when fix=true.
This is necessary so qemu-img check can call .bdrv_check() and mark the
image clean.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Paolo Bonzini [Thu, 9 Aug 2012 14:07:21 +0000 (16:07 +0200)]
blockdev: flip default cache mode from writethrough to writeback
Now all major device models (IDE, SCSI, virtio) can choose between
writethrough and writeback at run-time, and virtio will even revert
to writethrough if the guest is not capable of sending flushes. So
we can change the default to writeback at last.
Tested, for lack of a better idea, with a breakpoint on bdrv_open
and all cache choices one by one.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Paolo Bonzini [Thu, 9 Aug 2012 14:07:20 +0000 (16:07 +0200)]
virtio-blk: disable write cache if not negotiated
If the guest does not support flushes, we should run in writethrough mode.
The setting is temporary until the next reset, so that for example the
BIOS will run in writethrough mode while Linux will run with a writeback
cache.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Jason Baron [Fri, 3 Aug 2012 19:57:10 +0000 (15:57 -0400)]
ahci: Fix sglist memleak in ahci_dma_rw_buf()
I noticed that in hw/ide/ahci:ahci_dma_rw_buf() we do not free the sglist. Thus,
I've added a call to qemu_sglist_destroy() to fix this memory leak.
In addition, I've adeed a call in qemu_sglist_destroy() to 0 all of the sglist
fields, in case there is some other codepath that tries to free the sglist.
Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Jason Baron [Fri, 3 Aug 2012 19:57:06 +0000 (15:57 -0400)]
ahci: Fix ahci cdrom read corruptions for reads > 128k
While testing q35, which has its cdrom attached to the ahci controller, I found
that the Fedora 17 install would panic on boot. The panic occurs while
squashfs is trying to read from the cdrom. The errors are:
[ 8.622711] SQUASHFS error: xz_dec_run error, data probably corrupt
[ 8.625180] SQUASHFS error: squashfs_read_data failed to read block
0x20be48a
I was also able to produce corrupt data reads using an installed piix based
qemu machine, using 'dd'. I found that the corruptions were only occuring when
then read size was greater than 128k. For example, the following command
results in corrupted reads:
dd if=/dev/sr0 of=/tmp/blah bs=256k iflag=direct
The > 128k size reads exercise a different code path than 128k and below. In
ide_atapi_cmd_read_dma_cb() s->io_buffer_size is capped at 128k. Thus,
ide_atapi_cmd_read_dma_cb() is called a second time when the read is > 128k.
However, ahci_dma_rw_buf() restart the read from offset 0, instead of at 128k.
Thus, resulting in a corrupted read.
To fix this, I've introduced 'io_buffer_offset' field in IDEState to keep
track of the offset. I've also modified ahci_populate_sglist() to take a new
3rd offset argument, so that the sglist is property initialized.
I've tested this patch using 'dd' testing, and Fedora 17 now correctly boots
and installs on q35 with the cdrom ahci controller.
Signed-off-by: Jason Baron <jbaron@redhat.com> Tested-by: Andreas Färber <afaerber@suse.de> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Avi Kivity [Mon, 6 Aug 2012 12:49:03 +0000 (15:49 +0300)]
virtio-blk: fix use-after-free while handling scsi commands
The scsi passthrough handler falls through after completing a
request into the failure path, resulting in a use after free.
Reproducible by running a guest with aio=native on a block device.
Reported-by: Stefan Priebe <s.priebe@profihost.ag> Signed-off-by: Avi Kivity <avi@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Bruce Rogers [Thu, 9 Aug 2012 18:47:40 +0000 (12:47 -0600)]
handle device help before accelerator set up
A command line device probe using just -device "?" gets processed
after qemu-kvm initializes the accelerator. If /dev/kvm is not
present, the accelerator check will fail (kvm is defaulted to on),
which causes libvirt to not be set up to handle qemu guests.
Moving the device help handling before the accelerator set up allows
the device probe to work in this configuration and libvirt succeeds
in setting up for a qemu hypervisor mode.
Signed-off-by: Bruce Rogers <brogers@suse.com> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Blue Swirl [Thu, 9 Aug 2012 18:44:49 +0000 (18:44 +0000)]
Merge branch 'x86cpu_qom_tcg_v2' of git://github.com/imammedo/qemu
* 'x86cpu_qom_tcg_v2' of git://github.com/imammedo/qemu:
target-i386: move tcg initialization into x86_cpu_initfn()
cleanup cpu_set_debug_excp_handler
target-xtensa: drop usage of prev_debug_excp_handler
target-i386: drop usage of prev_debug_excp_handler
qemu_rearm_alarm_timer: do not call rearm if the next deadline is INT64_MAX
qemu_rearm_alarm_timer partially duplicates the code in
qemu_next_alarm_deadline to figure out if it needs to rearm the timer.
If it calls qemu_next_alarm_deadline, it always rearms the timer even if
the next deadline is INT64_MAX.
This patch simplifies the behavior of qemu_rearm_alarm_timer and removes
the duplicated code, always calling qemu_next_alarm_deadline and only
rearming the timer if the deadline is less than INT64_MAX.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Reviewed-by: Stefan Weil <sw@weilnetz.de> Tested-by: Andreas Färber <andreas.faerber@web.de> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Nathan Froyd [Thu, 7 Jun 2012 15:04:14 +0000 (15:04 +0000)]
MIPS: Correct FCR0 initialization
This change addresses a problem where QEMU incorrectly traps on
floating-point MADD group instructions with SIGILL, at least while
emulating MIPS32r2 processors. These instructions use the COP1X major
opcode and include ones like:
madd.d $f2,$f4,$f2,$f6
Here's Nathan's original analysis of the problem:
"QEMU essentially does:
d = find_cpu (cpu_string) // get CPU definition
fpu_init (env, d) // initialize fpu state (init FCR0, basically)
cpu_reset (env)
...and the cpu_reset call clears all interesting state that fpu_init
setup, then proceeds to reinitialize all the CP0 registers...but not
FCR0."
I have verified this change with system emulation running the GDB test
suite for the mips-sde-elf target (o32, big endian, 24Kf CPU emulated),
there were 55 progressions and no regressions.
Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com> Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Blue Swirl [Mon, 30 Jul 2012 15:24:23 +0000 (15:24 +0000)]
sparc: fix floppy TC line setup
The qemu_irq for Terminal Count (TC) line between FDC and Slavio misc
device was created only after use, spotted by Clang compiler. Also,
it was not created if the FDC didn't exist.
Rearrange code to fix order. Always create the TC line.
Difference with AMD PCscsi is that DC-390 contains a EEPROM,
and that a romfile is available to add INT13 support.
This has been successfully tested on:
- MS DOS 6.22 (using DC390 ASPI driver)
- MS Windows 98 SE (using DC390 driver)
- MS Windows NT 3.1 (using DC390 driver)
- MS Windows NT 4.0 (using DC390 driver)
- hard disk and cdrom boot
Signed-off-by: Hervé Poussineau <hpoussin@reactos.org> Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Paolo Bonzini [Thu, 9 Aug 2012 09:00:22 +0000 (11:00 +0200)]
scsi-disk: add support for the UNMAP command
The unmap command can reuse the same infrastructure as MODE SELECT
for reading the descriptor list into memory. The descriptors are
processed sequentially.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Thu, 9 Aug 2012 09:33:26 +0000 (11:33 +0200)]
scsi-disk: more assertions and resets for aiocb
Leaving the aiocb to a non-NULL value leads to an assertion failure when
rerror/werror are set to stop or enospc, and the operation is retried.
scsi-disk checks that the aiocb member is NULL before filling it.
This patch correctly resets the aiocb to NULL values everywhere,
and adds the dual assertion that the aiocb was non-NULL before
calling bdrv_acct_done.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Paolo Bonzini [Wed, 8 Aug 2012 14:26:16 +0000 (16:26 +0200)]
virtio-scsi: do not compare 32-bit QEMU tags against 64-bit virtio-scsi tags
This patch fixes a problem in handling task management functions
in virtio-scsi. The cause of the problem is a mismatch between
the size of the tag in QEMU (32-bit) and virtio-scsi (64-bit).
Changing the QEMU size is hard because the migration format
uses 32 bits to store the tag; so just don't use the QEMU tag
(virtio-scsi only uses the tag for task management functions
anyway) and look up the full 64-bit tag in the hba_private field.
The reproducer is a bit obscure. If you cause an I/O timeout
(for example with rerror=stop and doing 'cont' on the monitor
continuously without fixing the error), sooner or later the
guest will try to abort the command and reissue it. At this
point, QEMU will report _two_ errors instead of one when you
hit 'c', because the first error has not been canceled correctly.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Ronnie Sahlberg [Mon, 6 Aug 2012 08:24:55 +0000 (18:24 +1000)]
iscsi: Pick default initiator-name based on the name of the VM
This patch updates the iscsi layer to automatically pick a 'unique'
initiator-name based on the name of the vm in case the user has not set
an explicit iqn-name to use.
Create a new function qemu_get_vm_name() that returns the name of the VM,
if specified.
This way we can thus create default names to use as the initiator name
based on the guest session.
If the VM is not named via the '-name' command line argument, the iscsi
initiator-name used wiull simply be
iqn.2008-11.org.linux-kvm
If a name for the VM was specified with the '-name' option, iscsi will
use a default initiatorname of
iqn.2008-11.org.linux-kvm:<name>
These names are just the default iscsi initiator name that qemu will
generate/use only when the user has not set an explicit initiator name
via the commandlines or config files.
Signed-off-by: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Anthony Liguori [Tue, 7 Aug 2012 14:46:24 +0000 (09:46 -0500)]
Merge remote-tracking branch 'kwolf/for-anthony' into staging
* kwolf/for-anthony:
qemu-img: use QemuOpts instead of QEMUOptionParameter in resize function
qemu-iotests: Be more flexible with image creation options
qemu-iotests: add 039 qcow2 lazy refcounts test
qemu-io: add "abort" command to simulate program crash
qcow2: implement lazy refcounts
qemu-iotests: ignore qemu-img create lazy_refcounts output
docs: add lazy refcounts bit to qcow2 specification
qcow2: introduce dirty bit
docs: add dirty bit to qcow2 specification
qemu-iotests: add qed.py image manipulation utility
qapi: generalize documentation of streaming commands
ide scsi: Mess with geometry only for hard disk devices
Commit 59310659073d85745854f2f10c4292555c5a1c51 is incomplete,
we'll arrive in the scsi command complete callback in CSW state
and must handle that case correctly.
Anthony Liguori [Tue, 7 Aug 2012 00:31:55 +0000 (19:31 -0500)]
slirp: fix build on mingw32
in_addr_t isn't available on mingw32. Just use an unsigned long instead. I
considered typedef'ing in_addr_t on mingw32 but this would potentially be
brittle if mingw32 did introduce the type.
Cc: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Dong Xu Wang [Mon, 6 Aug 2012 02:18:42 +0000 (10:18 +0800)]
qemu-img: use QemuOpts instead of QEMUOptionParameter in resize function
Signed-off-by: Dong Xu Wang <wdongxu@linux.vnet.ibm.com> Reviewed-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 27 Jul 2012 13:14:28 +0000 (15:14 +0200)]
qemu-iotests: Be more flexible with image creation options
qemu-iotests already filters out image creation options that may be
present or not in order to get the same output in both cases. However,
often it only considers the default value of the option. Cover all valid
values instead so that ./check -o name=value can be used successfull for
all of them.
Stefan Hajnoczi [Fri, 27 Jul 2012 08:05:24 +0000 (09:05 +0100)]
qemu-iotests: add 039 qcow2 lazy refcounts test
This tests establishes the basic post-conditions of the qcow2 lazy
refcounts features:
1. If the image was closed normally, it is marked clean.
2. If an allocating write was performed and the image was not closed
normally, then it is marked dirty.
a. Written data can be read back successfully.
b. The image file can be repaired and will be marked clean again.
c. The image file is automatically repaired when opened read/write.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Stefan Hajnoczi [Fri, 27 Jul 2012 08:05:23 +0000 (09:05 +0100)]
qemu-io: add "abort" command to simulate program crash
Avoiding data loss and corruption is the top requirement for image file
formats. The qemu-io "abort" command makes it possible to simulate
program crashes and does not give the image format a chance to cleanly
shut down. This command is useful for data integrity test cases.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Stefan Hajnoczi [Fri, 27 Jul 2012 08:05:22 +0000 (09:05 +0100)]
qcow2: implement lazy refcounts
Lazy refcounts is a performance optimization for qcow2 that postpones
refcount metadata updates and instead marks the image dirty. In the
case of crash or power failure the image will be left in a dirty state
and repaired next time it is opened.
Reducing metadata I/O is important for cache=writethrough and
cache=directsync because these modes guarantee that data is on disk
after each write (hence we cannot take advantage of caching updates in
RAM). Refcount metadata is not needed for guest->file block address
translation and therefore does not need to be on-disk at the time of
write completion - this is the motivation behind the lazy refcount
optimization.
The lazy refcount optimization must be enabled at image creation time:
projects / qemu.git / log