Anthony Liguori [Mon, 30 Sep 2013 22:13:32 +0000 (17:13 -0500)]
Merge remote-tracking branch 'kraxel/audio.1' into staging
# By Bandan Das (3) and Gerd Hoffmann (1)
# Via Gerd Hoffmann
* kraxel/audio.1:
audio: remove CONFIG_MIXEMU configure option
hda-codec: make mixemu selectable at runtime
hda-codec: refactor common definitions into a header file
audio maintainers update
Anthony Liguori [Mon, 30 Sep 2013 22:13:18 +0000 (17:13 -0500)]
Merge remote-tracking branch 'borntraeger/tags/s390-next-20130924' into staging
This is a bunch of fixes/changes for the s390 architecture. It also
contains the fixes from the previous pull request, which did not make
it yet.
Overall it contains
- a fix for kexec without kdump (which uses diag308 subcode 0 instead of 1)
- several sclp related fixes
- some initial sclp migration code
- the sclp line mode console
- A fix for a boot problem with the virtio ccw ipl bios
- zeroed out padding bytes for the notes section of dump-guest-memory
- some cleanups
# gpg: Signature made Tue 24 Sep 2013 02:18:44 AM CDT using RSA key ID B5A61C7C
# gpg: Can't check signature: public key not found
# By Christian Borntraeger (6) and others
# Via Christian Borntraeger
* borntraeger/tags/s390-next-20130924:
s390/sclplmconsole: Add support for SCLP line-mode console
s390/ebcdic: Move conversion tables to header file
s390/eventfacility: allow childs to handle more than 1 event type
s390/eventfacility: remove unused event_type variable
s390/eventfacility: Fix receive/send masks
s390/eventfacility: fix multiple Read Event Data sources
s390/sclp: add reset() functions
s390/sclpquiesce: Add code to support live migration
s390/sclpconsole: Add code to support live migration for sclpconsole
s390/sclpconsole: modify definition of input buffer
s390/kexec: Implement diag308 subcode 0
s390/ioinst: Moved the CC setting to the IO instruction handlers
s390/cpu: Make setcc() function available to other files
s390/ipl: Update the s390-ccw.img rom
s390/ipl: Fix waiting for virtio processing
s390/dump: zero out padding bytes in notes sections
s390/kvm: Add check for priviledged SCLP handler
Bandan Das [Sat, 7 Sep 2013 04:53:59 +0000 (00:53 -0400)]
hda-codec: make mixemu selectable at runtime
Define PARAM so that we have two versions of the "desc_codec
and family" structs. Add a property called "mixer" whose default
value depends on whether CONFIG_MIXEMU is defined or not which
will help us call the appropriate instance init functions.
Signed-off-by: Bandan Das <bsd@redhat.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Bandan Das [Sat, 7 Sep 2013 04:53:58 +0000 (00:53 -0400)]
hda-codec: refactor common definitions into a header file
Move common defines and structs to a header file.
The next commit will include it twice, once for a device with a
mixer, and once for device without a mixer.
Signed-off-by: Bandan Das <bsd@redhat.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
av1474@comtv.ru bounces, and I havn't seen malc @ qemu-devel for quite a
while (anyone knows what is up?). Adding myself as audio maintainer, so
audio patches don't fall through the cracks that easily.
With nptl enabled, atomic_cmpxchg_32 and atomic_barrier
system calls are needed. This patch enabled really dummy
versions of the system calls, modeled after the m68k
kernel code.
With this patch I am able to execute m68k binaries
with qemu linux-user (busybox compiled for coldfire).
[v2] que an segfault instead of returning a EFAULT
to keep in line with kernel code.
linux-user: Check type of microMIPS break instruction
microMIPS instructions that cause breakpoint exceptions come in
16-bit and 32-bit variants. When handling exceptions caused by
such instructions, the instruction type needs to be taken into
account when extracting the break code.
The code has also been restructured for better clarity.
Laurent Vivier [Thu, 29 Aug 2013 23:46:44 +0000 (01:46 +0200)]
linux-user: add support of binfmt_misc 'O' flag
The binfmt_misc module can calculate the credentials and security
token according to the binary instead of to the interpreter if the
'C' flag is enabled.
To be able to execute non-readable binaries, this flag implies 'O'
flag. When 'O' flag is enabled, bintfmt_misc opens the file for
reading and pass the file descriptor to the interpreter.
It has been pointed out on LKML that the alpha umount syscall numbers
are named wrong, and a patch to rectify that has been posted for 3.11.
Glibc works around this by treating NR_umount as NR_umount2 if
NR_oldumount exists. That's more complicated than we need in QEMU,
given that we control linux-user/*/syscall_nr.h.
This is the last instance of TARGET_NR_oldumount, so delete that from
the strace.list.
Signed-off-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Anthony Liguori [Mon, 23 Sep 2013 16:53:11 +0000 (11:53 -0500)]
Merge remote-tracking branch 'stefanha/net' into staging
# By Aurelien Jarno (1) and Vincenzo Maffione (1)
# Via Stefan Hajnoczi
* stefanha/net:
e1000: NetClientInfo.receive_iov implemented
pcnet-pci: mark I/O and MMIO as LITTLE_ENDIAN
Anthony Liguori [Mon, 23 Sep 2013 16:53:05 +0000 (11:53 -0500)]
Merge remote-tracking branch 'stefanha/block' into staging
# By Stefan Hajnoczi (4) and others
# Via Stefan Hajnoczi
* stefanha/block:
virtio-blk: do not relay a previous driver's WCE configuration to the current
blockdev: do not default cache.no-flush to true
block: don't lose data from last incomplete sector
qcow2: Correct snapshots size for overlap check
coroutine: fix /perf/nesting coroutine benchmark
coroutine: add qemu_coroutine_yield benchmark
qemu-timer: do not take the lock in timer_pending
qemu-timer: make qemu_timer_mod_ns() and qemu_timer_del() thread-safe
qemu-timer: drop outdated signal safety comments
osdep: warn if open(O_DIRECT) on fails with EINVAL
libcacard: link against qemu-error.o for error_report()
Anthony Liguori [Mon, 23 Sep 2013 16:52:55 +0000 (11:52 -0500)]
Merge remote-tracking branch 'mjt/trivial-patches' into staging
# By Stefan Weil (8) and others
# Via Michael Tokarev
* mjt/trivial-patches:
tests/.gitignore: ignore test-throttle
exec: Fix broken build for MinGW (regression)
kvm: Fix compiler warning (clang)
tcg-sparc: Fix parenthesis warning
Makefile: Remove some more files when cleaning
target-i386: Fix segment cache dump
iov: avoid "orig_len may be used unitialized" warning
vscclient: remove unnecessary use of uninitialized variable
trace-events: Clean up with scripts/cleanup-trace-events.pl again
tci: Fix qemu-alpha on 32 bit hosts (wrong assertions)
*-user: Improve documentation for lock_user function
MAINTAINERS: Add missing entry to filelist for TCI target
translate-all: Fix formatting of dump output
*-user: Fix typo in comment (ulocking -> unlocking)
docs: Fix IO port number for CPU present bitmap.
q35: Fix typo in constant DEFUALT -> DEFAULT.
configure: Undefine _FORTIFY_SOURCE prior using it
Anthony Liguori [Mon, 23 Sep 2013 16:52:49 +0000 (11:52 -0500)]
Merge remote-tracking branch 'qemu-kvm/uq/master' into staging
# By Alexey Kardashevskiy (3) and others
# Via Paolo Bonzini
* qemu-kvm/uq/master:
target-i386: add feature kvm_pv_unhalt
linux-headers: update to 3.12-rc1
target-i386: forward CPUID cache leaves when -cpu host is used
linux-headers: update to 3.11
kvm: fix traces to use %x instead of %d
kvmvapic: Clear also physical ROM address when entering INACTIVE state
kvmvapic: Enter inactive state on hardware reset
kvmvapic: Catch invalid ROM size
kvm irqfd: support direct msimessage to irq translation
fix steal time MSR vmsd callback to proper opaque type
kvm: warn if num cpus is greater than num recommended
cpu: Move cpu state syncs up into cpu_dump_state()
exec: always use MADV_DONTFORK
Anthony Liguori [Mon, 23 Sep 2013 16:52:32 +0000 (11:52 -0500)]
Merge remote-tracking branch 'bonzini/scsi-next' into staging
# By Hervé Poussineau (5) and Stefan Weil (1)
# Via Paolo Bonzini
* bonzini/scsi-next:
block/iscsi: Drop iscsi_co_get_block_status for older versions of libiscsi
lsi: add 53C810 variant
lsi: remove todo
lsi: ignore write accesses to CTEST0 registers
lsi: check ssid versus sdid only if ssid is valid
lsi: use constant name instead of its value
This patch implements the NetClientInfo.receive_iov method for the
e1000 device emulation. In this way a network backend that uses
qemu_sendv_packet() can deliver the fragmented packet without
requiring an additional copy in the frontend/backend network code
(nc_sendv_compat() function).
The existing method NetClientInfo.receive has been reimplemented
using the new method.
Signed-off-by: Vincenzo Maffione <v.maffione@gmail.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Aurelien Jarno [Wed, 28 Aug 2013 12:17:39 +0000 (14:17 +0200)]
pcnet-pci: mark I/O and MMIO as LITTLE_ENDIAN
Now that the memory subsystem is propagating the endianness correctly,
the pcnet-pci device should have its I/O ports and MMIO memory marked
as LITTLE_ENDIAN, as PCI devices are little endian.
This makes the pcnet-pci NIC to work again on big endian MIPS Malta
(default NIC).
Cc: qemu-stable@nongnu.org Signed-off-by: Aurelien Jarno <aurelien@aurel32.net> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Paolo Bonzini [Fri, 20 Sep 2013 15:31:55 +0000 (17:31 +0200)]
virtio-blk: do not relay a previous driver's WCE configuration to the current
The following sequence happens:
- the SeaBIOS virtio-blk driver does not support the WCE feature, which
causes QEMU to disable writeback caching
- the Linux virtio-blk driver resets the device, finds WCE is available
but writeback caching is disabled; tells block layer to not send cache
flush commands
- the Linux virtio-blk driver sets the DRIVER_OK bit, which causes
writeback caching to be re-enabled, but the Linux virtio-blk driver does
not know of this side effect and cache flushes remain disabled
The bug is at the third step. If the guest does know about CONFIG_WCE,
QEMU should ignore the WCE feature's state. The guest will control the
cache mode solely using configuration space. This change makes Linux
do flushes correctly, but Linux will keep SeaBIOS's writethrough mode.
Hence, whenever the guest is reset, the cache mode of the disk should
be reset to whatever was specified in the "-drive" option. With this
change, the Linux virtio-blk driver finds that writeback caching is
enabled, and tells the block layer to send cache flush commands
appropriately.
Reported-by: Rusty Russell <rusty@au1.ibm.com Cc: qemu-stable@nongnu.org Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
block: don't lose data from last incomplete sector
To read the last sector that is not aligned to sector boundary, current
code for growable backends, since commit 893a8f6 "block: Produce zeros
when protocols reading beyond end of file", drops the data and directly
returns zeroes. That is incorrect.
Signed-off-by: Fam Zheng <famz@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
exec.c: In function ‘qemu_ram_free’:
exec.c:1239:17: warning:
implicit declaration of function ‘munmap’ [-Wimplicit-function-declaration]
exec.c:1239:17: warning:
nested extern declaration of ‘munmap’ [-Wnested-externs]
exec.c:1239: undefined reference to `munmap'
Add some ifdeffery again to fix this.
Signed-off-by: Stefan Weil <sw@weilnetz.de> Reviewed-by: Markus Armbruster <armbru@redhat.com> Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Tobias Markus [Sun, 25 Aug 2013 10:20:06 +0000 (12:20 +0200)]
target-i386: Fix segment cache dump
When in Long Mode, cpu_x86_seg_cache() logs "DS16" because the Default
operation size bit (D/B bit) is not set for Long Mode Data Segments since
there are only Data Segments in Long Mode and no explicit 16/32/64-bit
Descriptors.
This patch fixes this by checking the Long Mode Active bit of the hidden
flags variable and logging "DS" if it is set. (I.e. in Long Mode all Data
Segments are logged as "DS")
Signed-off-by: Tobias Markus <tobias@markus-regensburg.de> Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Stefan Weil [Thu, 12 Sep 2013 18:17:50 +0000 (20:17 +0200)]
tci: Fix qemu-alpha on 32 bit hosts (wrong assertions)
Debian busybox-static for alpha has a load address of 0x0000000120000000
which is mapped to 0x0000000020000000 for 32 bit hosts.
qemu-alpha uses the TCG opcodes qemu_ld32, qemu_ld64, qemu_st32 and
qemu_st64 which all raise the assertion (taddr == host_addr).
Remove all assertions of this type because they are either wrong or
unnecessary (when sizeof(tcg_target_ulong) >= sizeof(target_ulong)).
Cc: qemu-stable <qemu-stable@nongnu.org> Signed-off-by: Stefan Weil <sw@weilnetz.de> Reviewed-by: Richard Henderson <rth@twiddle.net> Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Anthony PERARD [Tue, 10 Sep 2013 16:36:18 +0000 (17:36 +0100)]
docs: Fix IO port number for CPU present bitmap.
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com> Reviewd-By: Igor Mammedov <imammedo@redhat.com> Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
configure: Undefine _FORTIFY_SOURCE prior using it
Currently, we are enforcing the _FORTIFY_SOURCE=2 without any
previous detection if the macro has been already defined, e.g.
by environment, or is just enabled by compiler by default.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Jan Vesely <jano.vesely@gmail.com> Reviewed-by: Peter Maydell <peter.maydell@linaro.org> Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Anthony Liguori [Fri, 20 Sep 2013 13:08:08 +0000 (08:08 -0500)]
Merge remote-tracking branch 'kraxel/usb.90' into staging
# By Hans de Goede (6) and Gerd Hoffmann (1)
# Via Gerd Hoffmann
* kraxel/usb.90:
usb: Fix iovec memleak on combined-packet free
usb: Also reset max_packet_size on ep_reset
xhci: Fix memory leak on xhci_disable_ep
xhci: Add xhci_epid_to_usbep helper function
xhci: Init a transfers xhci, slotid and epid member on epctx alloc
xhci: Fix number of streams allocated when using streams
usb: remove old usb-host code
Heinz Graalfs [Wed, 22 May 2013 12:11:36 +0000 (14:11 +0200)]
s390/sclplmconsole: Add support for SCLP line-mode console
Add simple support for SCLP line-mode also known as operating
system messages. This can be added in addition to or instead of
the SCLP full screen console with -device sclplmconsole.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Reviewed-by: Alexander Graf <agraf@suse.de> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Heinz Graalfs [Tue, 21 May 2013 15:04:58 +0000 (17:04 +0200)]
s390/ebcdic: Move conversion tables to header file
Move conversion tables to header file.
- In SCLP line mode processing EBCDIC/ASCII conversion is needed.
- An additional EBCDIC to ASCII conversion function is added.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Reviewed-by: Alexander Graf <agraf@suse.de> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
s390/eventfacility: allow childs to handle more than 1 event type
Currently all handlers (quiesce, console) only handle one event type.
Some drivers will handle multiple (compatible) event types. Rework the
code accordingly.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Reviewed-by: Alexander Graf <agraf@suse.de>
Currently we announce interchanged receive/send masks. This did not
trigger a bug, since the sclp console has the same masks for
send/receive and the Linux guest does not check the sclp mask for simple
events like quiesce. With other event users like the sclp line mode
console, we will have different send/receive bits. Fix it.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Reviewed-by: Alexander Graf <agraf@suse.de>
Add reset() functions for event-facility, sclpconsole, and sclpquiesce.
The reset() functions perform variable initialization
at IPL and e.g. when monitor system_reset is called.
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Reviewed-by: Thomas Huth <thuth@linux.vnet.ibm.com> Reviewed-by: Alexander Graf <agraf@suse.de> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Heinz Graalfs [Wed, 28 Aug 2013 14:30:28 +0000 (16:30 +0200)]
s390/sclpconsole: modify definition of input buffer
To use VMState for migration, we need to adapt some sclp code:
- allocate console buffer as part of the console
- change semantic of sclpconsole offset fields
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com> Reviewed-by: Alexander Graf <agraf@suse.de> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
This patch implements subcode 0 of diag 308. This is necessary for kexec
(without kdump). The main difference to subcode 1 is that all CPUs get
a full reset, instead of the architectured CPU reset (which leaves all
registers untouched).
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Max Reitz [Thu, 19 Sep 2013 10:29:15 +0000 (12:29 +0200)]
qcow2: Correct snapshots size for overlap check
Using s->snapshots_size instead of snapshots_size for the metadata
overlap check in qcow2_write_snapshots leads to the detection of an
overlap with the main qcow2 image header when deleting the last
snapshot, since s->snapshots_size has not yet been updated and is
therefore non-zero. However, the offset returned by qcow2_alloc_clusters
will be zero since snapshots_size is zero. Therefore, an overlap is
detected albeit no such will occur.
This patch fixes this by replacing s->snapshots_size by snapshots_size
when calling qcow2_pre_write_overlap_check.
Signed-off-by: Max Reitz <mreitz@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Thomas Huth [Mon, 1 Jul 2013 13:44:18 +0000 (15:44 +0200)]
s390/ioinst: Moved the CC setting to the IO instruction handlers
The IO instruction handlers now take care of setting the CC value on
their own, so that the confusing return code magic in kvm_handle_css_inst()
is not needed anymore.
Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Thomas Huth [Tue, 2 Jul 2013 11:43:38 +0000 (13:43 +0200)]
s390/cpu: Make setcc() function available to other files
Moved the setcc() function to cpu.h so that it can be used by other
files, too. It now also does not modify the kvm state anymore since
this gets updated during kvm_arch_put_registers() anyway.
Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Andrew Jones [Wed, 18 Sep 2013 14:41:45 +0000 (16:41 +0200)]
target-i386: add feature kvm_pv_unhalt
I don't know yet if want this feature on by default, so for now I'm
just adding support for "-cpu ...,+kvm_pv_unhalt".
Signed-off-by: Andrew Jones <drjones@redhat.com> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com> Reviewed-by: Andreas Färber <afaerber@suse.de> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
target-i386: forward CPUID cache leaves when -cpu host is used
Some users running cpu intensive tasks checking the cache CPUID leaves at
startup and making decisions based on the result reported that the guest was
not reflecting the host CPUID leaves when -cpu host is used.
This patch fix this.
Signed-off-by: Benoît Canet <benoit@irqsave.net>
[Rename new field to cache_info_passthrough - Paolo] Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Jan Kiszka [Tue, 3 Sep 2013 16:08:51 +0000 (18:08 +0200)]
kvmvapic: Enter inactive state on hardware reset
ROM layout may change after reset of devices are hotplugged, so we have
to pick up the physical address again when the ROM is initialized. This
is best achieved by resetting the state to INACTIVE.
CC: qemu-stable@nongnu.org Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Jan Kiszka [Tue, 3 Sep 2013 16:08:50 +0000 (18:08 +0200)]
kvmvapic: Catch invalid ROM size
If not caught early, a zero-length ROM will cause a NULL-pointer access
later on in patch_hypercalls when allocating a zero-length ROM copy and
trying to read from it.
CC: qemu-stable@nongnu.org Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm irqfd: support direct msimessage to irq translation
On PPC64 systems MSI Messages are translated to system IRQ in a PCI
host bridge. This is already supported for emulated MSI/MSIX but
not for irqfd where the current QEMU allocates IRQ numbers from
irqchip and maps MSIMessages to IRQ in the host kernel.
This adds a new direct mapping flag which tells
the kvm_irqchip_add_msi_route() function that a new VIRQ
should not be allocated, instead the value from MSIMessage::data
should be used. It is up to the platform code to make sure that
this contains a valid IRQ number as sPAPR does in spapr_pci.c.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Andrew Jones [Fri, 23 Aug 2013 13:24:37 +0000 (15:24 +0200)]
kvm: warn if num cpus is greater than num recommended
The comment in kvm_max_vcpus() states that it's using the recommended
procedure from the kernel API documentation to get the max number
of vcpus that kvm supports. It is, but by always returning the
maximum number supported. The maximum number should only be used
for development purposes. qemu should check KVM_CAP_NR_VCPUS for
the recommended number of vcpus. This patch adds a warning if a user
specifies a number of cpus between the recommended and max.
Signed-off-by: Andrew Jones <drjones@redhat.com> Acked-by: Marcelo Tosatti <mtosatti@redhat.com> Signed-off-by: Gleb Natapov <gleb@redhat.com>
James Hogan [Tue, 27 Aug 2013 11:19:10 +0000 (12:19 +0100)]
cpu: Move cpu state syncs up into cpu_dump_state()
The x86 and ppc targets call cpu_synchronize_state() from their
*_cpu_dump_state() callbacks to ensure that up to date state is dumped
when KVM is enabled (for example when a KVM internal error occurs).
Move this call up into the generic cpu_dump_state() function so that
other KVM targets (namely MIPS) can take advantage of it.
This requires kvm_cpu_synchronize_state() and cpu_synchronize_state() to
be moved out of the #ifdef NEED_CPU_H in <sysemu/kvm.h> so that they're
accessible to qom/cpu.c.
Signed-off-by: James Hogan <james.hogan@imgtec.com> Cc: Andreas Färber <afaerber@suse.de> Cc: Alexander Graf <agraf@suse.de> Cc: Gleb Natapov <gleb@redhat.com> Cc: qemu-ppc@nongnu.org Cc: kvm@vger.kernel.org Signed-off-by: Gleb Natapov <gleb@redhat.com>
Andrea Arcangeli [Thu, 25 Jul 2013 10:11:15 +0000 (12:11 +0200)]
exec: always use MADV_DONTFORK
MADV_DONTFORK prevents fork to fail with -ENOMEM if the default
overcommit heuristics decides there's too much anonymous virtual
memory allocated. If the KVM secondary MMU is synchronized with MMU
notifiers or not, doesn't make a difference in that regard.
Secondly it's always more efficient to avoid copying the guest
physical address space in the fork child (so we avoid to mark all the
guest memory readonly in the parent and so we skip the establishment
and teardown of lots of pagetables in the child).
In the common case we can ignore the error if MADV_DONTFORK is not
available. Leave a second invocation that errors out in the KVM path
if MMU notifiers are missing and KVM is enabled, to abort in such
case.
Signed-off-by: Andrea Arcangeli <aarcange@redhat.com> Tested-By: Benoit Canet <benoit@irqsave.net> Acked-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Gleb Natapov <gleb@redhat.com>
Gabriel Kerneis [Tue, 17 Sep 2013 16:26:48 +0000 (18:26 +0200)]
coroutine: fix /perf/nesting coroutine benchmark
The /perf/nesting benchmark is broken because the counters are
not reset after each iteration. Therefore, nesting is done only
on the first iteration, and skipped on every other.
This patch fixes the issue, and reduces the number of iterations
to make it possible to run the benchmark in a reasonable amount of
time.
Signed-off-by: Gabriel Kerneis <gabriel@kerneis.info> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Gabriel Kerneis [Tue, 17 Sep 2013 15:09:39 +0000 (17:09 +0200)]
coroutine: add qemu_coroutine_yield benchmark
Current coroutine performance benchmarks test only coroutine creation,
either directly or in a nested way. This patch adds a benchmark to
evaluate the performance of qemu_coroutine_yield.
Signed-off-by: Gabriel Kerneis <gabriel@kerneis.info> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Hans de Goede [Tue, 17 Sep 2013 19:44:49 +0000 (21:44 +0200)]
xhci: Init a transfers xhci, slotid and epid member on epctx alloc
Transfers are part of an epctx, which is part of a slot, which is part of
a xhci. Transfers cannot dynamically be moved from one epctx to another,
so once created their xhci, slotid and epid are constant, so lets set these
up at creation time, rather then re-initializing them with the same
value each time a transfer gets submitted.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The usb-host code has been rewritten for qemu 1.5 to use libusb,
the old code has been left in as temporary fallback. Now we are
two releases further out, targeting the 1.7 release. No major
issues with the new code poped up until now. Time to remove it
from tre tree. Should we ever need it again for some reason --
git has a copy for us in the history.
Paolo Bonzini [Thu, 12 Sep 2013 09:02:20 +0000 (11:02 +0200)]
qemu-timer: do not take the lock in timer_pending
We can deduce the result from expire_time, by making it always -1 if
the timer is not in the active_timers list. We need to check against
negative times passed to timer_mod_ns; clamping them to zero is not
a problem because the only clock that has a zero value at VM startup
is QEMU_CLOCK_VIRTUAL, and it is monotonic so it cannot be non-zero.
QEMU_CLOCK_HOST, instead, is not monotonic but it cannot go to negative
values unless the host time is seriously screwed up and points to
the 1960s.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Stefan Hajnoczi [Thu, 12 Sep 2013 09:02:18 +0000 (11:02 +0200)]
qemu-timer: drop outdated signal safety comments
host_alarm_handler() is invoked from the signal processing thread
(currently the iothread). Previously we did processing in a real signal
handler with signalfd and therefore needed signal-safe timer code.
Today host_alarm_handler() just marks the alarm timer as expired/pending
and notifies the main loop using qemu_notify_event().
Therefore these outdated comments about signal safety can be dropped.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Stefan Hajnoczi [Thu, 22 Aug 2013 09:29:03 +0000 (11:29 +0200)]
osdep: warn if open(O_DIRECT) on fails with EINVAL
Print a warning when opening a file O_DIRECT fails with EINVAL. This
saves users a lot of time trying to figure out the EINVAL error, which
is typical when attempting to open a file O_DIRECT on Linux tmpfs.
Reported-by: Deepak C Shetty <deepakcs@linux.vnet.ibm.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
projects / qemu.git / log