Anthony Liguori [Mon, 19 Nov 2012 15:27:13 +0000 (09:27 -0600)]
Merge remote-tracking branch 'bonzini/build-urgent' into staging
* bonzini/build-urgent:
Makefile: Add missing dependency (fix parallel builds)
tests: link in stubs
libcacard: link in stubs
libcacard: make unnesting rules available to Makefile.objs
Anthony Liguori [Mon, 19 Nov 2012 15:27:00 +0000 (09:27 -0600)]
Merge remote-tracking branch 'kwolf/for-anthony' into staging
* kwolf/for-anthony: (26 commits)
qemu-io: Use bdrv_drain_all instead of qemu_aio_flush
megasas: Use bdrv_drain_all instead of qemu_aio_flush
vmdk: Fix data corruption bug in WRITE and READ handling
fdc: remove last usage of FD_STATE_SEEK
fdc: fix typo in zero constant
fdc: remove double affectation of FD_MSR_CMDBUSY flag
fdc-tests: add tests for VERIFY command
fdc: implement VERIFY command
fdc-test: Check READ ID
fdc: fix false FD_SR0_SEEK
fdc: fix FD_SR0_SEEK for initial seek on DMA transfers
fdc: fix FD_SR0_SEEK for non-DMA transfers and multi sectors transfers
fdc: use status0 field instead of a local variable
fdc-test: add tests for non-DMA READ command
fdc-test: insert media before fuzzing registers
fdc-test: split test_media_change() test, so insert part can be reused
fdc: Remove status0 parameter from fdctrl_set_fifo()
aio: rename AIOPool to AIOCBInfo
aio: use g_slice_alloc() for AIOCB pooling
aio: switch aiocb_size type int -> size_t
...
Anthony Liguori [Mon, 19 Nov 2012 15:26:48 +0000 (09:26 -0600)]
Merge remote-tracking branch 'kraxel/usb.71' into staging
* kraxel/usb.71:
usb-host: fix splitted transfers
usb-host: update tracing
usb-redir: Set default debug level to warning
usb-redir: Only add actually in flight packets to the in flight queue
ehci: handle dma errors
ehci: keep the frame timer running in case the guest asked for frame list rollover interrupts
ehci: Don't verify the next pointer for periodic qh-s and qtd-s
ehci: Better detection for qtd-s linked in circles
ehci: Fixup q->qtdaddr after cancelling an already completed packet
ehci: Don't access packet after freeing it
usb: host-linux: Ignore parsing errors of the device descriptors
usb-host: scan for usb devices when the vm starts
usb: Fix (another) bug in usb_packet_map() for IOMMU handling
fix live migration
Anthony Liguori [Mon, 19 Nov 2012 14:31:24 +0000 (08:31 -0600)]
Merge remote-tracking branch 'afaerber/qom-cpu' into staging
* afaerber/qom-cpu:
target-i386: Add Haswell CPU model
target-i386/cpu: Add new Opteron CPU model
target-i386/cpu: Name new CPUID bits
qapi-types.h: Don't include qemu-common.h
osdep: Move qemu_{open,close}() prototypes
qemu-config.h: Include headers it needs
vnc-palette.h: Include <stdbool.h>
qemu-fsdev-dummy.c: Include module.h
qdev: Split up header so it can be used in cpu.h
Move qemu_irq typedef out of qemu-common.h
qemu-common.h: Comment about usage rules
The TCG interpreter still does not support this kind of
optimization. Therefore builds with TCI fail with an
unresolved symbol tcg_out_tb_finalize. This is fixed here.
Stefan Weil [Sun, 18 Nov 2012 19:06:19 +0000 (20:06 +0100)]
tci: fix build breakage for target MIPS
commit 5f7319cd introduced GETPC() usage for MIPS, which is currently
not defined when building with --enable-tcg-interpreter. Add MIPS to
the list of targets we selectively define GETPC() for.
Paolo Bonzini [Fri, 16 Nov 2012 17:35:27 +0000 (18:35 +0100)]
build: replace weak symbols with a static library
Weak symbols were a nice idea, but they turned out not to be a good one.
Toolchain support is just too sparse, in particular llvm-gcc is totally
broken.
This patch uses a surprisingly low-tech approach: a static library.
Symbols in a static library are always overridden by symbols in an
object file. Furthermore, if you place each function in a separate
source file, object files for unused functions will not be taken in.
This means that each function can use all the dependencies that it needs
(especially QAPI stuff such as error_setg).
Thus, all stubs are placed in separate object files and put together in
a static library. The library then is linked to all programs.
Amit Shah [Wed, 20 Jun 2012 06:59:32 +0000 (12:29 +0530)]
virtio-rng: hardware random number generator device
The Linux kernel already has a virtio-rng driver, this is the device
implementation.
When the guest asks for entropy from the virtio hwrng, it puts a buffer
in the vq. We then put entropy into that buffer, and push it back to
the guest.
Signed-off-by: Amit Shah <[email protected]> Signed-off-by: Anthony Liguori <[email protected]>
---
aliguori: converted to new RngBackend interface
aliguori: remove entropy needed event
aliguori: fix migration
Anthony Liguori [Mon, 25 Jun 2012 19:36:33 +0000 (14:36 -0500)]
vl: add -object option to create QOM objects from the command line
This will create a new QOM object in the '/objects' path. Note that properties
are set in order which allows for simple objects to be initialized entirely
with this option and then realized.
This option is roughly equivalent to -device but for things that are not
devices.
Over time various systemtap reserved words have been blacklisted
in the trace backend generator. The list is not complete though,
so there is continued risk of problems in the future. Preempt
such problems by specifying the full list of systemtap keywords
listed in its parser as identified here:
Gerd Hoffmann [Thu, 15 Nov 2012 12:07:49 +0000 (13:07 +0100)]
ehci: handle dma errors
Starting with commit 1c380f9460522f32c8dd2577b2a53d518ec91c6d dma
transfers can actually fail. This patch makes ehci keep track
of the busmaster bit in pci config space, by setting/clearing the
dma_context pointer. Attempts to dma without context will result
in raising HSE (Host System Error) interrupt and stopping the host
controller.
This patch fixes WinXP not booting with a usb stick attached to ehci.
Root cause is seabios activating ehci so you can boot from the stick,
and WinXP clearing the busmaster bit before resetting the host
controller, leading to ehci actually trying dma while it is disabled.
Hans de Goede [Thu, 15 Nov 2012 13:14:58 +0000 (14:14 +0100)]
ehci: Don't verify the next pointer for periodic qh-s and qtd-s
While testing the move to async packet handling for interrupt endpoints I
noticed that Windows-XP likes to play tricks with the next pointer for
periodic qh-s, so we should not fail qh / qtd verification when it changes.
Hans de Goede [Wed, 14 Nov 2012 16:21:38 +0000 (17:21 +0100)]
ehci: Better detection for qtd-s linked in circles
Windows links interrupt qtd-s in circles, which means that when interrupt
endpoints return USB_RET_ASYNC, combined with the recent
"ehci: Retry to fill the queue while waiting for td completion" patch,
we keep adding the tds to the queue over and over again, as we detect the
circle from fill_queue, but we call it over and over again ...
This patch fixes this by changing the circle detection to also detect
circling into tds already queued up previously.
Hans de Goede [Wed, 14 Nov 2012 16:21:37 +0000 (17:21 +0100)]
ehci: Fixup q->qtdaddr after cancelling an already completed packet
This avoids the q->qtdaddr == p->qtdaddr asserts we have triggering, when
a queue contains multiple completed packages when we cancel the queue.
I triggered this with windows7 + async interrupt endpoint handling (*)
+ not detecting circles in ehci_fill_queue() properly, which makes the qtd
validation in ehci_fill_queue fail, causing cancellation of the queue on every
mouse event ...
*) Which is not going upstream as it will cause loss of interrupt events on
migration.
Jan Kiszka [Thu, 15 Nov 2012 08:23:30 +0000 (09:23 +0100)]
usb: host-linux: Ignore parsing errors of the device descriptors
The Linux is more tolerant here as well: Just stop parsing the device
descriptors when an error is detected but do not reset what was found
so far. This allows to run buggy devices with partially invalid
descriptors.
Gerd Hoffmann [Mon, 12 Nov 2012 11:40:46 +0000 (12:40 +0100)]
vga: fix bochs alignment issue
The bochs dispi interface traditionally uses port 0x1ce as 16bit index
register and port 0x1cf as 16bit data register. The later is unaligned,
and probably for that reason the the data register was moved to 0x1d0
for non-x86 archs.
This patch makes the data register available at 0x1d0 on x86 too. The
old x86 location is kept for compatibility reasons, so both 0x1cf and
0x1d0 can be used as data register on x86.
Gerd Hoffmann [Wed, 14 Nov 2012 14:51:18 +0000 (15:51 +0100)]
usb-host: scan for usb devices when the vm starts
Commit a844ed842d9a9d929645c09ae0f52f753d7a02e0 leads to usb-host
detecting devices not right after qemu startup because the guest
isn't running yet. Instead they are found on the first of the
regular usb device poll runs. Which is too late for seabios to see
them, so booting from usb sticks fails.
Fix this by adding a vm state change handler which triggers a device
scan when the vm is started.
David Gibson [Wed, 14 Nov 2012 05:23:50 +0000 (16:23 +1100)]
usb: Fix (another) bug in usb_packet_map() for IOMMU handling
Elements in qemu SGLists can cross IOMMU page boundaries. So, in commit 39c138c8420f51a7da7b35233a8d7400a0b589ac "usb: Fix usb_packet_map() in the
presence of IOMMUs", I changed usb_packet_map() to split up each SGList
element on IOMMU page boundaries and each resulting piece of qemu's memory
space separately to the iovec the usb code uses internally.
That was correct in concept, but the patch has a bug. The 'base' variable
correctly steps through the dma address of each piece, but then we call
the dma_memory_map() function on the base address of the whole SGList
element every time.
This patch fixes at least one problem using XHCI on the pseries guest
machine. It didn't affect OHCI because that doesn't use usb_packet_map().
In theory it also affects EHCI, but we haven't observed that in practice.
I think the transfers were small enough on EHCI that they never crossed an
IOMMU page boundary in practice.
Gerd Hoffmann [Thu, 8 Nov 2012 14:54:05 +0000 (15:54 +0100)]
fix live migration
Commit 1c380f9460522f32c8dd2577b2a53d518ec91c6d breaks live migration.
DMA stops working for ehci (and probably for any pci device) after
restoring the guest because the bus master region never gets enabled.
Add code doing that after loading the pci config space from vmstate.
Alon Levy [Wed, 14 Nov 2012 13:09:07 +0000 (15:09 +0200)]
virtio-serial-bus: post_load send_event when vm is running
Alexander Larsson found irq injection to Windows guests stopped after a
migration. The symptom was the mouse stopped working.
Reproduction steps are:
1. On src, start qemu with a virtio-serial port without any backend
2. On dest, start qemu with a virtio-serial port with a backend
3. Migrate.
Upon migration, the older code detected the change in backend connection
status, and sent a notification to the guest. However, it's not
guaranteed that the apic is ready to inject irqs into the guest, and the
irq line remained high, resulting in any future interrupts going
unnoticed by the guest as well.
Add a new timer based on vm_clock for 1 ns in the future from post_load
to do the event send in case host_connected differs between migration
source and target.
While reading microMIPS decoding, I found a possible wrong opcode
encoding. According to [1] page 166, the bits 13..12 for MULTU is
0x01 rather than 0x00. Please review, thanks.
[1] MIPS Architecture for Programmers VolumeIV-e: The MIPS DSP
Application-Specific Extension to the microMIPS32 Architecture
Aurelien Jarno [Wed, 14 Nov 2012 14:04:42 +0000 (15:04 +0100)]
mips/malta: fix CBUS UART interrupt pin
According to the MIPS Malta Developement Platform User's Manual, the
i8259 interrupt controller is supposed to be connected to the hardware
IRQ0, and the CBUS UART to the hardware interrupt 2.
In QEMU they are both connected to hardware interrupt 0, the CBUS UART
interrupt being wrong. This patch fixes that. It should be noted that
the irq array in QEMU includes the software interrupts, hence
env->irq[2] is the first hardware interrupt.
Andre Przywara [Wed, 14 Nov 2012 18:28:53 +0000 (16:28 -0200)]
target-i386/cpu: Add new Opteron CPU model
Add a new base CPU model called Opteron_G5 to model the latest
Opteron CPUs. This increases the model value and model numbers and
adds TBM, F16C and FMA over the latest G4 model.
Andre Przywara [Wed, 14 Nov 2012 18:28:52 +0000 (16:28 -0200)]
target-i386/cpu: Name new CPUID bits
Update QEMU's knowledge of CPUID bit names. This allows to
enable/disable those new features on QEMU's command line when
using KVM and prepares future feature enablement in QEMU.
This adds F16C, RDRAND, LWP, TBM, TopoExt, PerfCtr_Core, PerfCtr_NB,
FSGSBASE, BMI1, AVX2, BMI2, ERMS, PCID, InvPCID, RTM, RDSeed and ADX.
Sources where the AMD BKDG for Family 15h/Model 10h, Intel Software
Developer Manual, and the Linux kernel for the leaf 7 bits.
Signed-off-by: Andre Przywara <[email protected]> Signed-off-by: Boris Ostrovsky <[email protected]>
[ehabkost: added CPUID_EXT_PCID]
[ehabkost: edited commit message]
[ehabkost: rebased against latest qemu.git master] Signed-off-by: Eduardo Habkost <[email protected]> Reviewed-by: Igor Mammedov <[email protected]> Signed-off-by: Andreas Färber <[email protected]>
Eduardo Habkost [Tue, 23 Oct 2012 23:29:38 +0000 (21:29 -0200)]
qemu-config.h: Include headers it needs
Include:
- <stdio.h> for FILE
- qemu-option.h for QemuOptsList
Some of those headers were probably being included by accident because
some other headers were including qemu-common.h, but those headers
should eventually stop including qemu-common.h.
Eduardo Habkost [Tue, 23 Oct 2012 23:01:36 +0000 (21:01 -0200)]
vnc-palette.h: Include <stdbool.h>
<stdbool.h> is needed for the 'bool' type, used in the header.
The header is probably being included by accident because some other
headers are including qemu-common.h, but those headers should eventually
stop including qemu-common.h.
Eduardo Habkost [Tue, 23 Oct 2012 21:43:01 +0000 (19:43 -0200)]
qemu-fsdev-dummy.c: Include module.h
module.h is where machine_init() is defined, but qemu-fsdev-dummy.c
doesn't include it.
The header is probably being included by accident because some other
headers are including qemu-common.h, but those headers should eventually
stop including qemu-common.h.
Anthony Liguori [Fri, 10 Aug 2012 17:00:43 +0000 (12:00 -0500)]
qdev: Split up header so it can be used in cpu.h
Header file dependency is a frickin' nightmare right now. cpu.h tends
to get included in our 'include everything' header files but qdev also
needs to include those headers mainly for qdev-properties since it knows
about CharDriverState and friends.
We can solve this for now by splitting out qdev.h along the same lines
that we previously split the C file. Then cpu.h just needs to include
qdev-core.h.
hw/qdev.h is split into following new headers:
hw/qdev-core.h
hw/qdev-properties.h
hw/qdev-monitor.h
Signed-off-by: Anthony Liguori <[email protected]>
[ehabkost: re-add DEFINE_PROP_PCI_HOST_DEVADDR, that was removed on the
original patch (by mistake, I guess)]
[ehabkost: kill qdev_prop_set_vlan() declaration]
[ehabkost: moved get_fw_dev_path() comment to the original location
(I don't know why it was moved)]
[ehabkost: removed qdev_exists() declaration]
[ehabkost: keep using 'QemuOpts' instead of 'struct QemuOpts', as
qdev-core.h includes qemu-option.h]
Igor Mammedov [Tue, 16 Oct 2012 01:57:21 +0000 (03:57 +0200)]
Move qemu_irq typedef out of qemu-common.h
It's necessary for making CPU child of DEVICE without
causing circular header deps.
Signed-off-by: Igor Mammedov <[email protected]>
[ehabkost: re-added the typedef to hw/irq.h after rebasing] Signed-off-by: Eduardo Habkost <[email protected]> Signed-off-by: Andreas Färber <[email protected]>
Eduardo Habkost [Tue, 23 Oct 2012 21:44:02 +0000 (19:44 -0200)]
qemu-common.h: Comment about usage rules
Every time we make a tiny change on a header file, we often find
circular header dependency problems. To avoid this nightmare, we need to
stop including qemu-common.h from other headers, and we should gradually
move the declarations from the catch-all qemu-common.h header to their
specific headers.
This simply adds a comment documenting the rules about qemu-common.h,
hoping that people will see it before including qemu-common.h from other
header files, and before adding more declarations to qemu-common.h.
Jan Kiszka [Mon, 12 Nov 2012 14:04:35 +0000 (15:04 +0100)]
kvm: Actually remove software breakpoints from list on cleanup
So far we only removed them from the guest, leaving its states in the
list. This made it impossible for gdb to re-enable breakpoints on the
same address after re-attaching.
Marcelo Tosatti [Tue, 30 Oct 2012 23:14:19 +0000 (21:14 -0200)]
acpi_piix4: fix migration of gpe fields
Migrate 16 bytes for en/sts fields (which is the correct size),
increase version to 3, and document how to support incoming
migration from qemu-kvm 1.2.
Kevin Wolf [Tue, 13 Nov 2012 15:41:22 +0000 (16:41 +0100)]
qemu-io: Use bdrv_drain_all instead of qemu_aio_flush
This is harmless as of today because I/O throttling is not used in
qemu-io, however as soon as .bdrv_drain handlers will be introduced,
qemu-io must be sure to call bdrv_drain_all().
vmdk: Fix data corruption bug in WRITE and READ handling
Fixed a MAJOR BUG in VMDK files on file boundaries on reads
and ALSO ON WRITES WHICH MIGHT CORRUPT THE IMAGE AND DATA!!!!!!
Triggered for example with the following VMDK file (partly listed):
RW 4193792 FLAT "XP-W1-f001.vmdk" 0
RW 2097664 FLAT "XP-W1-f002.vmdk" 0
RW 4193792 FLAT "XP-W1-f003.vmdk" 0
RW 512 FLAT "XP-W1-f004.vmdk" 0
RW 4193792 FLAT "XP-W1-f005.vmdk" 0
RW 2097664 FLAT "XP-W1-f006.vmdk" 0
RW 4193792 FLAT "XP-W1-f007.vmdk" 0
RW 512 FLAT "XP-W1-f008.vmdk" 0
Patch includes:
1.) Patch fixes wrong calculation on extent boundaries. Especially it
fixes the relativeness of the sector number to the current extent.
Verfied correctness with:
1.) Converted either with Virtualbox to VDI and then with qemu-img and
then with qemu-img only:
VBoxManage clonehd --format vdi /VM/XP-W/new/XP-W1.vmdk ~/.VirtualBox/Harddisks/XP-W1-new-test.vdi
./qemu-img convert -O raw ~/.VirtualBox/Harddisks/XP-W1-new-test.vdi /root/QEMU/VM-XP-W1/XP-W1-via-VBOX.img
md5sum /root/QEMU/VM-XP-W/XP-W1-direct.img
md5sum /root/QEMU/VM-XP-W/XP-W1-via-VBOX.img
=> same MD5 hash
2.) Verified debug log files
3.) Run Windows XP successfully
4.) chkdsk run successfully without any errors
VERIFY command is like a READ command, except that read data is not
transfered by DMA.
As DMA engine is not used, so we have to start data transfer ourselves.
Kevin Wolf [Mon, 3 Sep 2012 23:46:54 +0000 (23:46 +0000)]
fdc: Remove status0 parameter from fdctrl_set_fifo()
It decided whether an interrupt is triggered. Only one caller made use
of this functionality, so move the code there.
In this one caller, the interrupt must actually be triggered
unconditionally, like it was before commit 2fee0088. For example, a
successful read without an implied seek can result in st0 = 0, but still
triggers the interrupt.
Stefan Hajnoczi [Wed, 31 Oct 2012 15:34:37 +0000 (16:34 +0100)]
aio: rename AIOPool to AIOCBInfo
Now that AIOPool no longer keeps a freelist, it isn't really a "pool"
anymore. Rename it to AIOCBInfo and make it const since it no longer
needs to be modified.
Stefan Hajnoczi [Wed, 31 Oct 2012 15:34:36 +0000 (16:34 +0100)]
aio: use g_slice_alloc() for AIOCB pooling
AIO control blocks are frequently acquired and released because each aio
request involves at least one AIOCB. Therefore, we pool them to avoid
heap allocation overhead.
The problem with the freelist approach in AIOPool is thread-safety. If
we want BlockDriverStates to associate with AioContexts that execute in
multiple threads, then a global freelist becomes a problem.
This patch drops the freelist and instead uses g_slice_alloc() which is
tuned for per-thread fixed-size object pools. qemu_aio_get() and
qemu_aio_release() are now thread-safe.
Note that the change from g_malloc0() to g_slice_alloc() should be safe
since the freelist reuse case doesn't zero the AIOCB either.
Stefan Hajnoczi [Wed, 31 Oct 2012 15:34:35 +0000 (16:34 +0100)]
aio: switch aiocb_size type int -> size_t
Using appropriate types for variables is a good thing :). All users
simply do sizeof(MyType) and the value is passed to a memory allocator,
it should be size_t.
Nick Thomas [Fri, 2 Nov 2012 13:01:23 +0000 (13:01 +0000)]
tests: allow qemu-iotests to be run against nbd backend
To do this, we start a qemu-nbd process at _make_test_img and kill
it in _cleanup_test_img. $TEST_IMG is changed to point at the TCP
server. We also remove the checks for existence of binaries from
common.config - they're duplicated in common, and we can make the
qemu-nbd check conditional on $IMGPROTO being "nbd" if we do it there.
Stefan Weil [Sun, 4 Nov 2012 11:09:34 +0000 (12:09 +0100)]
block: Workaround for older versions of MinGW gcc
Versions before gcc-4.6 don't support unnamed fields in initializers
(see http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676).
Offset and OffsetHigh belong to an unnamed struct which is part of an
unnamed union. Therefore the original code does not work with older
versions of gcc.
Kevin Wolf [Fri, 26 Oct 2012 18:31:15 +0000 (20:31 +0200)]
qemu-iotests: qcow2: Test growing large refcount table
Actually writing all the content with 512 byte sector size would take
forever, therefore build the image file with a Python script and use
qemu-io for the last write that actually triggers the refcount table
growth.
Kevin Wolf [Fri, 26 Oct 2012 14:33:32 +0000 (16:33 +0200)]
qcow2: Fix refcount table size calculation
A missing factor for the refcount table entry size in the calculation
could mean that too little memory was allocated for the in-memory
representation of the table, resulting in a buffer overflow.
Anthony Liguori [Wed, 14 Nov 2012 14:53:03 +0000 (08:53 -0600)]
Merge remote-tracking branch 'kraxel/pixman.v5' into staging
* kraxel/pixman.v5:
pixman: cleanup properly on make distclean
pixman: add licensing info
pixman: build internal version early
pixman: pass cflags, add -fPIC
pixman: disable gtk
pixman: set --host for cross builds
pixman: add output dir to include path
Anthony Liguori [Wed, 14 Nov 2012 14:51:06 +0000 (08:51 -0600)]
Merge remote-tracking branch 'bonzini/nbd-next' into staging
* bonzini/nbd-next:
nbd: fixes to read-only handling
hmp: add NBD server commands
nbd: disallow nbd-server-add before nbd-server-start
nbd: force read-only export for read-only devices
nbd: fix nbd_server_stop crash when no server was running
nbd: accept URIs
nbd: accept relative path to Unix socket
qemu-nbd: initialize main loop before block layer
Anthony Liguori [Wed, 14 Nov 2012 14:50:45 +0000 (08:50 -0600)]
Merge remote-tracking branch 'bonzini/scsi-next' into staging
* bonzini/scsi-next:
virtio-scsi: use dma_context_memory
dma: Define dma_context_memory and use in sysbus-ohci
megasas: Correct target/lun mapping
scsi-disk: flush cache after disabling it
megasas: do not include block_int.h
scsi: remove superfluous call to scsi_device_set_ua
virtio-scsi: factor checks for VIRTIO_SCSI_S_DRIVER_OK when reporting events
scsi: do not return short responses for emulated commands
projects / qemu.git / log