Set the ZynqMP number of priority queues to 2.
Signed-off-by: Alistair Francis
Reviewed-by: Peter Maydell
---
hw/arm/xlnx-zynqmp.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/hw/arm/xlnx-zynqmp.c b/hw/arm/xlnx-zynqmp.c
index 23c7199..0d86ba3 100644
--- a/hw/arm/xlnx-zynqmp.c
+++ b/hw
Le 28/07/2016 à 13:57, Peter Maydell a écrit :
> For i386, the ABI specifies that 'long long' (8 byte values)
> need only be 4 aligned, but we were requiring them to be
> 8-aligned. This meant we were laying out the target_epoll_event
> structure wrongly. Add a suitable ifdef to abitypes.h to
> s
Le 28/07/2016 à 13:57, Peter Maydell a écrit :
> For i386, the ABI specifies that 'long long' (8 byte values)
> need only be 4 aligned, but we were requiring them to be
> 8-aligned. This meant we were laying out the target_epoll_event
> structure wrongly. Add a suitable ifdef to abitypes.h to
> s
On 28 July 2016 at 20:19, Laurent Vivier wrote:
> Why the following program from commit
>
> c2e3dee linux-user: Define target alignment size
>
> int main(void)
> {
> printf("alignof(short) %ld\n", __alignof__(short));
> printf("alignof(int) %ld\n", __alignof__(int));
> printf("alig
On 28 July 2016 at 21:36, Laurent Vivier wrote:
>
>
> Le 28/07/2016 à 13:57, Peter Maydell a écrit :
>> For i386, the ABI specifies that 'long long' (8 byte values)
>> need only be 4 aligned, but we were requiring them to be
>> 8-aligned. This meant we were laying out the target_epoll_event
>> str
http://lists.nongnu.org/archive/html/qemu-devel/2016-07/msg04839.html
If there is any way this can be pulled in, we FreeBSD nerds would
appreciate it.
sean
signature.asc
Description: OpenPGP digital signature
On Tue, May 17, 2016 at 06:18:45PM +0800, Wei Jiangang wrote:
> The first had been reviewed.
> The second had been posted last month, but no feedback.
> They're similar, so resend them together.
>
> Wei Jiangang (2):
> hw/pci-bridge: Convert pxb initialization functions to Error
> apb: convert
On 07/28/2016 09:29 AM, Halil Pasic wrote:
>> You mean va_start, not start_va. And actually, C11 is clear that errno
>> is unspecified after library functions (but not macros) that don't
>> explicitly state otherwise. Since va_start() is a macro and not a
>> library function, that means va_start
On 07/28/2016 08:37 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Free the list, not just the elements.
>
> Signed-off-by: Marc-André Lureau
> ---
> qga/main.c | 8 ++--
> 1 file changed, 2 insertions(+), 6 deletions(-)
>
> diff --git a/qga/main.c b/qga/main.c
> ind
On 07/28/2016 08:37 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Signed-off-by: Marc-André Lureau
> ---
> qga/guest-agent-command-state.c | 7 +++
> qga/guest-agent-core.h | 1 +
> qga/main.c | 6 ++
> 3 files changed, 14 insertions(
On 07/28/2016 09:24 AM, Peter Maydell wrote:
> Currently the -version command line argument prints a string ending
> with "Copyright (c) 2003-2008 Fabrice Bellard". This is now some
> eight years out of date; abstract it out of the several places that
> print the string and update it to:
>
> Copy
- Original Message -
> On 07/28/2016 08:37 AM, marcandre.lur...@redhat.com wrote:
> > From: Marc-André Lureau
> >
> > Free the list, not just the elements.
> >
> > Signed-off-by: Marc-André Lureau
> > ---
> > qga/main.c | 8 ++--
> > 1 file changed, 2 insertions(+), 6 deletions(-
On Thu, Jul 28, 2016 at 07:50:52AM +, Li, Liang Z wrote:
> > > }
> > >
> > > +static void update_free_pages_stats(struct virtio_balloon *vb,
> >
> > why _stats?
>
> Will change.
>
> > > + max_pfn = get_max_pfn();
> > > + mutex_lock(&vb->balloon_lock);
> > > + while (pfn < max_pfn) {
> > > +
On 07/28/2016 03:36 PM, Marc-André Lureau wrote:
>>> +g_list_foreach(config->blacklist, (GFunc)g_free, NULL);
>>
>> What an ugly cast - undefined behavior according to C. But it happens to
>> work on all ABI that we support (calling a unary function as cast to
>> appear as a binary function si
On 07/28/2016 08:37 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> srcfifo && dstfifo must be freed in error case, however unlink() may
> delete a file from a different context. Instead, use mkdtemp()/rmdir()
> for the temporary files.
>
> Signed-off-by: Marc-André Lureau
On 07/28/2016 08:37 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> chr_disconnect callback is used to cleanup server socket peer. This
> removes some leaks when the chardev is freed.
>
> Signed-off-by: Marc-André Lureau
> ---
> qemu-char.c | 1 +
> 1 file changed, 1 inser
On Thu, Jul 28, 2016 at 06:36:18AM +, Li, Liang Z wrote:
> > > > This ends up doing a 1MB kmalloc() right? That seems a _bit_ big.
> > > > How big was the pfn buffer before?
> > >
> > > Yes, it is if the max pfn is more than 32GB.
> > > The size of the pfn buffer use before is 256*4 = 1024 Byt
On 07/28/2016 08:37 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Hi,
>
> Thanks to AddressSanitizer (ASAN), I found a number of direct leaks
> worth fixing. Note that there are probably many indirect leaks left (I
> am adding some here), I haven't investigated much yet.
*** This bug is a duplicate of bug 1581936 ***
https://bugs.launchpad.net/bugs/1581936
** This bug has been marked a duplicate of bug 1581936
Frozen Windows 7 VMs with VGA CVE-2016-3712 fix (2.6.0 and 2.5.1.1)
--
You received this bug notification because you are a member of qemu-
devel-m
On Thu, 2016-07-28 at 21:32 +0530, Richard Henderson wrote:
> For the benefit of Peter's ./scripts/clean-includes and friends,
> please name these with *.inc.c, like in tcg/*/.
Ok. David merged the series in for-2.8 but I will send a subsequent
patch to fix them up.
Cheers,
Ben.
On Thu, 2016-07-28 at 21:36 +0530, Richard Henderson wrote:
> On 07/27/2016 03:51 AM, Benjamin Herrenschmidt wrote:
> >
> > +/* GETPC() works here because this is inline */
> > +raise_exception_err_ra(env, POWERPC_EXCP_PROGRAM,
> > + POWERP
On 07/28/2016 08:37 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Similarly to other chr_close callbacks, free char type specific data.
>
> Signed-off-by: Marc-André Lureau
> ---
> qemu-char.c | 9 +
> 1 file changed, 9 insertions(+)
>
> diff --git a/qemu-char.c
On 07/28/2016 08:38 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Simplify a bit the code by using g_strdup_printf() and store it in a
> non-const value so casting is no longer needed, and ownership is
> clearer.
>
> Signed-off-by: Marc-André Lureau
> ---
> hw/core/bus.c
On Fri, 2016-07-29 at 07:57 +1000, Benjamin Herrenschmidt wrote:
> On Thu, 2016-07-28 at 21:36 +0530, Richard Henderson wrote:
> >
> > On 07/27/2016 03:51 AM, Benjamin Herrenschmidt wrote:
> > >
> > >
> > > +/* GETPC() works here because this is inline */
> > > +raise_exc
On Thu, Jul 28, 2016 at 03:30:09AM +, Li, Liang Z wrote:
> > Subject: Re: [PATCH v2 repost 4/7] virtio-balloon: speed up inflate/deflate
> > process
> >
> > On Wed, Jul 27, 2016 at 09:03:21AM -0700, Dave Hansen wrote:
> > > On 07/26/2016 06:23 PM, Liang Li wrote:
> > > > + vb->pfn_limit
On 07/28/2016 08:38 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Allows to specify a destroy function for the test data.
"Allows to" is not idiomatic English. Alternatives that sound better are
"Allows $who to specify" (most simply, "Allows one to"), or "Allows
specifying
On Thu, Jul 28, 2016 at 03:06:37AM +, Li, Liang Z wrote:
> > > + * VIRTIO_BALLOON_PFNS_LIMIT is used to limit the size of page bitmap
> > > + * to prevent a very large page bitmap, there are two reasons for this:
> > > + * 1) to save memory.
> > > + * 2) allocate a large bitmap may fail.
> > >
On 07/28/2016 08:38 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> A few strings are allocated and never freed.
>
> Signed-off-by: Marc-André Lureau
> ---
> tests/postcopy-test.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
Reviewed-by: Eric Blake
--
Eri
On 07/28/2016 08:38 AM, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> The path is allocated and should be freed.
>
> The qmp response should be unref, but then 'machine' must be duplicated.
>
> Use a destroy function for the PCTestData.
>
> Signed-off-by: Marc-André Lureau
float_invalid_op_excp() and float_check_status() must be inline
due to their use of GETPC().
Use __attribute__((__always_inline__)) to enforce it
Signed-off-by: Benjamin Herrenschmidt
---
target-ppc/fpu_helper.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/target-p
The following changes since commit 21a21b853a1bb606358af61e738abfb9aecbd720:
Merge remote-tracking branch 'remotes/ehabkost/tags/x86-pull-request' into
staging (2016-07-27 18:18:21 +0100)
are available in the git repository at:
https://github.com/jnsnow/qemu.git tags/ide-pull-request
for y
If one attempts to perform a system_reset after a failed IO request
that causes the VM to enter a paused state, QEMU will segfault trying
to free up the pending IO requests.
These requests have already been completed and freed, though, so all
we need to do is NULL them before we enter the paused s
The target-ppc/excp_helper.c:powerpc_excp() case POWERPC_EXCP_FP fails
to set "env->spr[SPR_BOOKE_ESR] = ESR_FP;". I can submit a patch for
that, or anyone can add it, but I notice that in the other cases where
SPR_BOOKE_ESR is set, the "msr" is ALSO set. Since the "msr" is used
to initialize SRR
> From: Alex Williamson [mailto:alex.william...@redhat.com]
> Sent: Thursday, July 28, 2016 11:42 PM
>
> On Thu, 28 Jul 2016 10:15:24 +
> "Tian, Kevin" wrote:
>
> > Hi, Alex,
> >
> > Along with recent enhancement on virtual IOMMU (vIOMMU) in Qemu, I'm
> > thinking whether there is any issue
Le 28/07/2016 à 13:57, Peter Maydell a écrit :
> For i386, the ABI specifies that 'long long' (8 byte values)
> need only be 4 aligned, but we were requiring them to be
> 8-aligned. This meant we were laying out the target_epoll_event
> structure wrongly. Add a suitable ifdef to abitypes.h to
> s
> On Thu, Jul 28, 2016 at 03:06:37AM +, Li, Liang Z wrote:
> > > > + * VIRTIO_BALLOON_PFNS_LIMIT is used to limit the size of page
> > > > +bitmap
> > > > + * to prevent a very large page bitmap, there are two reasons for this:
> > > > + * 1) to save memory.
> > > > + * 2) allocate a large bitm
On 2016/7/29 3:07, Dr. David Alan Gilbert wrote:
* zhanghailiang (zhang.zhanghaili...@huawei.com) wrote:
This is the 17th version of COLO FT feature.
Here is only COLO frame part, you can get the whole codes from github:
https://github.com/coloft/qemu/commits/colo-v3.0-periodic-mode
Migration
> On Thu, Jul 28, 2016 at 06:36:18AM +, Li, Liang Z wrote:
> > > > > This ends up doing a 1MB kmalloc() right? That seems a _bit_ big.
> > > > > How big was the pfn buffer before?
> > > >
> > > > Yes, it is if the max pfn is more than 32GB.
> > > > The size of the pfn buffer use before is 256*
On 07/27/2016 03:51 AM, Benjamin Herrenschmidt wrote:
-tcg_gen_andi_tl(EA, EA, ~0xf);\
-/* We only need to swap high and low halves. gen_qemu_ld64 does necessary \
- 64-bit byteswap already. */\
On 07/29/2016 08:41 AM, Hailiang Zhang wrote:
On 2016/7/29 3:07, Dr. David Alan Gilbert wrote:
* zhanghailiang (zhang.zhanghaili...@huawei.com) wrote:
This is the 17th version of COLO FT feature.
Here is only COLO frame part, you can get the whole codes from github:
https://github.com/coloft/q
On 2016/7/29 8:55, Changlong Xie wrote:
On 07/29/2016 08:41 AM, Hailiang Zhang wrote:
On 2016/7/29 3:07, Dr. David Alan Gilbert wrote:
* zhanghailiang (zhang.zhanghaili...@huawei.com) wrote:
This is the 17th version of COLO FT feature.
Here is only COLO frame part, you can get the whole codes
On 07/28/2016 11:44 PM, Nikunj A Dadhania wrote:
From: Sandipan Das
Adds following instructions:
vabsdub: Vector Absolute Difference Unsigned Byte
vabsduh: Vector Absolute Difference Unsigned Halfword
vabsduw: Vector Absolute Difference Unsigned Word
Signed-off-by: Sandipan Das
[ use ISA300
On 07/28/2016 11:44 PM, Nikunj A Dadhania wrote:
From: Swapnil Bokade
Adds following instructions:
vcmpnezb[.]: Vector Compare Not Equal or Zero Byte
vcmpnezh[.]: Vector Compare Not Equal or Zero Halfword
vcmpnezw[.]: Vector Compare Not Equal or Zero Word
Signed-off-by: Swapnil Bokade
[ coll
On 2016/7/29 0:38, Wei Huang wrote:
> This patch adds a pmu=[on/off] option to enable/disable vpmu support
> in guest vm. There are several reasons to justify this option. First
> vpmu can be problematic for cross-migration between different SoC as
> perf counters is architecture-dependent. It is
> > > On Wed, Jul 27, 2016 at 09:03:21AM -0700, Dave Hansen wrote:
> > > > On 07/26/2016 06:23 PM, Liang Li wrote:
> > > > > + vb->pfn_limit = VIRTIO_BALLOON_PFNS_LIMIT;
> > > > > + vb->pfn_limit = min(vb->pfn_limit, get_max_pfn());
> > > > > + vb->bmap_len = ALIGN(vb->pfn_limit, BITS_P
On Thu, 2016-07-28 at 09:51 +0200, Cédric Le Goater wrote:
> On 07/28/2016 04:14 AM, Andrew Jeffery wrote:
> >
> > On Wed, 2016-07-27 at 18:46 +0200, Cédric Le Goater wrote:
> > >
> > > The SCU controler holds the board revision number in its 0x7C
> > > register. Let's use an alias to link a "sil
On 07/28/2016 11:32 PM, Eric Blake wrote:
On 07/28/2016 04:50 AM, Cao jin wrote:
Follow CODING_STYLE
Cc: Daniel P. Berrange
Cc: Gerd Hoffmann
Cc: Paolo Bonzini
Signed-off-by: Cao jin
---
util/qemu-sockets.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
Daniel P. Be
Follow CODING_STYLE
Cc: Daniel P. Berrange
Cc: Gerd Hoffmann
Cc: Paolo Bonzini
Cc: Eric Blake
Signed-off-by: Cao jin
---
util/qemu-sockets.c | 12 +++-
1 file changed, 7 insertions(+), 5 deletions(-)
v2 changelog:
1. fix to Eric's comments.
diff --git a/util/qemu-sockets.c b/util/q
On Fri, 2016-07-29 at 06:19 +0530, Richard Henderson wrote:
> This, I'm not so keen on.
>
> (1) The helper, since it writes to registers controlled by tcg, must be
> described to clobber all registers. Which will noticeably increase memory
> traffic to ENV. For instance, you won't be able to h
On Sat, Jul 9, 2016 at 4:38 AM, Richard Henderson wrote:
> From: Samuel Damashek
>
> As it currently stands, QEMU does not properly handle self-modifying code
> when the write is unaligned and crosses a page boundary. The procedure
> for handling a write to the current translation block is to wri
The following changes since commit 2d2e632ad00d11867c6c5625605b1fbc022dd62f:
Update version for v2.7.0-rc0 release (2016-07-22 15:32:42 +0100)
are available in the git repository at:
git://git.kernel.org/pub/scm/virt/kvm/mst/qemu.git tags/for_upstream
for you to fetch changes up to f077f889
From: Marcel Apfelbaum
Specify the root port interrupt pin as part of the init
process for cases when msi/msix are not enabled.
Fixes "hw/pci/pci.c:196:23: runtime error: shift exponent -1 is negative"
warning from clang's sanitizer.
Reported-by: Peter Maydell
Signed-off-by: Marcel Apfelbaum
From: Marcel Apfelbaum
Prevent future issues when hotplug will work for devices
attached to pxbs.
Suggested-by: Igor Mammedov
Signed-off-by: Marcel Apfelbaum
Tested-by: Laszlo Ersek
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/pci-bridge/pci_expander_bridge.c |
From: Marcel Apfelbaum
PXBs do not support hotplug so they don't have a PCNT function.
Since the PXB's PCI root-bus is a child bus of bus 0, the
build_dsdt code will add a call to the corresponding PCNT function.
Fix this by skipping the PCNT call for the above case.
While at it skip also PCIe c
We changed link status register in pci express endpoint capability
over time. Specifically,
commit b2101eae63ea57b571cee4a9075a4287d24ba4a4 ("pcie: Set the "link
active" in the link status register") set data link layer link active
bit in this register without adding compatibility to old machine t
From: Marcel Apfelbaum
Enable transitional virtio devices by default.
Enable virtio-1.0 for devices plugged into
PCIe ports (Root ports or Downstream ports).
Using the virtio-1 mode will remove the limitation
of the number of devices that can be attached to a machine
by removing the need for the
From: Marcel Apfelbaum
Instead of always passing both IO and MEM ranges when
computing CRS ranges, define a new CrsRangeSet structure
that include them both.
This is done before introducing a third type of range,
64-bit MEM, so it will be easier to pass them all around.
Reviewed-by: Igor Mammed
From: Wei Jiangang
Firstly, convert pxb_dev_init_common() to Error and rename
it to pxb_dev_realize_common().
Actually, pxb_register_bus() is converted as well.
And then,
convert pxb_dev_initfn() and pxb_pcie_dev_initfn() to Error,
rename them to pxb_dev_realize() and pxb_pcie_dev_realize()
resp
From: Marc-André Lureau
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/net/vhost_net.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
index f92d3f8..3677a82 100644
--- a/hw/
From: Marcel Apfelbaum
In build_crs(), the calculation and merging of the ranges already happens
in 64-bit, but the entry boundaries are silently truncated to 32-bit in the
call to aml_dword_memory(). Fix it by handling the 64-bit MMIO ranges
separately.
This fixes 64-bit BARs behind PXBs.
Repo
From: Wei Jiangang
Convert a device model where initialization obviously can't fail,
make it implement realize() rather than init().
Signed-off-by: Wei Jiangang
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
Reviewed-by: Markus Armbruster
---
hw/pci-host/apb.c | 5 ++---
From: Marc-André Lureau
Shorten the code and make it more clear by using the specialized
function g_str_has_prefix().
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
net/vhost-user.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-
From: Prasad J Pandit
virtio back end uses set of buffers to facilitate I/O operations.
An infinite loop unfolds in virtqueue_pop() if a buffer was
of zero size. Add check to avoid it.
Reported-by: Li Qiang
Signed-off-by: Prasad J Pandit
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael
From: Marc-André Lureau
Although not strictly required, it is nice to have vhost_log_put()
safely callable multiple times.
Clear dev->log* when calling vhost_log_put() to make the function
idempotent. This also simplifies a bit the caller work.
Signed-off-by: Marc-André Lureau
Reviewed-by: Mic
From: Marc-André Lureau
vhost-dev opaque isn't necessarily an fd, it can be a chardev when using
vhost-user. Goto fail, so vhost_backend_cleanup() is called to handle
backend cleanup appropriately.
vhost_set_backend_type() should never fail, use an assert().
Signed-off-by: Marc-André Lureau
Re
From: Marc-André Lureau
In some cases, qemu_chr_fe_read_all() on HUP event doesn't raise
CHR_EVENT_CLOSED because the read/recv function returns -1 on
disconnected peers (for example with tch_chr_recv, an ECONNRESET errno
overwritten as EIO).
It is simpler to explicitely disconnect on HUP, risin
From: Marc-André Lureau
All these functions must be called only after the backend is connected.
They are called from virtio-net.c, after either virtio or link status
change.
The check for nc->peer->link_down should ensure vhost_net_{start,stop}()
are always called between vhost_user_{start,stop}
From: Marc-André Lureau
It is called on multiple code path, so make it safe to call several
times (note: I don't remember a reproducer here, but a function called
'cleanup' should probably be idempotent in my book)
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by:
From: Marc-André Lureau
Add missing VHOST_OPS_DEBUG() logs, for completeness.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/virtio/vhost.c | 17 +++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/hw/vir
From: Marc-André Lureau
Make sure the log was released on cleanup, or it will leak (the
alternative is to call vhost_log_put() unconditionally, but it may hide
some dev state issues).
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/vi
From: Marc-André Lureau
Add a chr_wait_connected for the tcp backend, and use it in the
open_socket() function.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
qemu-char.c | 63 ++--
From: Marc-André Lureau
Calling a vhost operation may fail, for example with disconnected
vhost-user backend, but qemu shouldn't abort in this case.
Log an error instead, except on error and cleanup code paths where it
can be mostly ignored.
Let's use a VHOST_OPS_DEBUG macro to easily disable t
From: Marc-André Lureau
If vhost_dev_init() failed, caller may still call vhost_dev_cleanup()
later. However, vhost_dev_cleanup() tries to remove the device from the
list even if it wasn't yet added, which may lead to crashes. Similarly
for the memory listener.
Signed-off-by: Marc-André Lureau
From: Marc-André Lureau
vhost_net_init() calls vhost_dev_init() and in case of failure, calls
vhost_dev_cleanup() directly. However, the structure is already
partially cleaned on error. Calling vhost_dev_cleanup() again will call
vhost_virtqueue_cleanup() on already clean queues, and causing pote
From: Marc-André Lureau
Either the callback is mandatory to implement, in which case an assert()
is more appropriate, or it's not and we can't tell much whether the
function should fail or not (given it's name, I guess it should silently
success by default). Instead, make the implementation manda
From: Marc-André Lureau
vhost_dev_init(), calling vhost backend initialization, should be
cleaned up after failure too. Call vhost_dev_cleanup() in all failure
cases. First, it needs to zero-alloc the struct to avoid the initial
garbage.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S.
From: Marc-André Lureau
Spotted by valgrind.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
tests/vhost-user-test.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/tests/vhost-user-test.c b/tests/vhost-user-test.c
index 46d0588..27b
From: Marc-André Lureau
Let's use qemu proper error reporting API, this ensures the error is
reported at the right place (stderr or monitor), with a conventional
format.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/virtio/vhost.c |
From: Marc-André Lureau
A function to wait on the backend to be connected, to be used in the
following patches.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
include/sysemu/char.h | 8
qemu-char.c | 9 +
2 fi
From: Marc-André Lureau
Calling qemu_chr_fe_set_msgfds() on unconnected socket leads to crash
since s->ioc is NULL in this case. Return an error earlier instead.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
qemu-char.c | 10 ++
From: Marc-André Lureau
Many code paths assume get_vhost_net() returns non-null.
Keep VhostUserState.vhost_net after a successful vhost_net_init(),
instead of freeing it in vhost_net_cleanup().
VhostUserState.vhost_net is thus freed before after being recreated or
on final vhost_user_cleanup()
On Thu, Jul 28, 2016 at 07:27:02PM +0200, Cédric Le Goater wrote:
> Hello,
>
> On 07/26/2016 08:23 AM, David Gibson wrote:
> > On Mon, Jul 25, 2016 at 04:24:43PM +0200, Cédric Le Goater wrote:
> >> From: Benjamin Herrenschmidt
> >>
> >> No devices yet, not even an interrupt controller, just to ge
From: Marc-André Lureau
It is fine to call set_msgfds() with 0 fd, and ensures any previous fd
array is cleared.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/virtio/vhost-user.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions
From: Marc-André Lureau
Not all vhost-user backends support ops->vhost_net_set_backend(). It is
a nicer to provide an assert/error than to crash trying to
call. Furthermore, it improves a bit the code by hiding vhost_ops
details.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
On Fri, Jul 29, 2016 at 12:13:01PM +1000, Benjamin Herrenschmidt wrote:
> On Fri, 2016-07-29 at 06:19 +0530, Richard Henderson wrote:
> > This, I'm not so keen on.
> >
> > (1) The helper, since it writes to registers controlled by tcg, must be
> > described to clobber all registers. Which will n
From: Marc-André Lureau
Add a few assertions to be more explicit about the runtime behaviour
after the previous patch: get_vhost_net() is non-null after
net_vhost_user_init().
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/net/vhost_
From: Cao jin
My previous commit 2e2aa316 removed internal flag msi_in_use, which
exists in vmstate, use VMSTATE_UNUSED for migration compatibility.
Reported-by: Amit Shah
Suggested-by: Amit Shah
Cc: Markus Armbruster
Cc: Marcel Apfelbaum
Cc: Paolo Bonzini
Cc: Michael S. Tsirkin
Cc: Amit S
From: Marc-André Lureau
Check qemu_chr_fe_set_msgfds() for errors, to make sure the message to
be sent is correct.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/virtio/vhost-user.c | 4 +++-
1 file changed, 3 insertions(+), 1 deleti
From: Marc-André Lureau
Now that get_vhost_net() returns non-null after a successful
vhost_net_init(), we no longer need to check this case.
This reverts commit ecd34898596c60f79886061618dd7e01001113ad.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S.
On Wed, Jul 27, 2016 at 01:14:54AM +0400, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Hi,
>
> Since 'vhost-user: simple reconnection support' has been merged, it is
> possible to disconnect and reconnect a vhost-user backend. However,
> many code paths in qemu may trigger as
From: Marc-André Lureau
The vhost-user code is quite inconsistent with error handling. Instead
of ignoring some return values of read/write and silently going on with
invalid state (invalid read for example), break the code flow when the
error happened.
Signed-off-by: Marc-André Lureau
Reviewed
From: Marc-André Lureau
Found thanks to valgrind.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
tests/virtio-net-test.c | 12 +---
1 file changed, 9 insertions(+), 3 deletions(-)
diff --git a/tests/virtio-net-test.c b/tests/vi
On Fri, Jul 29, 2016 at 08:33:45AM +1000, Benjamin Herrenschmidt wrote:
> float_invalid_op_excp() and float_check_status() must be inline
> due to their use of GETPC().
>
> Use __attribute__((__always_inline__)) to enforce it
>
> Signed-off-by: Benjamin Herrenschmidt
Do you want me to fold this
From: Marc-André Lureau
The chardev waits for an initial connection before starting qemu, and
vhost-user should wait for the backend negotiation to be completed
before starting qemu too.
vhost-user is started in the net_vhost_user_event callback, which is
synchronously called after the socket is
From: Marc-André Lureau
Similar to vhost_user_read() error report, it is useful to have early
error report.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/virtio/vhost-user.c | 13 ++---
1 file changed, 10 insertions(+), 3 de
From: Marc-André Lureau
The state.num value will probably be 0 in this case, but that
doesn't make sense to update.
Signed-off-by: Marc-André Lureau
Reviewed-by: Michael S. Tsirkin
Signed-off-by: Michael S. Tsirkin
---
hw/virtio/vhost.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
Hi Xin,
Thanks for your comments firstly...
Regards,
-Gonglei
> -Original Message-
> From: Zeng, Xin [mailto:xin.z...@intel.com]
> Sent: Thursday, July 28, 2016 1:29 PM
> To: Gonglei (Arei); Michael S. Tsirkin
> Cc: qemu-devel@nongnu.org; virtio-...@lists.oasis-open.org; Ola Liljedahl;
On Thu, Jul 28, 2016 at 06:22:05PM +0530, Richard Henderson wrote:
> On 07/28/2016 12:19 PM, Nikunj A Dadhania wrote:
> > +r->element[i] = abs(a->element[i] - b->element[i]); \
> > +} \
> > +}
> > +
> > +/* VA
On Fri, 2016-07-29 at 12:44 +1000, David Gibson wrote:
> On Fri, Jul 29, 2016 at 08:33:45AM +1000, Benjamin Herrenschmidt
> wrote:
> >
> > float_invalid_op_excp() and float_check_status() must be inline
> > due to their use of GETPC().
> >
> > Use __attribute__((__always_inline__)) to enforce it
On Fri, 2016-07-29 at 13:34 +1000, David Gibson wrote:
>
> What should I do with this in the short term? Leave it in
> ppc-for-2.8, or remove it for now pending possible changes?
I think I'm still measuring a performance improvement with this, I'll
test a bit more and will get back to you.
It w
201 - 300 of 321 matches
Mail list logo
