On Wed, Aug 16, 2017 at 05:46:18PM -0400, John Snow wrote:
>
>
> On 08/14/2017 05:57 PM, Eduardo Habkost wrote:
> > Example output when using "-machine q35":
> >
> > {
> > "available": true,
> > "count": 1,
> > "device-types": [
> > "ide-device"
> > ],
> >
On 08/18/2017 01:51 PM, Philippe Mathieu-Daudé wrote:
On 08/18/2017 11:15 AM, Mark Cave-Ayland wrote:
Separate out the standard ethernet CRC32 calculation into a new
net_crc32()
function, renaming the constant POLYNOMIAL to POLYNOMIAL_BE to make it
clear
that this is a big-endian CRC32
Changes v1->v2:
- Style fixes to make checkpatch.pl happy.
- Rebased.
Changes v2->v3:
- Fix typo in patch 3 ("qemu.py: make 'args' public") commit message.
Changes v3->v4:
- Squash the 2 first commits since they are co-dependant.
- Cleanup launch() and shutdown().
- Reorder the commits,
On Fri, Aug 18, 2017 at 12:08:38PM +0200, Igor Mammedov wrote:
> Move cpu_model +-feat parsing into a separate file so that it
> could be reused later for parsing similar format of sparc target
>
> Signed-off-by: Igor Mammedov
[...]
> diff --git a/include/qom/cpu.h
On Fri, Aug 18, 2017 at 12:08:35PM +0200, Igor Mammedov wrote:
> SPARC is the last target that uses legacy way of parsing
> and initializing cpu features, drop legacy approach and
> convert features to properties so that SPARC could as minimum
> benefit from generic cpu_generic_init(), common with
The 'm->numa_auto_assign_ram = numa_legacy_auto_assign_ram;' line
was supposed to be in pc_i440fx_2_9_machine_options() (see commit
3bfe5716 "numa: equally distribute memory on nodes"), but the
merge commit adb354dd ("Merge remote-tracking branch
'mst/tags/for_upstream' into staging") moved it to
Now that shutdown() is guaranteed to always execute self._load_io_log()
and self._post_shutdown(), their calls in 'except' became redundant and
we can safely replace it by a call to shutdown().
Signed-off-by: Amador Pahim
---
scripts/qemu.py | 6 +-
1 file changed, 1
On Thu, Aug 17, 2017 at 11:26:55AM +0800, Wei Wang wrote:
> This patch adds support to walk through the free page blocks in the
> system and report them via a callback function. Some page blocks may
> leave the free list after zone->lock is released, so it is the caller's
> responsibility to
On Wed, Aug 16, 2017 at 09:26:51AM +0800, Dou Liyang wrote:
> Currently, Using the fisrt node without memory on the machine makes
> QEMU unhappy. With this example command line:
> ... \
> -m 1024M,slots=4,maxmem=32G \
> -numa node,nodeid=0 \
> -numa node,mem=1024M,nodeid=1 \
> -numa
On Fri, Aug 18, 2017 at 09:35:00PM +0300, Michael S. Tsirkin wrote:
> On Wed, Aug 16, 2017 at 09:45:35AM +0800, Dou Liyang wrote:
> > Currently, Using the fisrt node without memory on the machine makes
> > QEMU unhappy. With this example command line:
> > ... \
> > -m 1024M,slots=4,maxmem=32G
On Thu, Aug 17, 2017 at 01:58:40PM +0800, Dou Liyang wrote:
> Hi Igor,
>
> I tested this patch with following guests:
>
> 1. RHEL 6.5 with Linux 2.6.32
> 2. RHEL 7.0 with Linux 3.10.0
> 3. Fedora 23 with Linux 4.13.0-rc5
> 4. window 2003 service
> 5. window 7
> 6. window 10
What's the
On 08/18/2017 11:15 AM, Mark Cave-Ayland wrote:
This provides a standard ethernet CRC32 little-endian implementation.
Signed-off-by: Mark Cave-Ayland
---
include/net/net.h |2 ++
net/net.c | 22 ++
2 files changed, 24
On 18.08.2017 18:08, Thomas Huth wrote:
> On 18.08.2017 13:43, David Hildenbrand wrote:
>> s390-stattrib.c needs definition of TARGET_PAGE_SIZE, solve it via cpu.h.
>
> Why not simply #include "exec/cpu-all.h" ? If cpu.h is not really needed
> here...
>
> Thomas
>
>
Git grep showed me that
As discussed with Paolo and Markus let's combine libqemustub.a into
libqemuutil.a to avoid circular dependencies.
Alistair Francis (2):
Makefile: Remove libqemustub.a
Convert remaining single line fprintf() to warn_report()
Makefile| 7 +++
Makefile.target
On Fri, Aug 18, 2017 at 09:28:55PM +0300, Michael S. Tsirkin wrote:
> On Wed, Aug 16, 2017 at 09:26:51AM +0800, Dou Liyang wrote:
> > Currently, Using the fisrt node without memory on the machine makes
> > QEMU unhappy. With this example command line:
> > ... \
> > -m 1024M,slots=4,maxmem=32G
On Fri, Aug 18, 2017 at 12:29 PM, Philippe Mathieu-Daudé
wrote:
> Hi Alistair,
>
>
> On 08/18/2017 03:47 PM, Alistair Francis wrote:
>>
>> Using two libraries (libqemuutil.a and libqemustub.a) would sometimes
>> result in circular dependencies. To avoid these issues let's just
>>
On Wed, Aug 16, 2017 at 08:25:41AM +0200, Markus Armbruster wrote:
> Eduardo Habkost writes:
>
> > On Tue, Aug 15, 2017 at 12:03:53PM +0200, Markus Armbruster wrote:
> >> Eduardo Habkost writes:
> >>
> >> Suggest to insert here:
> >>
> >> If
The current message shows 'self._args', which contains only part of the
options used in the Qemu command line.
This patch makes the qemu full args list an instance variable and then
uses it in the negative exit code message.
Message was moved outside the 'if is_running' block to make sure it
This is just an refactor to separate the exception handler from the
actual launch procedure, improving the readability and making future
maintenances in this piece of code easier.
Signed-off-by: Amador Pahim
---
scripts/qemu.py | 33 -
1 file
For increased portability, let's use os.path.devnull.
Signed-off-by: Amador Pahim
---
scripts/qemu.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/scripts/qemu.py b/scripts/qemu.py
index ef531bb23b..51545f7f97 100644
--- a/scripts/qemu.py
+++
is_running() returns None when called before the first time we
call launch():
>>> import qemu
>>> vm = qemu.QEMUMachine('qemu-system-x86_64')
>>> vm.is_running()
>>>
It should return False instead. This patch fixes that.
For consistence, this patch removes the parenthesis from
On 18.08.2017 18:11, Thomas Huth wrote:
> Suggesting to add a patch description like: "The functions are not used
> in target/s390x/ so a header in hw/s390x/ is a better place" ?
Sure, I will include that, thanks!
>
> On 18.08.2017 13:43, David Hildenbrand wrote:
>> Signed-off-by: David
On 18/08/17 18:06, Philippe Mathieu-Daudé wrote:
> On 08/18/2017 01:51 PM, Philippe Mathieu-Daudé wrote:
>> On 08/18/2017 11:15 AM, Mark Cave-Ayland wrote:
>>> Separate out the standard ethernet CRC32 calculation into a new
>>> net_crc32()
>>> function, renaming the constant POLYNOMIAL to
On Fri, Aug 18, 2017 at 04:41:41PM +0800, Wei Wang wrote:
> On 08/18/2017 10:13 AM, Michael S. Tsirkin wrote:
> > On Thu, Aug 17, 2017 at 11:26:56AM +0800, Wei Wang wrote:
> > > Add a new vq to report hints of guest free pages to the host.
> > Please add some text here explaining the
* Michael S. Tsirkin (m...@redhat.com) wrote:
> On Wed, Jun 28, 2017 at 08:00:43PM +0100, Dr. David Alan Gilbert (git) wrote:
> > From: "Dr. David Alan Gilbert"
> >
> > When we receive a GET_VRING_BASE message set enable = false
> > to stop any new received packets modifying
On 08/18/2017 09:36 AM, Programmingkid wrote:
>
>> On Aug 18, 2017, at 4:46 AM, Artyom Tarasenko wrote:
>>
>> On Fri, Aug 18, 2017 at 6:31 AM, Programmingkid
>> wrote:
>>>
On Aug 15, 2017, at 6:27 PM, Paolo Bonzini
> On Aug 18, 2017, at 3:23 PM, John Snow wrote:
>
>
>
> On 08/18/2017 09:36 AM, Programmingkid wrote:
>>
>>> On Aug 18, 2017, at 4:46 AM, Artyom Tarasenko wrote:
>>>
>>> On Fri, Aug 18, 2017 at 6:31 AM, Programmingkid
>>>
Hi Igor,
On 08/17/2017 12:06 PM, Philippe Mathieu-Daudé wrote:
no logical change, only code movement (and fix a comment typo).
while at that fix checkpatch errors or
first fix checkpatch noted issues in cpu.h and then move it in next
patch to internal.h
After asking confirmation on IRC
When launching a VM, if an exception happens and the VM is not
initiated, it might be useful to see the qemu command line and
the qemu command output.
This patch creates that message. Notice that self._iolog needs to be
cleaned up in the beginning of the launch() to make sure we will not
expose
On 08/18/2017 11:15 AM, Mark Cave-Ayland wrote:
Signed-off-by: Mark Cave-Ayland
---
hw/net/eepro100.c | 19 +--
1 file changed, 1 insertion(+), 18 deletions(-)
diff --git a/hw/net/eepro100.c b/hw/net/eepro100.c
index 5a4774a..4226572 100644
If a VM is launched, files are created and a cleanup is required before
a new launch. This cleanup is executed by shutdown(), so shutdown() must
be called even if the VM is manually terminated (i.e. using kill).
This patch creates a control to make sure launch() will not be executed
again if
The 'returncode' Popen attribute is not guaranteed to be updated. It
actually depends on a call to either poll(), wait() or communicate().
On the other hand, poll() will: "Check if child process has terminated.
Set and return returncode attribute."
Let's use the poll() to check whether the
A new call to launch() with a running VM will fall in exception and
consequently call shutdown().
This patch makes launch() to raise an exception when it's called with VM
already running.
Signed-off-by: Amador Pahim
---
scripts/qemu.py | 3 +++
1 file changed, 3
On 08/18/2017 02:32 AM, Markus Armbruster wrote:
Philippe Mathieu-Daudé writes:
On 08/17/2017 02:55 PM, Alistair Francis wrote:
On 15/08/2017 09:30, Markus Armbruster wrote:
The stupid fix is to repeat libraries until the link succeeds:
test-util-obj-y =
On Tue, Aug 15, 2017 at 08:06:18PM -0400, Lan Tianyu wrote:
> Signed-off-by: Lan Tianyu
> ---
> hw/i386/pc_piix.c| 14 +++---
> hw/i386/pc_q35.c | 14 +++---
> include/hw/compat.h | 3 +++
> include/hw/i386/pc.h | 3 +++
> 4 files changed, 28
On Thu, Aug 17, 2017 at 09:37:10AM +0800, Lan Tianyu wrote:
> On 2017年08月16日 19:21, Paolo Bonzini wrote:
> > On 16/08/2017 02:22, Lan Tianyu wrote:
> >> Xen vIOMMU device model will be in Xen hypervisor. Skip vIOMMU
> >> check for Xen here when vcpu number is more than 255.
> >
> > I think you
On 18/08/2017 18:38, Eduardo Habkost wrote:
>>> I think you still need to do a check for vIOMMU being enabled.
>>
>> Yes, this will be done in the Xen tool stack and Qemu doesn't have such
>> knowledge. Operations of create, destroy Xen vIOMMU will be done in the
>> Xen tool stack.
>
> Shouldn't
On 08/18/2017 01:59 PM, Philippe Mathieu-Daudé wrote:
On 08/18/2017 11:15 AM, Mark Cave-Ayland wrote:
Signed-off-by: Mark Cave-Ayland
---
hw/net/pcnet.c | 16 +---
1 file changed, 1 insertion(+), 15 deletions(-)
diff --git a/hw/net/pcnet.c
On Thu, Aug 17, 2017 at 10:32 PM, Markus Armbruster wrote:
> Philippe Mathieu-Daudé writes:
>
>> On 08/17/2017 02:55 PM, Alistair Francis wrote:
> On 15/08/2017 09:30, Markus Armbruster wrote:
>> The stupid fix is to repeat libraries until the link
To launch a VM, we need to create basically two files: the monitor
socket (if it's a UNIX socket) and the qemu log file.
For the qemu log file, we currently just open the path, which will
create the file if it does not exist or overwrite the file if it does
exist.
For the monitor socket, if it
This module should not write directly to stdout/stderr. Instead, it
should either raise exceptions or just log the messages and let the
callers handle them and decide what to do. For example, scripts could
choose to send the log messages stderr or/and write them to a file if
verbose or debugging
On Thu, Aug 17, 2017 at 04:07:56PM +0200, Igor Mammedov wrote:
> Since
> (commit d4a606b3 i386: Don't override -cpu options on -cpu host/max)
> it became possible to delete hack where it was necessary to
> postpone applying plus/minus features to realize time
> after max_features were applied to
On Fri, Aug 18, 2017 at 04:19:57PM +0200, Igor Mammedov wrote:
> > > >
> > > > Clean it up after 2.10.
> > > >
> >
> > So is the v2 good enough or do I need to resend it?
> Do you really need it in 2.10?
> it's only 2 days left till release so unless it's blocker
> I'd wait till after release
On Fri, Aug 18, 2017 at 04:09:43PM -0300, Eduardo Habkost wrote:
> The 'm->numa_auto_assign_ram = numa_legacy_auto_assign_ram;' line
> was supposed to be in pc_i440fx_2_9_machine_options() (see commit
> 3bfe5716 "numa: equally distribute memory on nodes"), but the
> merge commit adb354dd ("Merge
On 08/18/2017 03:47 PM, Alistair Francis wrote:
Convert any remaining uses of fprintf(stderr, "warning:"...
to use warn_report() instead. This helps standardise on a single
method of printing warnings to the user.
All of the warnings were changed using this command:
find ./* -type f -exec
At least in some configurations, setting c-file-style is not enough to
conform to the QEMU coding style, so explicitly set c-basic-offset as well.
Signed-off-by: Thiago Jung Bauermann
---
My emacs was using indentation level of 8 spaces and this patch convinced
it
Using two libraries (libqemuutil.a and libqemustub.a) would sometimes
result in circular dependencies. To avoid these issues let's just
combine both into a single library that functions as both.
Signed-off-by: Alistair Francis
---
Makefile| 7
Convert any remaining uses of fprintf(stderr, "warning:"...
to use warn_report() instead. This helps standardise on a single
method of printing warnings to the user.
All of the warnings were changed using this command:
find ./* -type f -exec sed -i 's|fprintf(.*".*warning[,:] |warn_report("|Ig'
hi Balaton,
> I know about this and if you read the cover message (RFC PATCH 00/12) I link
> to a fixed version of that U-Boot firmware (rebuilt from source with the
> mentioned patches) which works with this emulation. (The original one from
> the updater also starts but cannot boot due to
On 08/18/2017 11:15 AM, Mark Cave-Ayland wrote:
Separate out the standard ethernet CRC32 calculation into a new net_crc32()
function, renaming the constant POLYNOMIAL to POLYNOMIAL_BE to make it clear
that this is a big-endian CRC32 calculation.
Then remove the existing implementation from
On Wed, Aug 16, 2017 at 09:45:35AM +0800, Dou Liyang wrote:
> Currently, Using the fisrt node without memory on the machine makes
> QEMU unhappy. With this example command line:
> ... \
> -m 1024M,slots=4,maxmem=32G \
> -numa node,nodeid=0 \
> -numa node,mem=1024M,nodeid=1 \
> -numa
I succeed booting a SPARC image once applying 1 to 8 (this commit),
so not being able to boot the image previously is probably due to patch
order, which mean current order is not bisect-able.
On 08/18/2017 07:08 AM, Igor Mammedov wrote:
it's just a wrapper, drop it and use cpu_generic_init()
On Thu, Aug 17, 2017 at 04:32:23PM +0200, Igor Mammedov wrote:
> On Wed, 16 Aug 2017 16:56:40 -0300
> Eduardo Habkost wrote:
>
> > On Fri, Jul 14, 2017 at 03:52:09PM +0200, Igor Mammedov wrote:
> > > call xtensa_irq_init() at realize time which makes
> > > cpu_xtensa_init()
On 08/18/2017 11:15 AM, Mark Cave-Ayland wrote:
Signed-off-by: Mark Cave-Ayland
---
hw/net/pcnet.c | 16 +---
1 file changed, 1 insertion(+), 15 deletions(-)
diff --git a/hw/net/pcnet.c b/hw/net/pcnet.c
index 6544553..c050993 100644
---
The allwinner-a10 device uses serial_hds[0] without checking whether
it is available or not. So using the cubieboard with -nodefaults
currently results in a segmentation fault. Fix it by adding a
proper check here.
And while we're at it, mark the device as "user_creatable = false"
since this
On 18 August 2017 at 18:08, Thomas Huth wrote:
> The allwinner-a10 device uses serial_hds[0] without checking whether
> it is available or not. So using the cubieboard with -nodefaults
> currently results in a segmentation fault. Fix it by adding a
> proper check here.
> And
On Fri, Aug 18, 2017 at 04:36:06PM +0800, Wei Wang wrote:
> On 08/18/2017 10:28 AM, Michael S. Tsirkin wrote:
> > On Thu, Aug 17, 2017 at 11:26:56AM +0800, Wei Wang wrote:
> > > Add a new vq to report hints of guest free pages to the host.
> > >
> > > Signed-off-by: Wei Wang
The 'host doesn't support requested feature' is probably a red-herring in this
case
The fact it's failing with an IO error but nothing else suggests either:
a) it's something closing the socket between the two qemu's
b) The I/O error is from storage/NBD
Assuming it fails on precopy, I'd look
On Fri, Aug 18, 2017 at 05:00:18PM +0100, Anthony PERARD wrote:
> > > > >
> > > > > Clean it up after 2.10.
> > > > >
> > >
> > > So is the v2 good enough or do I need to resend it?
> > Do you really need it in 2.10?
> > it's only 2 days left till release so unless it's blocker
> > I'd wait
Hi Igor,
On 08/18/2017 07:08 AM, Igor Mammedov wrote:
QOMfy cpu models handling introducing propper cpu types
for each cpu model.
Signed-off-by: Igor Mammedov
---
with this and conversion of features to properties,
it would be possible to replace cpu_sparc_init() with
Hi Alistair,
On 08/18/2017 03:47 PM, Alistair Francis wrote:
Using two libraries (libqemuutil.a and libqemustub.a) would sometimes
result in circular dependencies. To avoid these issues let's just
combine both into a single library that functions as both.
Signed-off-by: Alistair Francis
On 08/17/2017 12:22 PM, Igor Mammedov wrote:
On Thu, 17 Aug 2017 01:30:54 -0300
Philippe Mathieu-Daudé wrote:
[...]
Also I couldn't test it with KVM.
Tested in TCG mode (boots debian mips/mips64 kernel with different cpu types),
and new CPU leaf types show up on QOM tree as
Le 18/08/2017 à 21:43, luigi burdo a écrit :
> can i ask you something ? why you dont try to integrate in qemu the
> pegasos 2 or the efika machine. i have the feeling that probably it
> can be more simple because more old machine and components.
>
Except they are antique, and the Pegasos at
Don't open-code something that has a convenient helper available.
Signed-off-by: Eric Blake
Reviewed-by: Markus Armbruster
---
tests/numa-test.c | 21 +++--
1 file changed, 3 insertions(+), 18 deletions(-)
diff --git a/tests/numa-test.c
On 08/18/2017 05:15 PM, Eric Blake wrote:
> libqtest provides two layers of functions: qtest_*() that operate
> on an explicit object, and a plain version that operates on the
> 'global_qtest' object. However, very few tests care about the
> distinction, and even the tests that manipulate
On Aug 16, 2017 11:41 PM, "Markus Armbruster" wrote:
Eric Blake writes:
> On 08/16/2017 11:13 AM, Markus Armbruster wrote:
>> Markus Armbruster writes:
>>
>
>>>
>>> Conclusion: no consensus, yet.
>>
>> All right, let's start over and
On 19/08/17 12:46, Alexey Kardashevskiy wrote:
> On 19/08/17 01:18, Eric Blake wrote:
>> On 08/18/2017 08:31 AM, Stefan Hajnoczi wrote:
>>> Most qcow2 files are uncompressed so it is wasteful to allocate (32 + 1)
>>> * cluster_size + 512 bytes upfront. Allocate s->cluster_cache and
>>>
This will keep checkpatch happy when the next patch does code motion.
Fix the include order to match HACKING when adding the needed header.
Signed-off-by: Eric Blake
---
tests/libqtest.c | 19 +--
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git
Rather than make multiple callers call strlen(), it's easier if
socket_send() itself can compute a length via strlen() if none
was provided (caller passes -1). Callers that can get at the
length more efficiently are left that way.
Signed-off-by: Eric Blake
---
On 08/18/2017 06:15 PM, Eric Blake wrote:
Prior to commit 063c23d9, we were tracking a list of parallel
qtest objects, in order to safely clean up a SIGABRT handler
only after the last connection quits. But when we switched to
more of glib's infrastructure, the list became dead code that
is
On 08/18/2017 06:15 PM, Eric Blake wrote:
Don't open-code something that has a convenient helper available.
Signed-off-by: Eric Blake
Reviewed-by: Markus Armbruster
Reviewed-by: Philippe Mathieu-Daudé
---
tests/numa-test.c | 21
On 18/08/17 23:31, Stefan Hajnoczi wrote:
> Most qcow2 files are uncompressed so it is wasteful to allocate (32 + 1)
> * cluster_size + 512 bytes upfront. Allocate s->cluster_cache and
> s->cluster_data when the first read operation is performance on a
> compressed cluster.
>
> The buffers are
There was only one caller; it's easier to inline things.
Signed-off-by: Eric Blake
---
tests/libqtest.c | 22 ++
1 file changed, 6 insertions(+), 16 deletions(-)
diff --git a/tests/libqtest.c b/tests/libqtest.c
index 5d16351e24..b6dd26e54a 100644
---
qtest_rtas_call() and qtest_big_endian() did not have a short
version with an implied global_qtest; but changing these two
functions fits with the theme of the previous patch.
It doesn't hurt that we are now no longer ambiguous with the
qtest_rtas_call() of include/hw/ppc/spapr_rtos.h.
None of the tests were calling the long qtest_*() form, except via
the static inline short forms. Remove a layer of indirection by
only supporting the short form, and using global_qtest directly in
the .c file. (Yes, this flies in the face of thread-safety, by
relying on a global instead of
On 08/18/2017 04:39 PM, Eric Blake wrote:
> On 08/18/2017 04:33 PM, Philippe Mathieu-Daudé wrote:
>> Hi Eric,
>>
>> On 08/18/2017 06:15 PM, Eric Blake wrote:
>>> Assertions should be separate from the side effects, since in
>>> theory, g_assert() can be disabled (in practice, we can't really
>>>
On 19/08/17 01:18, Eric Blake wrote:
> On 08/18/2017 08:31 AM, Stefan Hajnoczi wrote:
>> Most qcow2 files are uncompressed so it is wasteful to allocate (32 + 1)
>> * cluster_size + 512 bytes upfront. Allocate s->cluster_cache and
>> s->cluster_data when the first read operation is performance on
Markus gave some good advice on my 'v4: Clean up around qmp() and hmp()'
series [1]. Among other things, we agreed that if I'm going to get rid
of the qtest_* layer of function wrappers, I should do it all the way
(rather than just on the qmp() functions), and up front. So that's what
this
Back when the test was introduced, in commit 62c39b307, the
test was set up to run qemu-ga directly on the host performing
the test, and defaults to limiting itself to safe commands. At
the time, it was envisioned that setting QGA_TEST_SIDE_EFFECTING
in the environment could cover a few more
Assertions should be separate from the side effects, since in
theory, g_assert() can be disabled (in practice, we can't really
ever do that).
Signed-off-by: Eric Blake
---
qtest.c | 80 ++---
1 file changed, 57
All callers are now passing global_qtest, just remove the parameter
instead. While at it, improve the naming of the va_list variant
of socket_sendf() to be more like the printf/vprintf naming scheme.
Signed-off-by: Eric Blake
---
tests/libqtest.c | 77
On 08/18/2017 06:15 PM, Eric Blake wrote:
This will keep checkpatch happy when the next patch does code motion.
Fix the include order to match HACKING when adding the needed header.
Signed-off-by: Eric Blake
Reviewed-by: Philippe Mathieu-Daudé
---
On 08/18/2017 06:15 PM, Eric Blake wrote:
There was only one caller; it's easier to inline things.
What's the benefit of this change? Having a separate function ease code
reading. You might add the 'inline' qualifier but the compiler is smart
enough to inline it.
Instead of this change I'd
Prior to commit 063c23d9, we were tracking a list of parallel
qtest objects, in order to safely clean up a SIGABRT handler
only after the last connection quits. But when we switched to
more of glib's infrastructure, the list became dead code that
is never assigned to.
Signed-off-by: Eric Blake
On 08/18/2017 04:58 PM, Eric Blake wrote:
>> NDEBUG doesn't affect g_assert() (only assert(), but that wasn't in use
>> here) - I have to double-check glib documentation to see whether
>> g_assert() can be crippled in a manner similar to how I know assert()
>> can be crippled. Ideally, we have a
We already have several files that knowingly require assert()
to work. While we do NOT want to encourage the use of
'assert(side-effects)' (that is a bad practice that prevents
copy-and-paste of code to other projects that CAN disable
assertions; plus it costs unnecessary reviewer mental cycles
On 08/18/2017 07:23 PM, Eric Blake wrote:
We already have several files that knowingly require assert()
to work. While we do NOT want to encourage the use of
'assert(side-effects)' (that is a bad practice that prevents
copy-and-paste of code to other projects that CAN disable
assertions; plus
Hello Hannes,
On 08/18/17 11:37, Hannes Reinecke wrote:
> According to SPC-3 INQUIRY and REQUEST SENSE should return GOOD
> even on unsupported LUNS.
>
> Signed-off-by: Hannes Reinecke
> ---
> hw/scsi/scsi-bus.c | 29 +
> 1 file changed, 25
Put static functions prior to public headers, in part so that
improvements to qtest_start() can benefit from the static
helpers without needing forward references. Code motion, with
no semantic change.
Signed-off-by: Eric Blake
---
tests/libqtest.c | 263
Most of our tests were using qtest_start() to initialize the
connection and then set global_qtest; the few tests that were
using qtest_init() instead are still setting global_qtest shortly
afterwards. So it makes more sense to just always set
global_qtest, and have all callers go through a single
libqtest provides two layers of functions: qtest_*() that operate
on an explicit object, and a plain version that operates on the
'global_qtest' object. However, very few tests care about the
distinction, and even the tests that manipulate multiple qtest
connections at once are just fine
On 08/18/2017 04:33 PM, Philippe Mathieu-Daudé wrote:
> Hi Eric,
>
> On 08/18/2017 06:15 PM, Eric Blake wrote:
>> Assertions should be separate from the side effects, since in
>> theory, g_assert() can be disabled (in practice, we can't really
>> ever do that).
>
> What about the suggestion on
On 08/18/2017 06:39 PM, Eric Blake wrote:
On 08/18/2017 04:33 PM, Philippe Mathieu-Daudé wrote:
Hi Eric,
On 08/18/2017 06:15 PM, Eric Blake wrote:
Assertions should be separate from the side effects, since in
theory, g_assert() can be disabled (in practice, we can't really
ever do that).
On 18/08/2017 05:38, Fam Zheng wrote:
> On Thu, 08/17 16:23, Paolo Bonzini wrote:
>>> +if (scsi_sense_matches(r, SENSE_CODE(NO_MEDIUM))) {
>>> +error = ENOMEDIUM;
>>> +} else if (scsi_sense_matches(r, SENSE_CODE(TARGET_FAILURE))) {
>>> +error = ENOMEM;
>>> +
Hi Eric,
On 08/18/2017 06:15 PM, Eric Blake wrote:
Assertions should be separate from the side effects, since in
theory, g_assert() can be disabled (in practice, we can't really
ever do that).
What about the suggestion on your "Hacks for building on gcc 7 / Fedora
26" series about avoid
On 08/18/2017 04:46 PM, Philippe Mathieu-Daudé wrote:
> On 08/18/2017 06:15 PM, Eric Blake wrote:
>> There was only one caller; it's easier to inline things.
>
> What's the benefit of this change? Having a separate function ease code
> reading. You might add the 'inline' qualifier but the
On 18.08.2017 09:54, Cornelia Huck wrote:
> On Thu, 17 Aug 2017 16:02:00 +0200
> Thomas Huth wrote:
>
>> On 17.08.2017 10:41, Cornelia Huck wrote:
[...]
>>> I'm wondering if we could unify selection of the network device
>>> somehow. There's probably two cases:
>>> - Test a
On Fri, Aug 18, 2017 at 6:31 AM, Programmingkid
wrote:
>
>> On Aug 15, 2017, at 6:27 PM, Paolo Bonzini wrote:
>>
>> On 15/08/2017 20:46, Programmingkid wrote:
>>>
On Aug 14, 2017, at 2:51 AM, Paolo Bonzini wrote:
* Philipp Hahn (h...@univention.de) wrote:
> Hello,
>
> Am 15.08.2017 um 13:25 schrieb Laszlo Ersek:
> > On 08/14/17 20:39, Dr. David Alan Gilbert wrote:
> >> * Philipp Hahn (h...@univention.de) wrote:
> >>> I'm currently investigating a problem, were a Linux VM does not reboot
> >>> and gets
On Fri, Aug 18, 2017 at 10:23:09AM +0200, Alberto Garcia wrote:
On Fri 18 Aug 2017 05:10:18 AM CEST, Manos Pitsidianakis wrote:
block/throttle.c uses existing I/O throttle infrastructure inside a
block filter driver. I/O operations are intercepted in the filter's
read/write coroutines, and
1 - 100 of 282 matches
Mail list logo