It's not necessary to lock the address translation portion of the
vcpu_mem_access callback.
Signed-off-by: Mahmoud Mandour
---
contrib/plugins/cache.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/contrib/plugins/cache.c b/contrib/plugins/cache.c
index 4a71602639..695fb96
Multicore L1 cache modelling is introduced and is supported for both
full system emulation and linux-user.
For full-system emulation, L1 icache and dcache are maintained for each
available core, since this information is exposed to the plugin through
`qemu_plugin_n_vcpus()`.
For linux-user, a sta
Hello,
This series removes passing arguments to plugins through "arg=" since
it's redundant and reduces readability especially when the argument
itself is composed of a name and a value.
Also, passing arguments through "arg=" still works but is marked as
deprecated and will produce a deprecation
Signed-off-by: Mahmoud Mandour
---
contrib/plugins/hotblocks.c | 14 --
1 file changed, 12 insertions(+), 2 deletions(-)
diff --git a/contrib/plugins/hotblocks.c b/contrib/plugins/hotblocks.c
index 4b08340143..062200a7a4 100644
--- a/contrib/plugins/hotblocks.c
+++ b/contrib/plugins/
This commit makes the plugin adhere to the new plugins arg-passing
scheme by expecting full-form boolean args instead of short-form
booleans. This necessitates that we introduce a new argument, here
"track", to accept "r", "w", or "rw".
Also, it makes arguments not positional and we only care abou
On 2021/7/20 14:57, Cornelia Huck wrote:
On Mon, Jul 19 2021, Yanan Wang wrote:
Currently the only difference between smp_parse and pc_smp_parse
is the support of multi-dies and the related error reporting code.
With an arch compat variable "bool smp_dies_supported", we can
easily make smp_par
This call will help boolean argument parsing since arguments are now
passed to plugins as a name and value.
Signed-off-by: Mahmoud Mandour
---
include/qemu/qemu-plugin.h | 13 +
plugins/api.c | 5 +
2 files changed, 18 insertions(+)
diff --git a/include/qemu/qemu-p
Passing arguments to plugins had to be done through "arg=".
This is redundant and introduces confusion especially when the argument
has a name and value (e.g. `-plugin plugin_name,arg="argname=argvalue"`).
This allows passing plugin arguments directly e.g:
`-plugin plugin_name,argname=argvalu
Signed-off-by: Mahmoud Mandour
---
tests/plugin/syscall.c | 23 ---
1 file changed, 16 insertions(+), 7 deletions(-)
diff --git a/tests/plugin/syscall.c b/tests/plugin/syscall.c
index 6dd71092e1..484b48de49 100644
--- a/tests/plugin/syscall.c
+++ b/tests/plugin/syscall.c
@@ -
Signed-off-by: Mahmoud Mandour
---
contrib/plugins/lockstep.c | 31 ++-
docs/devel/tcg-plugins.rst | 2 +-
2 files changed, 23 insertions(+), 10 deletions(-)
diff --git a/contrib/plugins/lockstep.c b/contrib/plugins/lockstep.c
index 7fd35eb669..a41ffe83fa 100644
---
Parsing boolean arguments correctly (e.g. pattern=on or source=false).
Introduced a new "track" argument that takes a [read|write] value. This
substitutes passing read or write to "arg=" that is deprecated.
Also, matches are now taken one by one through the "match" argument.
Signed-off-by: Mahmou
Since plugin arguments now expect boolean arguments, a plugin argument
name "sortby" now expects a value of "read", "write", or "address".
"io" arg is now expected to be passed as a full-form boolean parameter,
i.e. "io=on|true|yes|off|false|no"
Signed-off-by: Mahmoud Mandour
---
contrib/plugin
Signed-off-by: Mahmoud Mandour
---
tests/plugin/bb.c | 15 +++
1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/tests/plugin/bb.c b/tests/plugin/bb.c
index de09bdde4e..7d470a1011 100644
--- a/tests/plugin/bb.c
+++ b/tests/plugin/bb.c
@@ -104,10 +104,17 @@ QEMU_PLUGIN_EXP
Correctly parsing plugin argument since they now must be provided as
full-form boolean parameters, e.g.:
-plugin ./contrib/plugins/libhowvec.so,verbose=on,inline=on
Also, introduced the argument "count" that accepts one opt to count
individually at a time.
Signed-off-by: Mahmoud Mandour
---
Signed-off-by: Mahmoud Mandour
---
docs/devel/tcg-plugins.rst | 18 +-
1 file changed, 9 insertions(+), 9 deletions(-)
diff --git a/docs/devel/tcg-plugins.rst b/docs/devel/tcg-plugins.rst
index 9377bc51d8..7b1a3921b6 100644
--- a/docs/devel/tcg-plugins.rst
+++ b/docs/devel/tcg-pl
Made argument "inline" not positional, this has two benefits. First is
that we adhere to how QEMU passes args generally, by taking the last
value of an argument and drop the others. And the second is that this
sets up a framework for potentially adding new args easily.
Signed-off-by: Mahmoud Mando
Signed-off-by: Mahmoud Mandour
---
docs/system/deprecated.rst | 7 +++
1 file changed, 7 insertions(+)
diff --git a/docs/system/deprecated.rst b/docs/system/deprecated.rst
index e2e0090878..7ae6f1f727 100644
--- a/docs/system/deprecated.rst
+++ b/docs/system/deprecated.rst
@@ -126,6 +126,13
On 2021/7/20 0:53, Daniel P. Berrangé wrote:
On Mon, Jul 19, 2021 at 11:20:34AM +0800, Yanan Wang wrote:
Currently the only difference between smp_parse and pc_smp_parse
is the support of multi-dies and the related error reporting code.
With an arch compat variable "bool smp_dies_supported", we
data might point into the middle of a larger buffer, there is a separate
free_on_destroy pointer passed into bufp_alloc() to handle that. It is
only used in the normal workflow though, not when dropping packets due
to the queue being full. Fix that.
Resolves: https://gitlab.com/qemu-project/qemu
Hi, Philippe,
On 07/22/2021 01:49 AM, Philippe Mathieu-Daudé wrote:
> On 7/21/21 7:38 PM, Philippe Mathieu-Daudé wrote:
>> On 7/21/21 11:53 AM, Song Gao wrote:
>>> This patch implement fixed point arithemtic instruction translation.
>
> Typo arithmetic.
>
>>>
>>> This includes:
>>> - ADD.{W/D},
Hi
On Thu, Jul 22, 2021 at 11:28 AM Gerd Hoffmann wrote:
> data might point into the middle of a larger buffer, there is a separate
> free_on_destroy pointer passed into bufp_alloc() to handle that. It is
> only used in the normal workflow though, not when dropping packets due
> to the queue be
Hi all,
I see a strange effect on master, with
cd build
../configure --enable-kvm --enable-debug --enable-tcg
--target-list=x86_64-softmmu
make -j
make -j check
...
Not run: 220 287
Passed all 121 iotests
[buzzes here]
ps -a
PID TTY TIME CMD
3609 pts/000:00:00 make
4196 pts
On 7/21/21 2:38 PM, wangyanan (Y) wrote:
On 2021/7/20 0:57, Cornelia Huck wrote:
On Mon, Jul 19 2021, Yanan Wang wrote:
Hi,
This is v2 of the series [1] that I have posted to introduce some smp
parsing
fixes and improvement, much more work has been processed compared to
RFC v1.
[1] ht
> FWIW, this mail seems to have been accepted by the qemu-arm list:
> https://lists.gnu.org/archive/html/qemu-arm/2021-07/msg00341.html
> but it didn't get to the qemu-devel list.
Thank you for letting us know.
As you said, it seems that it is not listed in qemu-devel.
We are checking with the adm
Hi, Philippe.
> Meanwhile you can send your series via sourcehut: https://sr.ht/
Thank you for providing useful information.
We will use the above sites as needed.
Best regards.
> -Original Message-
> From: Philippe Mathieu-Daudé
> Sent: Tuesday, July 20, 2021 8:25 PM
> To: Ishii, Shuu
On Tue, Jul 20, 2021 at 02:54:00PM +0200, Markus Armbruster wrote:
> There is nothing to delete after migrate_add_blocker() failed. Trying
> anyway is safe, but useless. Don't.
>
> Cc: Sunil Muthuswamy
> Cc: Kamil Rytarowski
> Cc: Reinoud Zandijk
> Signed-off-by: Markus Armbruster
> ---
> t
On 2021/7/20 0:53, Andrew Jones wrote:
On Mon, Jul 19, 2021 at 11:20:37AM +0800, Yanan Wang wrote:
We totally have two requirements for a valid SMP configuration:
s/totally//
the sum of "sockets * dies * cores * threads" must represent all
the product
the possible cpus, i.e., max_cpus, and
On 7/22/21 9:49 AM, Claudio Fontana wrote:
> Hi all,
>
> I see a strange effect on master, with
>
> cd build
> ../configure --enable-kvm --enable-debug --enable-tcg
> --target-list=x86_64-softmmu
> make -j
>
> make -j check
> ...
>
> Not run: 220 287
> Passed all 121 iotests
> [buzzes here]
>
** Changed in: qemu
Status: Expired => New
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1860553
Title:
cmake crashes on qemu-alpha-user with Illegal Instruction
Status in QEMU:
New
Bug
Hi, Philippe
On 07/22/2021 01:46 AM, Philippe Mathieu-Daudé wrote:
> On 7/21/21 11:53 AM, Song Gao wrote:
>> This patch implement fixed point bit instruction translation.
>>
>> This includes:
>> - EXT.W.{B/H}
>> - CL{O/Z}.{W/D}, CT{O/Z}.{W/D}
>> - BYTEPICK.{W/D}
>> - REVB.{2H/4H/2W/D}
>> - REVH.{2
CC libvir-list
On Tue, Jul 06, 2021 at 12:01:37 +0200, Paolo Bonzini wrote:
> Switch from QemuOpts to keyval. This enables the introduction
> of non-scalar machine properties, and JSON syntax in the future.
>
> For JSON syntax to be supported right now, we would have to
> consider what would hap
在 2021/7/21 下午12:15, Bin Meng 写道:
From: Christina Wang
The initial value of VLAN Ether Type (VET) register is 0x8100, as per
the manual and real hardware.
While Linux e1000e driver always writes VET register to 0x8100, it is
not always the case for everyone. Drivers relying on the reset valu
On 2021/7/22 15:51, Pierre Morel wrote:
On 7/21/21 2:38 PM, wangyanan (Y) wrote:
On 2021/7/20 0:57, Cornelia Huck wrote:
On Mon, Jul 19 2021, Yanan Wang wrote:
Hi,
This is v2 of the series [1] that I have posted to introduce some
smp parsing
fixes and improvement, much more work has been
在 2021/7/21 下午12:15, Bin Meng 写道:
From: Christina Wang
The initial value of VLAN Ether Type (VET) register is 0x8100, as per
the manual and real hardware.
While Linux e1000 driver always writes VET register to 0x8100, it is
not always the case for everyone. Drivers relying on the reset value
ping
On Tue, Jul 13, 2021 at 6:38 PM Andrew Melnychenko
wrote:
> Libvirt usually launches qemu with strict permissions.
> To enable eBPF RSS steering, qemu-ebpf-rss-helper was added.
>
> Added property "ebpf_rss_fds" for "virtio-net" that allows to
> initialize eBPF RSS context with passed progr
properties set by function x86_cpu_apply_props, including
kvm_default_props, tcg_default_props,
and the "vendor" property for KVM and HVF,
are actually to be set only for cpu models in builtin_x86_defs,
registered with x86_register_cpu_model_type, and not for
cpu models "base", "max", and the subc
On 7/22/21 10:12 AM, Philippe Mathieu-Daudé wrote:
> On 7/22/21 9:49 AM, Claudio Fontana wrote:
>> Hi all,
>>
>> I see a strange effect on master, with
>>
>> cd build
>> ../configure --enable-kvm --enable-debug --enable-tcg
>> --target-list=x86_64-softmmu
>> make -j
>>
>> make -j check
>> ...
>>
On Thu, 22 Jul 2021 at 01:36, Richard Henderson
wrote:
>
> On 7/13/21 3:37 AM, Peter Maydell wrote:
> > +static bool do_ldst_sg(DisasContext *s, arg_vldst_sg *a, MVEGenLdStSGFn fn)
> > +{
> > +TCGv_i32 addr;
> > +TCGv_ptr qd, qm;
> > +
> > +if (!dc_isar_feature(aa32_mve, s) ||
> > +
On 7/10/21 3:30 PM, Peter Maydell wrote:
> I've noticed recently that intermittently 'make check' will hang on
> my aarch32 test system (really an aarch64 box with an aarch32 chroot).
>
> I think from grep that this must be the vhost-user-blk test.
>
> Here's the process tree:
>
> pmaydell 13126
Q35 has now ACPI hotplug enabled by default for PCI(e) devices.
As opposed to native PCIe hotplug, some guests will not assign
IO range to pcie-root-ports not supporting native hotplug,
resulting into a regression.
Fix it by setting the "reserve-io" hint capability of the
pcie-root-ports so the fi
Hi
Signed-off-by: zhuguanghong https://developer.gnome.org/glib/stable/glib-Threads.html#g-once-init-enter
I would have thought QOM usage is so widespread in QEMU that we'll
trigger initialization of this hash tble very early in startup
while we're still single threaded.
That said there's not re
在 2021/7/21 下午4:53, Jonah Palmer 写道:
Hi Jason. My apologies for the delayed response, several work-related
things came up recently, but they're slowing down now so I'm turning
my attention these patches to get taken care of.
A few questions and comments below (and in other following patche
在 2021/7/21 下午5:11, Jonah Palmer 写道:
On 7/13/21 10:40 PM, Jason Wang wrote:
在 2021/7/12 下午6:35, Jonah Palmer 写道:
+void hmp_virtio_queue_status(Monitor *mon, const QDict *qdict)
+{
+ Error *err = NULL;
+ const char *path = qdict_get_try_str(qdict, "path");
+ int queue = qdict_get_i
On Thu, Jul 22, 2021 at 05:15:46PM +0800, 朱光宏 wrote:
> Hi
> Signed-off-by: zhuguanghong ---
> qom/object.c | 4
> 1 file changed, 4 insertions(+)
>
>
> diff --git a/qom/object.c b/qom/object.c
> index 6a01d56546..dc0a363ed0 100644
> --- a/qom/object.c
> +++ b/qom/object.c
> @@ -73,15 +73,
在 2021/7/21 下午4:59, Jonah Palmer 写道:
On 7/13/21 10:37 PM, Jason Wang wrote:
在 2021/7/12 下午6:35, Jonah Palmer 写道:
From: Laurent Vivier
This new command shows internal status of a VirtQueue.
(vrings and indexes).
Signed-off-by: Laurent Vivier
Signed-off-by: Jonah Palmer
---
hw/virtio/
Signed-off-by: zhuguanghong
---
qom/object.c | 4
1 file changed, 4 insertions(+)
diff --git a/qom/object.c b/qom/object.c
index 6a01d56546..dc0a363ed0 100644
--- a/qom/object.c
+++ b/qom/object.c
@@ -73,15 +73,19 @@ struct TypeImpl
};
static Type type_interface;
+G_LOCK_DEFINE (GHashTa
在 2021/7/21 下午10:15, Thomas Huth 写道:
QEMU should never terminate unexpectedly just because the guest is
doing something wrong like specifying wrong queue numbers. Let's
simply refuse to set the device active in this case.
Buglink: https://bugs.launchpad.net/qemu/+bug/1890160
Signed-off-by: Tho
在 2021/7/16 上午3:32, Thomas Huth 写道:
QEMU currently crashes when it's started like this:
cat << EOF | ./qemu-system-i386 -device vmxnet3 -nodefaults -qtest stdio
outl 0xcf8 0x80001014
outl 0xcfc 0xe0001000
outl 0xcf8 0x80001018
outl 0xcf8 0x80001004
outw 0xcfc 0x7
outl 0xcf8 0x80001083
write 0x
When skipping free pages to send, their corresponding dirty bits in the
memory region dirty bitmap need to be cleared. Otherwise the skipped
pages will be sent in the next round after the migration thread syncs
dirty bits from the memory region dirty bitmap.
Cc: David Hildenbrand
Cc: Peter Xu
Cc
This is an automated cleanup. This bug report has been moved to QEMU's
new bug tracker on gitlab.com and thus gets marked as 'expired' now.
Please continue with the discussion here:
https://gitlab.com/qemu-project/qemu/-/issues/494
** Changed in: qemu
Status: New => Expired
** Bug watch
On 22.07.21 10:30, Wei Wang wrote:
When skipping free pages to send, their corresponding dirty bits in the
memory region dirty bitmap need to be cleared. Otherwise the skipped
pages will be sent in the next round after the migration thread syncs
dirty bits from the memory region dirty bitmap.
Cc
On 21/07/2021 18:37, Michael S. Tsirkin wrote:
> On Wed, Jul 21, 2021 at 06:27:33PM +0200, Igor Mammedov wrote:
>> On Wed, 21 Jul 2021 12:09:01 -0400
>> "Michael S. Tsirkin" wrote:
>>
>>> On Wed, Jul 21, 2021 at 05:49:16PM +0200, Laurent Vivier wrote:
On 21/07/2021 16:59, Igor Mammedov wrote:
On Thursday, July 22, 2021 5:48 PM, David Hildenbrand wrote:
> On 22.07.21 10:30, Wei Wang wrote:
> > When skipping free pages to send, their corresponding dirty bits in
> > the memory region dirty bitmap need to be cleared. Otherwise the
> > skipped pages will be sent in the next round after the m
Hi Richard,
I think you were the last person to fiddle with the prefetching
in buffer_zero_avx2 and friends; Joe (cc'd) wondered if explicit
prefetching still made sense on modern CPUs, and that their hardware
generally figures stuff out better on simple increments.
What was your thinking on t
This simple patch suppresses a warning message from qemu-nbd when a
client abruptly disconnects. There is a way to work around this in
the client (by shutting down the connection properly). Nevertheless
the "Broken pipe" error seems unnecessary to me as it does not convey
any useful information a
$ rm -f /tmp/sock /tmp/pid
$ qemu-img create -f qcow2 /tmp/disk.qcow2 1M
$ qemu-nbd -t --format=qcow2 --socket=/tmp/sock --pid-file=/tmp/pid
/tmp/disk.qcow2 &
$ nbdsh -u 'nbd+unix:///?socket=/tmp/sock' -c 'h.get_size()'
qemu-nbd: Disconnect client, due to: Failed to send reply: Unable to write to
On Wed, 21 Jul 2021 12:37:40 -0400
"Michael S. Tsirkin" wrote:
> On Wed, Jul 21, 2021 at 06:27:33PM +0200, Igor Mammedov wrote:
> > On Wed, 21 Jul 2021 12:09:01 -0400
> > "Michael S. Tsirkin" wrote:
> >
> > > On Wed, Jul 21, 2021 at 05:49:16PM +0200, Laurent Vivier wrote:
> > > > On 21/07/2
On 2021/7/22 14:02, Cornelia Huck wrote:
On Thu, Jul 22 2021, Yanan Wang wrote:
In the SMP configuration, we should either specify a topology
parameter with a reasonable value (equal to or greater than 1)
or just leave it omitted and QEMU will calculate its value.
Configurations which explicit
For full description see 2/2.
Tested hotplug on Q35 (see 2/2 for reproducer) and PC (with pci-bridge) machines
Igor Mammedov (2):
acpi: x86: pcihp: cleanup devfn usage in
build_append_pci_bus_devices()
acpi: x86: pcihp: add support hotplug on multifunction bridges
hw/i386/acpi-build.c |
Signed-off-by: Igor Mammedov
---
hw/i386/acpi-build.c | 13 ++---
1 file changed, 6 insertions(+), 7 deletions(-)
diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
index 17836149fe..b40e284b72 100644
--- a/hw/i386/acpi-build.c
+++ b/hw/i386/acpi-build.c
@@ -374,7 +374,7 @@ static
Commit 17858a1695 (hw/acpi/ich9: Set ACPI PCI hot-plug as default on Q35)
switched PCI hotplug from native to ACPI one by default.
That however breaks ihotplug on following CLI that used to work:
-nodefaults -machine q35 \
-device
pcie-root-port,id=pcie-root-port-0,multifunction=on,bus=pcie
Reviewed-by: Michael Rolnik
Tested-by: Michael Rolnik
On Wed, Jul 21, 2021 at 9:00 PM Philippe Mathieu-Daudé
wrote:
> +Michael/Alex/Pavel
>
> On 7/21/21 8:41 AM, Richard Henderson wrote:
> > GDB single-stepping is now handled generically.
> >
> > Signed-off-by: Richard Henderson
> > ---
> >
* David Hildenbrand (da...@redhat.com) wrote:
> virtio-mem exposes a dynamic amount of memory within RAMBlocks by
> coordinating with the VM. Memory within a RAMBlock can either get
> plugged and consequently used by the VM, or unplugged and consequently no
> longer used by the VM. Logical unplug i
On 22.07.21 13:29, Dr. David Alan Gilbert wrote:
* David Hildenbrand (da...@redhat.com) wrote:
virtio-mem exposes a dynamic amount of memory within RAMBlocks by
coordinating with the VM. Memory within a RAMBlock can either get
plugged and consequently used by the VM, or unplugged and consequentl
On Fri, Jul 02, 2021 at 04:32:21PM +0400, marcandre.lur...@redhat.com wrote:
> From: Marc-André Lureau
>
> Before commit 49afbca3b00e8e517d54964229a794b51768deaf ("virtio-gpu: drop
> use_virgl_renderer"), use_virgl_renderer was preventing calling GL
> functions from non-GL context threads. The in
[I know this is probably too late for 6.1 but maybe the first patch of the
series is good as it just adds a missing doc?]
By default, KVM doesn't limit the usage of paravirtualized feature (neither
native KVM nor Hyper-V) to what was exposed to the guest in CPUIDs making
it possible to use all of
By default, KVM allows the guest to use all currently supported PV features
even when they were not announced in guest visible CPUIDs. Introduce a new
"kvm-pv-enforce-cpuid" flag to limit the supported feature set to the
exposed features. The feature is supported by Linux >= 5.10 and is not
enabled
By default, KVM allows the guest to use all currently supported Hyper-V
enlightenments when Hyper-V CPUID interface was exposed, regardless of if
some features were not announced in guest visible CPUIDs. hv-enforce-cpuid
feature alters this behavior and only allows the guest to use exposed
Hyper-V
KVM PV features don't seem to be documented anywhere, in particular, the
fact that some of the features are enabled by default and some are not can
only be figured out from the code.
Signed-off-by: Vitaly Kuznetsov
---
docs/kvm-pv.txt | 92 +
1 fil
21.07.2021 23:46, Eric Blake wrote:
Add a regression test to make sure we detect attempts to use 'qemu-img
bitmap' to modify an in-use local file.
Suggested-by: Nir Soffer
Signed-off-by: Eric Blake
---
Sadly, this missed my bitmaps pull request today. If there's any
reason to respin that pul
Hi,
This is a rather complex series with changes that aren’t exactly local
to the mirror job, so maybe it’s too complex for 6.1.
However, it is a bug fix, and not an insignificant one, though probably
not a regression of any kind.
Bug report:
https://gitlab.com/qemu-project/qemu/-/issues/462
(I
An error does not take us out of the READY phase, which is what
s->synced signifies. It does of course mean that source and target are
no longer in sync, but that is what s->actively_sync is for -- s->synced
never meant that source and target are in sync, only that they were at
some point (and at
Callers should be able to specify whether they want job_cancel_sync() to
force-cancel the job or not.
In fact, almost all invocations do not care about consistency of the
result and just want the job to terminate as soon as possible, so they
should pass force=true. The replication block driver is
We must check whether the job is force-cancelled early in our main loop,
most importantly before any `continue` statement. For example, we used
to have `continue`s before our current checking location that are
triggered by `mirror_flush()` failing. So, if `mirror_flush()` kept
failing, force-canc
On Thu, Jul 22, 2021 at 12:42:55PM +0800, wangyanan (Y) wrote:
> On 2021/7/20 0:42, Andrew Jones wrote:
> > On Mon, Jul 19, 2021 at 11:20:36AM +0800, Yanan Wang wrote:
> > > Currently we directly calculate the omitted cpus based on the already
> > > provided parameters. This makes some cmdlines lik
We largely have two cancel modes for jobs:
First, there is actual cancelling. The job is terminated as soon as
possible, without trying to reach a consistent result.
Second, we have mirror in the READY state. Technically, the job is not
really cancelled, but it just is a different completion mo
Most callers of job_is_cancelled() actually want to know whether the job
is on its way to immediate termination. For example, we refuse to pause
jobs that are cancelled; but this only makes sense for jobs that are
really actually cancelled.
A mirror job that is cancelled during READY with force=f
#1 adds support for MADV_POPULATE_WRITE, #2 cleans up the code to avoid
global variables and prepare for concurrency, #3 and #4 optimize thread
handling, #5 makes os_mem_prealloc() safe to be called from multiple
threads concurrently and #6 makes the SIGBUS handler coexist cleanly with
the MCE SIGB
Let's sense support and use it for preallocation. MADV_POPULATE_WRITE
does not require a SIGBUS handler, doesn't actually touch page content,
and avoids context switches; it is, therefore, faster and easier to handle
than our current approach.
While MADV_POPULATE_WRITE is, in general, faster than
Let's minimize the number of global variables to prepare for
os_mem_prealloc() getting called concurrently and make the code a bit
easier to read.
The only consumer that really needs a global variable is the sigbus
handler, which will require protection via a mutex in the future either way
as we c
Add a mutex to protect the SIGBUS case, as we cannot mess concurrently
with the sigbus handler and we have to manage the global variable
sigbus_memset_context. The MADV_POPULATE_WRITE path can run
concurrently.
Note that page_mutex and page_cond are shared between concurrent
invocations, which sho
Test what happens when there is an I/O error after a mirror job in the
READY phase has been cancelled.
Signed-off-by: Max Reitz
---
.../tests/mirror-ready-cancel-error | 143 ++
.../tests/mirror-ready-cancel-error.out | 5 +
2 files changed, 148 insertions(+)
c
Let's limit the number of threads to something sane, especially that
- We don't have more threads than the number of pages we have
- We don't have threads that initialize small (< 64 MiB) memory
Signed-off-by: David Hildenbrand
---
util/oslib-posix.c | 12 ++--
1 file changed, 10 inserti
Let's simplify the case when we only want a single thread and don't have
to mess with signal handlers.
Signed-off-by: David Hildenbrand
---
util/oslib-posix.c | 8
1 file changed, 8 insertions(+)
diff --git a/util/oslib-posix.c b/util/oslib-posix.c
index a1d309d495..1483e985c6 100644
-
Temporarily modifying the SIGBUS handler is really nasty, as we might be
unlucky and receive an MCE SIGBUS while having our handler registered.
Unfortunately, there is no way around messing with SIGBUS when
MADV_POPULATE_WRITE is not applicable or not around.
Let's forward SIGBUS that don't belong
On Thu, Jul 22, 2021 at 04:10:32PM +0800, wangyanan (Y) wrote:
> On 2021/7/20 0:53, Andrew Jones wrote:
> > On Mon, Jul 19, 2021 at 11:20:37AM +0800, Yanan Wang wrote:
> > > We totally have two requirements for a valid SMP configuration:
> > s/totally//
> >
> > > the sum of "sockets * dies * cores
On 22/07/2021 12:59, Igor Mammedov wrote:
> Commit 17858a1695 (hw/acpi/ich9: Set ACPI PCI hot-plug as default on Q35)
> switched PCI hotplug from native to ACPI one by default.
>
> That however breaks ihotplug on following CLI that used to work:
>-nodefaults -machine q35 \
>-device
> pcie
Hi
On Thu, Jul 22, 2021 at 4:12 PM Gerd Hoffmann wrote:
> On Fri, Jul 02, 2021 at 04:32:21PM +0400, marcandre.lur...@redhat.com
> wrote:
> > From: Marc-André Lureau
> >
> > Before commit 49afbca3b00e8e517d54964229a794b51768deaf ("virtio-gpu: drop
> > use_virgl_renderer"), use_virgl_renderer was
On Wed, 21 Jul 2021 at 20:49, Eric Blake wrote:
>
> The following changes since commit e77c8b8b8e933414ef07dbed04e02973fccffeb0:
>
> Update version for v6.1.0-rc0 release (2021-07-21 17:10:15 +0100)
>
> are available in the Git repository at:
>
> https://repo.or.cz/qemu/ericb.git tags/pull-bit
On Thu, Jul 22, 2021 at 02:24:03PM +0800, wangyanan (Y) wrote:
> On 2021/7/20 1:20, Andrew Jones wrote:
> > On Mon, Jul 19, 2021 at 11:20:42AM +0800, Yanan Wang wrote:
> > > We are going to introduce an unit test for the parser smp_parse()
> > > in hw/core/machine.c, but now machine.c is only built
On Thu, Jul 22, 2021 at 02:15:18PM +0800, wangyanan (Y) wrote:
> On 2021/7/20 2:57, Andrew Jones wrote:
> > On Mon, Jul 19, 2021 at 11:20:43AM +0800, Yanan Wang wrote:
> > > Add a QEMU unit test for the parsing of given SMP configuration.
> > > Since all the parsing logic is in generic function smp
Hi
On Thu, Jul 22, 2021 at 4:51 PM Marc-André Lureau <
marcandre.lur...@gmail.com> wrote:
> Hi
>
> On Thu, Jul 22, 2021 at 4:12 PM Gerd Hoffmann wrote:
>
>> On Fri, Jul 02, 2021 at 04:32:21PM +0400, marcandre.lur...@redhat.com
>> wrote:
>> > From: Marc-André Lureau
>> >
>> > Before commit 49afb
Eric Auger wrote:
> Hi Dave,
>
> On 7/6/21 4:19 PM, Dr. David Alan Gilbert wrote:
>> * Eric Auger (eric.au...@redhat.com) wrote:
...
>> Well, I initially wanted to know more about this scenario to determine
>> whether
>> a normal shutdown would fall into it.😂
> I think it was for
Since patch submitters tend to submit code that works for the success
case, I like to test a few failure cases before anything else. Gotcha:
$ qemu-system-x86_64 -machine pc,pflash0=xxx qemu-system-x86_64:
Property 'cfi.pflash01.drive' can't find value 'xxx'
The error message is misleadi
On Thu, 22 Jul 2021 at 14:19, Juan Quintela wrote:
>
> Eric Auger wrote:
> > Hi Dave,
> >
> > On 7/6/21 4:19 PM, Dr. David Alan Gilbert wrote:
> > That's consistent with what you suggested in the past ans what is logged
> > in the commit message of
> >
> > cddafd8f353d2d251b1a5c6c948a577a85838582
On Thu, Jul 22, 2021 at 02:36:30PM +0200, David Hildenbrand wrote:
> Let's sense support and use it for preallocation. MADV_POPULATE_WRITE
> does not require a SIGBUS handler, doesn't actually touch page content,
> and avoids context switches; it is, therefore, faster and easier to handle
> than ou
On 22/07/21 15:25, Markus Armbruster wrote:
Since patch submitters tend to submit code that works for the success
case, I like to test a few failure cases before anything else. Gotcha:
$ qemu-system-x86_64 -machine pc,pflash0=xxx
qemu-system-x86_64: Property 'cfi.pflash01.drive' can't
On Thu, Jul 22, 2021 at 08:02:16AM +0200, Cornelia Huck wrote:
> On Thu, Jul 22 2021, Yanan Wang wrote:
>
> > In the SMP configuration, we should either specify a topology
> > parameter with a reasonable value (equal to or greater than 1)
> > or just leave it omitted and QEMU will calculate its v
On 22.07.21 15:31, Daniel P. Berrangé wrote:
On Thu, Jul 22, 2021 at 02:36:30PM +0200, David Hildenbrand wrote:
Let's sense support and use it for preallocation. MADV_POPULATE_WRITE
does not require a SIGBUS handler, doesn't actually touch page content,
and avoids context switches; it is, theref
On Thu, Jul 22, 2021 at 03:39:50PM +0200, David Hildenbrand wrote:
> On 22.07.21 15:31, Daniel P. Berrangé wrote:
> > On Thu, Jul 22, 2021 at 02:36:30PM +0200, David Hildenbrand wrote:
> > > Let's sense support and use it for preallocation. MADV_POPULATE_WRITE
> > > does not require a SIGBUS handle
1 - 100 of 267 matches
Mail list logo