When a PCI device lives behind an IOMMU, it should use 'pci_dma_*' family of
functions when any transfer from/to guest memory is required while
'cpu_physical_memory_*' family of functions completely bypass any MMU/IOMMU in
the system.
vmxnet3 in some places was using 'cpu_physical_memory_*' family
On 17/06/16 20:01, Paolo Bonzini wrote:
>
> On 17/06/2016 18:33, Alex Bennée wrote:
>> First we move the break/watchpoints into an array which is more
>> amenable to RCU control that the QLIST. We then control the life time
>> of references to break/watchpoint data by removing long held
>> referenc
On 06/20/2016 11:32 AM, Benjamin Herrenschmidt wrote:
On Mon, 2016-06-20 at 18:02 +1000, Benjamin Herrenschmidt wrote:
On Mon, 2016-06-20 at 17:08 +1000, Benjamin Herrenschmidt wrote:
That fixed, it dies elsewhere in something related to page faults,
still digging.
Next problem: Darwin k
Some architectures (e.g. ARMv8) need the address which is aligned
to a size more than the size of the memory access.
It's enougth the current costless alignment check implementation in QEMU,
but we need to support the alignment size specifying.
Signed-off-by: Sergey Sorokin
---
include/exec/cpu-
On 20 June 2016 at 04:44, Andrew Jeffery wrote:
> On Fri, 2016-06-17 at 15:22 +0100, Peter Maydell wrote:
>> +static Property aspeed_scu_properties[] = {
>> +DEFINE_PROP_ARRAY("reset", AspeedSCUState, num_resets, reset,
>> + qdev_prop_uint32, uint32_t),
>> +DEFINE_PROP
On 18.06.2016 02:50, alar...@ddci.com wrote:
> Note change of subject from "Determining interest in PPC e500spin,
> yield".
>
> Thomas Huth wrote on 06/16/2016 01:47:05 AM:
> Aaron Larson wrote on 15.06.2016 22:12
>
> in ppce500_spin.c
>
> AL> @@ -104,6 +108,16 @@
> AL>
> AL> cpu_synchro
The following changes since commit 482b61844ae7c6df39df0b48ac90ffbc87bed7d2:
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20160617'
into staging (2016-06-17 16:16:37 +0100)
are available in the git repository at:
git://github.com/stefanha/qemu.git tags/block-pull-requ
Avoid redefining device IDs. Use the standard Linux headers that are
already in the source tree.
Signed-off-by: Stefan Hajnoczi
Message-id: 1462798061-30382-2-git-send-email-stefa...@redhat.com
---
tests/libqos/virtio.h| 9 -
tests/virtio-blk-test.c | 7 ---
tests/virtio-net-te
Note that VIRTIO_F_ANY_LAYOUT and VIRTIO_F_NOTIFY_ON_EMPTY are bit
numbers in virtio_config.h but bit masks in qtest virtio.h. Therefore
it's necessary to change users from X to (1u << X).
Signed-off-by: Stefan Hajnoczi
Message-id: 1462798061-30382-4-git-send-email-stefa...@redhat.com
---
tests
qvirtqueue_setup() allocates the vring and virtqueue state. So far
there has been no function to free it. Callers have been using
guest_free() for the vring but forgot to free the QVirtQueue state.
This patch solves the memory leak by introducing qvirtqueue_cleanup().
Signed-off-by: Stefan Hajn
From: "Denis V. Lunev"
Actually we must set dirty bitmap dirty after we have written all our
zeroes for correct processing in drive mirror code. In the other case
we can face not zeroes in this area in mirror_iteration.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
Rev
Signed-off-by: Stefan Hajnoczi
Message-id: 1462798061-30382-3-git-send-email-stefa...@redhat.com
---
tests/libqos/virtio-pci.c | 3 ++-
tests/libqos/virtio.h | 2 --
2 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/tests/libqos/virtio-pci.c b/tests/libqos/virtio-pci.c
index 9d45
From: "Denis V. Lunev"
The request area is specified in bytes, not in sectors.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladimir Sementsov-Ogievskiy
Reviewed-by: Fam Zheng
Reviewed-by: Eric Blake
Message-id: 1466093381-6120-2-git-send-email-...@openvz.org
CC: Stefan Hajnoczi
CC: Kevin Wolf
Public bug reported:
When setreuid() or setegid() are called from x86_64 target code in user
mode, qemu crashes inside the NPTL signal handlers. x86 targets do not
directly use a syscall to handle setreuid() / setegid(); instead the x86
NPTL implementation sets up a temporary data region in memor
The QMP block-job-resume command and cancellation may want to reset the
job's iostatus. The next patches add a user who does not want to reset
iostatus so move it up to block_job_enter() callers.
Signed-off-by: Stefan Hajnoczi
Reviewed-by: Fam Zheng
Reviewed-by: Paolo Bonzini
Message-id: 14660
On 20 June 2016 at 15:11, Dr. David Alan Gilbert wrote:
> * Peter Maydell (peter.mayd...@linaro.org) wrote:
>> I prefer a "start, length" macro to "position, position",
>> because this matches what we already have for the deposit
>> and extract functions in this header.
>
> I think it depends on t
Move s->target to the new AioContext when there is an AioContext change.
The backup_run() coroutine does not use asynchronous I/O so there is no
need to wait for in-flight requests in a BlockJobDriver->pause()
callback.
Guest writes are intercepted by the backup job. Treat them as guest
activity
The descriptor element, used, and avail vring structs are defined in
virtio_ring.h. There is no need to duplicate them in libqos virtio.
Signed-off-by: Stefan Hajnoczi
Message-id: 1462798061-30382-6-git-send-email-stefa...@redhat.com
---
tests/libqos/virtio.c | 10 +-
tests/libqos/virti
Block jobs are coroutines that usually perform I/O but sometimes also
sleep or yield. Currently only sleeping or yielded block jobs can be
paused. This means jobs that do not sleep or yield (using
block_job_yield()) are unaffected by block_job_pause().
Add block_job_pause_point() so that block j
From: "Denis V. Lunev"
This is mandatory for correct backup creation. In the other case the
content under this area would be lost.
Dirty bits are set exactly like in bdrv_aligned_pwritev, i.e. they are set
even if notifier has returned a error.
Signed-off-by: Denis V. Lunev
Reviewed-by: Vladim
On Mon, Jun 20, 2016 at 4:12 PM, Alex Bennée wrote:
>
> alvise rigo writes:
>
> > Hi Alex,
> >
> > I'm looking into the worries that Sergey issued in his review of the
> > last LL/SC series. The target is to reduce the TLB flushes by using an
> > exclusive history of dynamic length. I don't have
Signed-off-by: Stefan Hajnoczi
Message-id: 1462798061-30382-7-git-send-email-stefa...@redhat.com
---
tests/virtio-blk-test.c | 50 -
1 file changed, 16 insertions(+), 34 deletions(-)
diff --git a/tests/virtio-blk-test.c b/tests/virtio-blk-test.c
in
* Peter Maydell (peter.mayd...@linaro.org) wrote:
> On 16 June 2016 at 18:12, Dr. David Alan Gilbert (git)
> wrote:
> > From: "Dr. David Alan Gilbert"
> >
> > e.g. BIT_RANGE(15, 0) gives 0xff00
> >
> > Suggested by: Paolo Bonzini
> > Signed-off-by: Dr. David Alan Gilbert
> > ---
> > include/qe
It's possible that an AioContext notifier user was close to finishing
when .detach_aio_context() or .attached_aio_context() is called. In
that case they may call bdrv_remove_aio_context_notifier() during the
callback.
Use safe iteration to avoid crashing when the notifier list is modified
during
Patch 2 fixes a wrong assertion in qcow2_get_cluster_offet(). Patch 1
fixes wrong range limitations I encountered in qemu-io while trying to
break that wrong assertion.
Not CC-ing qemu-stable because these issues were introduced after 2.6.0.
Max Reitz (2):
qemu-io: Use correct range limitation
alvise rigo writes:
> Hi Alex,
>
> I'm looking into the worries that Sergey issued in his review of the
> last LL/SC series. The target is to reduce the TLB flushes by using an
> exclusive history of dynamic length. I don't have anything ready yet
> though.
Are you also tackling the race condit
Signed-off-by: Stefan Hajnoczi
Message-id: 1462798061-30382-9-git-send-email-stefa...@redhat.com
---
tests/libqos/virtio-pci.c | 42 ++
tests/libqos/virtio-pci.h | 17 -
tests/virtio-blk-test.c | 11 ++-
tests/virtio-scsi-test.c |
Block jobs that use additional BDSes or event loop resources need a
callback to get their affairs in order when the AioContext is switched.
Simple block jobs don't need an attach callback, they automatically work
thanks to the generic attach/detach notifiers that this patch adds.
Signed-off-by: S
Whoops, I meant http://patchwork.ozlabs.org/patch/590640/.
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1594394
Title:
Using setreuid / setegid crashes x86_64 user-mode target
Status in QEMU:
N
On 20/06/2016 15:55, Sergey Fedorov wrote:
>> > I'm not sure why you say that arrays are more amenable than QTAILQ
>> > (though indeed include/qemu/rcu_queue.h only includes QLIST for now),
>> > but I feel bad asking you to redo all the work...
> Is there any realistic way to manage *doubly* link
The block_job_is_paused() function name is not great because callers
only use it to determine whether pausing has been requested. Rename it
to highlight those semantics and remove it from the public header file
as there are no external callers.
Signed-off-by: Stefan Hajnoczi
Reviewed-by: Paolo B
On Tue, 31 May 2016 20:48:38 +0300
Marcel Apfelbaum wrote:
> Re-use the pci acpi hotplug code and enable it only for
> the new machines using the 'acpi-pci-hotplug-with-bridge-support'
> compat property.
>
> Signed-off-by: Marcel Apfelbaum
> ---
> hw/acpi/ich9.c | 31 ++
create_iovec() has a comment lamenting the lack of SIZE_T_MAX. Since
there actually is a SIZE_MAX, use it.
Two places use INT_MAX for checking the upper bound of a sector count
that is used as an argument for a blk_*() function (blk_discard() and
blk_write_compressed(), respectively). BDRV_REQUEST
Signed-off-by: Stefan Hajnoczi
Message-id: 1462798061-30382-8-git-send-email-stefa...@redhat.com
---
tests/virtio-scsi-test.c | 45 +++--
1 file changed, 15 insertions(+), 30 deletions(-)
diff --git a/tests/virtio-scsi-test.c b/tests/virtio-scsi-test.c
ind
Add a helper function to document why block jobs sometimes run in the
QEMU main loop and to avoid code duplication in a following patch.
Signed-off-by: Stefan Hajnoczi
Reviewed-by: Paolo Bonzini
Reviewed-by: Fam Zheng
Message-id: 1466096189-6477-5-git-send-email-stefa...@redhat.com
---
blockjo
An asserted pirq can be disabled and the corresponding GSIs
should then go down to 0. However, because of the conditional in
ich9_lpc_update_by_pirq, the legacy 8259 pin could remain stuck to 1.
Signed-off-by: Paolo Bonzini
---
hw/isa/lpc_ich9.c | 3 ---
1 file changed, 3 deletions(-)
diff --g
Note that virtio_ring.h defines feature bits using their bit number:
#define VIRTIO_RING_F_INDIRECT_DESC 28
On the other hand libqos virtio.h uses the bit mask:
#define QVIRTIO_F_RING_INDIRECT_DESC0x1000
The patch makes the necessary adjustments.
I have used "1u << BITMASK" ins
Add block_job_pause_point() calls to mark quiescent points and make sure
to complete in-flight requests when switching AioContexts.
This patch solves undefined behavior in the mirror block job when the
BDS AioContext is changed by dataplane.
Signed-off-by: Stefan Hajnoczi
Signed-off-by: Fam Zhen
g_source_attach can return any value between 1 and UINT_MAX if you let
QEMU run long enough. However, qemu_chr_fe_add_watch can also return
a negative errno value when the device is disconnected or does not
support chr_add_watch. Change it to return zero to avoid overloading
these values.
Fix th
Recently, qcow2_get_cluster_offset() has been changed to work with bytes
instead of sectors. This invalidated some assertions and introduced a
possible integer multiplication overflow.
This could be reproduced using e.g.
$ qemu-img create -f qcow2 -o cluster_size=1M blub.qcow2 8G
Formatting 'foo.
serial_xmit starts transmission of whatever is in the FIFO or THR;
serial_watch_cb is a wrapper around it.
Signed-off-by: Paolo Bonzini
---
hw/char/serial.c | 21 +
1 file changed, 13 insertions(+), 8 deletions(-)
diff --git a/hw/char/serial.c b/hw/char/serial.c
index 6f0a49
I mostly filed the bug report since I was seeing multiple different
attempts to implement this, and even a proper patch series on the
mailing list, but no movement at all toward integrating this feature
into mainline qemu.
What would be needed to e.g. make the patch series on the mailing list
acce
ich9->pic and ich9->ioapic differ for the first 16 GSIs (because
ich9->pic is wired to 8259+IOAPIC but ich9->ioapic is wired to
IOAPIC only). However, ich9->ioapic is never used for the first
16 GSIs, so the two vectors can be merged.
Signed-off-by: Paolo Bonzini
---
hw/i386/pc_q35.c | 3
On Tue, 31 May 2016 20:48:37 +0300
Marcel Apfelbaum wrote:
subj doesn't match patch, it does opposite i.e. makes sure that
pcie isn't counted in when building acpi hotplug aml
> Emit acpi code for pci hotplug on all PC machines:
> - if legacy pci hotpug is enabled (pcihp_bridge_en)
> - if ther
On 18/06/2016 09:54, Ashijeet Acharya wrote:
> Use socket_*() functions from include/qemu/sockets.h instead of
> listen()/bind()/ connect()/parse_host_port(). socket_*() fucntions are
> QAPI based and this patch performs this api conversion since everything
> will be using QAPI based sockets in t
https://patches.linaro.org/patch/63313/ may be relevant here.
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1594394
Title:
Using setreuid / setegid crashes x86_64 user-mode target
Status in QEMU:
Blue hasn't been active in the QEMU project for a long time. Drop his
last MAINTAINERS entries.
As per Paolo's recommendation, downgrade status of "BSD user" from
Maintained to Orphan since the FreeBSD guys effectively forked it, and
"SPARC target" from Maintained to Odd Fixes, since we still hav
Currently the southbridge in the q35 machine types gets 40 different
qemu_irqs. The 16 lower GSIs (connected to both 8259 and IOAPIC) and
the 24 input pins of the IOAPIC (the higher 8 of which correspond to the
8 higher GSIs). However, there's no reason to use the 16 first input
pins of the IOAPI
This patchset fixes a couple of minor bugs in linux-user which
were manifesting as LTP test failures:
* syscalls which take a sigsetsize argument should check it
matches the target's sigset_t size, or fail EINVAL
* a typo meant we weren't locking enough data for the argument
to rt_sigqueuin
-dfilter overwrites any previous filter. The overwritten filter is
leaked. Leaks since the beginning (commit 3514552, v2.6.0). Free it
properly.
Signed-off-by: Markus Armbruster
Message-Id: <1466011636-6112-2-git-send-email-arm...@redhat.com>
Reviewed-by: Eric Blake
---
util/log.c | 7 ++
Many syscalls which take a sigset_t argument also take an argument
giving the size of the sigset_t. The kernel insists that this
matches its idea of the type size and fails EINVAL if it is not.
Implement this logic in QEMU. (This mostly just means some LTP test
cases which check error cases now p
Support the F_GETPIPE_SZ and F_SETPIPE_SZ fcntl operations.
Signed-off-by: Peter Maydell
---
linux-user/strace.c | 7 +++
linux-user/syscall.c | 6 ++
linux-user/syscall_defs.h | 2 ++
3 files changed, 15 insertions(+)
diff --git a/linux-user/strace.c b/linux-user/strace.c
in
*** This bug is a duplicate of bug 1591724 ***
https://bugs.launchpad.net/bugs/1591724
Thanks - so it's fixed upstream and in ubuntu yakkety. I'll mark it as
a dup of bug 1591724.
** Changed in: qemu (Ubuntu)
Status: Confirmed => Fix Released
** This bug has been marked a duplicate o
On Sat, Jun 18, 2016 at 1:48 AM, Richard Henderson wrote:
> On 06/17/2016 09:03 PM, Pranith Kumar wrote:
>>
>> case 0xe8 ... 0xef: /* lfence */
>> +tcg_gen_mb(TCG_MO_LD_LD | TCG_BAR_SC);
>> +break;
>> case 0xf0 ... 0xf7: /* mfence */
>> if (!(
On 18/06/2016 22:42, Ido Yariv wrote:
> The current code creates a whole page mmio region for the MSI-X table
> size.
>
> However, the page containing the MSI-X table may contain other registers
> not related to MSI-X. Creating an mmio region for the whole page masks
> such registers and may bre
I went through and quickly analysed all our LTP failures
when running under linux-user, which I figured might be of
interest to the list...
This is with current master plus the following patches (all on list):
linux-user: Avoid possible misalignment in host_to_target_siginfo()
linux-user: Use __g
The main fixes here are in patch 2 and patch 6: watches are lost
after migration, and not removed on reset.
The rest are cleanups; patch 5 fixes the qemu_chr_fe_add_watch
API, which botches its return value pretty badly.
Paolo
Paolo Bonzini (6):
serial: make tsr_retry unsigned
serial: reinst
From: Eduardo Habkost
error_propagate() already ignores local_err==NULL, so there's no
need to check it before calling.
Coccinelle patch used to perform the changes added to
scripts/coccinelle/error_propagate_null.cocci.
Reviewed-by: Eric Blake
Acked-by: Cornelia Huck
Reviewed-by: Markus Armb
The following changes since commit 5edbd4e30426d3a0d712510b2509a521e35192b1:
Merge remote-tracking branch 'remotes/otubo/tags/pull-seccomp-20160620' into
staging (2016-06-20 12:53:35 +0100)
are available in the git repository at:
git://repo.or.cz/qemu/armbru.git tags/pull-error-
Move common code outside the if, and reset tsr_retry even in loopback mode.
Right now it cannot become non-zero, but it will be possible as soon as
we start respecting the baud rate.
Signed-off-by: Paolo Bonzini
---
hw/char/serial.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff
http://git.qemu.org/?p=qemu.git;a=commitdiff;h=362f5fb5643a9cfcf4b5127f
** Changed in: qemu
Status: Fix Committed => Fix Released
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/984476
Title:
When qemu_set_log_filename() detects an invalid file name, it reports
an error, closes the log file (if any), and starts logging to stderr
(unless daemonized or nothing is being logged).
This is wrong. Asking for an invalid log file on the command line
should be fatal. Asking for one in the moni
Otherwise, a serial port can get stuck if it is migrated while flow control
is in effect.
Signed-off-by: Paolo Bonzini
---
hw/char/serial.c | 16 ++--
include/hw/char/serial.h | 1 +
2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/hw/char/serial.c b/hw/char/s
On 20 June 2016 at 15:55, Paolo Bonzini wrote:
>
>
> On 18/06/2016 09:54, Ashijeet Acharya wrote:
>> Use socket_*() functions from include/qemu/sockets.h instead of
>> listen()/bind()/ connect()/parse_host_port(). socket_*() fucntions are
>> QAPI based and this patch performs this api conversion s
From: Eduardo Habkost
Use Coccinelle script to replace 'ret = E; return ret' with
'return E'. The script will do the substitution only when the
function return type and variable type are the same.
Manual fixups:
* audio/audio.c: coding style of "read (...)" and "write (...)"
* block/qcow2-clust
Otherwise, this can cause serial_xmit to be entered with LSR.TEMT=0,
which is invalid and causes an assertion failure.
Reported-by: Bret Ketchum
Signed-off-by: Paolo Bonzini
---
hw/char/serial.c | 16
1 file changed, 12 insertions(+), 4 deletions(-)
diff --git a/hw/char/serial
http://git.qemu.org/?p=qemu.git;a=commitdiff;h=a5b71725c7067f6805eb30
** Changed in: qemu
Status: Fix Committed => Fix Released
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1036987
Title:
It can never become negative; reflect this in the type of the field
and simplify the conditions.
Signed-off-by: Paolo Bonzini
---
hw/char/serial.c | 12
include/hw/char/serial.h | 2 +-
2 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/hw/char/serial.c b/hw/cha
On 20 June 2016 at 15:05, Timothy Pearson wrote:
> I mostly filed the bug report since I was seeing multiple different
> attempts to implement this, and even a proper patch series on the
> mailing list, but no movement at all toward integrating this feature
> into mainline qemu.
>
> What would be
Commit 926cde5 ("scsi: esp: make cmdbuf big enough for maximum CDB size",
2016-06-16) changed the size of a migrated field. Split it in two
parts, and only migrate the second part in a new vmstate version.
Signed-off-by: Paolo Bonzini
---
hw/scsi/esp.c | 5 +++--
include/migration
From: Evgeny Yakovlev
Some guests (win2008 server for example) do a lot of unnecessary
flushing when underlying media has not changed. This adds additional
overhead on host when calling fsync/fdatasync.
This change introduces a dirty flag in BlockDriverState which is set
in bdrv_set_dirty and is
Make ich9_lpc_update_pic take care only of GSIs 0-15, and
ich9_lpc_update_apic take care only of GSIs 16-23. Assert
that they are called with the correct GSI indices.
Signed-off-by: Paolo Bonzini
---
hw/isa/lpc_ich9.c | 39 ---
1 file changed, 16 insertions(+
On 06/20/2016 08:26 AM, Max Reitz wrote:
> create_iovec() has a comment lamenting the lack of SIZE_T_MAX. Since
> there actually is a SIZE_MAX, use it.
>
> Two places use INT_MAX for checking the upper bound of a sector count
> that is used as an argument for a blk_*() function (blk_discard() and
On 06/20/2016 08:26 AM, Max Reitz wrote:
> Recently, qcow2_get_cluster_offset() has been changed to work with bytes
> instead of sectors. This invalidated some assertions and introduced a
> possible integer multiplication overflow.
>
> This could be reproduced using e.g.
>
> $ qemu-img create -f
From: Evgeny Yakovlev
This change adds hyperv feature words report through qom rpc.
When VM is configured with hyperv features enabled
libvirt will check that required feature words are set
in cpuid leaf 4003 through qom request.
Currently qemu does not report hyperv feature words
which pre
On 17/06/2016 16:03, Paolo Bonzini wrote:
> The modeling here was weird. ICH9 does not need both ->pic and
> ->ioapic, it can make do with just a 24-entry GSI array.
>
> If you change that in the previous patch, this one makes much more
> sense. As it is now, it seems like the ISA bus will not
On Mon, 20 Jun 2016, Aleksandar Markovic wrote:
> This patch is about NaN-2008 flavor of Mips instructions CEIL, CVT,
> FLOOR, ROUND, TRUNC only (its title is "Add nan2008 flavor...").
>
> Legacy-NaN flavors of the same Mips instructions already operate
> correctly, and there is nothing to be f
On 17 June 2016 at 09:35, Cédric Le Goater wrote:
> This test uses the palmetto platform and the AST2400 SPI controller to
> test the m25p80 flash module device model. The flash model is defined
> by the platform (n25q256a) and it would be nice to find way to control
> it, using a property probabl
The third argument to the rt_sigqueueinfo syscall is a pointer to
a siginfo_t, not a pointer to a sigset_t. Fix the error in the
arguments to lock_user(), which meant that we would not have
detected some faults that we should.
Signed-off-by: Peter Maydell
---
linux-user/syscall.c | 5 -
1 fi
On 06/20/2016 06:56 AM, Sergey Sorokin wrote:
/* Flags stored in the low bits of the TLB virtual address. These are
- defined so that fast path ram access is all zeros. */
+ * defined so that fast path ram access is all zeros.
+ * They start after address alignment bits.
+ */
+#define TLB_FL
On 06/20/2016 06:43 AM, Peter Maydell wrote:
On 18 June 2016 at 07:47, Richard Henderson wrote:
We fail to pass to AS all of the different flags that
may be required for a given set of CFLAGS. Rather than
figuring out the host-specific mapping, it's better to
allow the compiler driver to do th
Sounds very relevant, yes. Thanks for the link!
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1594394
Title:
Using setreuid / setegid crashes x86_64 user-mode target
Status in QEMU:
New
Bug de
This series includes multiple changes to the way errors are
handled by the global property system.
Git tree:
https://github.com/ehabkost/qemu-hacks.git work/global-error-handling
Changes v1 -> v2:
* Fix test-qdev-global-props unit test
* Remove subprocess code and "nouser" test case
* Repor
Move the compat_props handling to core machine code.
Reviewed-by: Marcel Apfelbaum
Signed-off-by: Eduardo Habkost
---
hw/core/machine.c | 16
include/hw/boards.h | 1 +
vl.c| 9 ++---
3 files changed, 19 insertions(+), 7 deletions(-)
diff --git a/hw/cor
We print a few fatal error messages to stdout instead of stderr.
Reproducer:
$ qemu-system-x86_64 -g 1024x768
Option g not supported for this target
$ qemu-system-x86_64 -g 1024x768 >/dev/null
Fix by printing them with error_report(). This also improves the messages.
The above one be
MachineClass::compat_props may point to class names that are not
compiled into the QEMU binary. Skip registering those as global
properties. This will allow the qdev global property code to
implement stricter checks on the global property values in the
future.
Signed-off-by: Eduardo Habkost
---
C
The function is not used by any QEMU code, except for the
test-qdev-global-props unit test.
Signed-off-by: Eduardo Habkost
---
Changes v1 -> v2:
* (New patch in series)
---
hw/core/qdev-properties.c | 9 -
include/hw/qdev-properties.h | 1 -
tests/test-qdev-global-props.c | 12 +
Since commit 25f8dd9, qdev_prop_set_globals_for_type() stops
applying global properties on the first error. It is a leftover
from when QEMU exited on any error when applying global property.
Now we print a warning about the first error, bug ignore all
other global properties after it.
For example,
Instead of just printing a warning very late, reject obviously
invalid -global arguments by validating the class name.
Update test-qdev-global-props to not expect class name validation
to be performed in qdev_prop_check_globals().
Reviewed-by: Markus Armbruster
Signed-off-by: Eduardo Habkost
--
From: Eduardo Habkost
This patch simplifies code that uses a local_err variable just to
immediately use it for an error_propagate() call.
Coccinelle patch used to perform the changes added to
scripts/coccinelle/remove_local_err.cocci.
Reviewed-by: Eric Blake
Acked-by: Cornelia Huck
Signed-off
qdev_prop_check_globals() tries to warn the user if a given
-global option was not used. But it does that only if the device
is not hotpluggable.
The warning also makes it harder for management code or people
that write their own scripts or config files: there's no way to
know if a given -global o
This was fixed in QEMU 2.5.
** Changed in: qemu
Status: New => Fix Released
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1356916
Title:
Too small argv limit
Status in QEMU:
Fix Release
g_error() is not an acceptable way to report errors to the user:
$ qemu-system-x86_64 -dfilter 1000+0
** (process:17187): ERROR **: Failed to parse range in: 1000+0
Trace/breakpoint trap (core dumped)
g_assert() isn't, either:
$ qemu-system-x86_64 -dfilter 1000x+64
**
ER
Use the new GlobalProperty.errp field to handle compat_props
errors.
Example output before this change:
(with an intentionally broken entry added to PC_COMPAT_1_3 just
for testing)
$ qemu-system-x86_64 -machine pc-1.3
qemu-system-x86_64: hw/core/qdev-properties.c:1091:
qdev_prop_set_globals_
The function is just a helper to handle the -global options, it
can stay in vl.c like most qemu_opts_foreach() calls.
Reviewed-by: Igor Mammedov
Signed-off-by: Eduardo Habkost
---
hw/core/qdev-properties-system.c | 21 +
include/qemu/config-file.h | 1 -
vl.c
Paolo Bonzini writes:
> On 13/06/2016 11:03, Markus Armbruster wrote:
>> Paolo, will you pick this up, or do you want me to take it through my
>> (currently empty) error-next branch?
>
> I can pick it up.
Since other error-related patches have since accumulated, I posted a
pull request for all o
Fix committed: b4850e5ae9607f9f31932
** Changed in: qemu
Status: New => Fix Committed
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1572329
Title:
ARM bootloader does not set r0 to 0
Stat
On 06/20/2016 05:38 PM, Peter Maydell wrote:
> On 17 June 2016 at 13:15, Cédric Le Goater wrote:
>> A set of SPI flash slaves is attached under the flash controllers of
>> the palmetto platform. "n25q256a" flash modules are used for the BMC
>> and "mx25l25635e" for the host. These types are common
** Changed in: qemu
Status: New => Fix Committed
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1590336
Title:
qemu-arm does not reject vrintz on non-v8 cpu
Status in QEMU:
Fix Committed
On 16.06.2016 18:06, Peter Maydell wrote:
> Don't use the cpu_to_*w() functions, which we are trying to deprecate.
> Instead either just use cpu_to_*() to do the byteswap, or use
> st*_be_p() if we need to do the store somewhere other than to a
> variable that's already the correct type.
>
> Signe
101 - 200 of 389 matches
Mail list logo