On 2013-05-08 00:42, Jan Kiszka wrote:
> On 2013-05-07 23:03, Jordan Justen wrote:
>> On Sun, Mar 24, 2013 at 11:06 AM, Jan Kiszka wrote:
>>> On 2013-02-25 16:44, Jan Kiszka wrote:
On 2013-02-25 16:39, Anthony Liguori wrote:
> Jan Kiszka writes:
>
>> This is in fact very simply:
Currently RTAS and hypercalls are registered in the XICS class init
function. The upcoming XICS-KVM will inherit from XICS but will use
another API to register RTAS tokens with KVM so registration has
to move from the class init function (common for both XICS and
XICS-KVM) to the _realize function
From: David Gibson
Recent (host) kernels support emulating the PAPR defined "XICS" interrupt
controller system within KVM. This patch allows qemu to initialize and
configure the in-kernel XICS, and keep its state in sync with qemu's XICS
state as necessary.
This should give considerable perform
The upcoming support of in-kernel XICS will redefine migration callbacks
for both ICS and ICP so classes and callback pointers are added.
This adds a cpu_setup callback to the XICS device class (as XICS-KVM
will do it different) and xics_dispatch_cpu_setup(). This also moves
the place where xics_d
From: David Gibson
Recent PowerKVM allows the kernel to intercept some RTAS calls from the
guest directly. This is used to implement the more efficient in-kernel
XICS for example. qemu is still responsible for assigning the RTAS token
numbers however, and needs to tell the kernel which RTAS fun
This is rework of in-kernel XICS on top of "[PATCH 00/11] pseries: migration
and QOM support" + compile fix patch + XICS migration fix patch.
Migration from XICS to XICS-KVM and vice versa works.
In this series, XICS-KVM inherits from XICS. I do not really see the point of
adding one more XICS-c
On Tue, Jul 16, 2013 at 10:24:38PM +0100, Alex Bligh wrote:
> --On 16 July 2013 18:55:15 +0200 Paolo Bonzini wrote:
>
> >>What do you think? In the end I thought the schedule_bh_at stuff
> >>was simpler.
> >
> >It is simpler, but I'm not sure it is the right API. Of course, if
> >Kevin and Stefa
On Tue, Jul 16, 2013 at 05:31:01PM +0200, Paolo Bonzini wrote:
> Brown paper bag for me. Commit 2c9b15c added a line setting mr->owner,
> but there was already one that initialized it to NULL (added in the
> earlier commit 803c0816).
What I read from log is the reverse, Commit 2c9b15c added a lin
On Tue, Jul 16, 2013 at 02:17:28PM +0200, Michael Mueller wrote:
> On Tue, 16 Jul 2013 11:05:11 +0800
> Stefan Hajnoczi wrote:
>
> > On Mon, Jul 15, 2013 at 09:41:19PM +0200, Christian Borntraeger wrote:
> > > When running with trace backend e.g. "simple" the writer thread
> > > needs to be imple
On Mon, Jul 01, 2013 at 01:29:17PM +0200, Stefan Hajnoczi wrote:
> Load the virtio.c state into vring.c when we start dataplane mode and
> vice versa when stopping dataplane mode. This patch makes it possible
> to start and stop dataplane any time while the guest is running.
>
> This will eventua
On Tue, Jul 16, 2013 at 09:47:40PM +0530, Bharata B Rao wrote:
> Hi,
>
> This is the v2 of the patchset that changes gluster driver in QEMU
> to use pkg-config and adds discard support.
>
> Changes in v2
> -
> - Ensure that gluster libraries don't get linked to linux-user targets.
>
On Tue, Jul 16, 2013 at 02:45:16PM +0200, Jan Kiszka wrote:
> This restore the behavior prior to b018ddf633 which accidentally changed
> the return code to 0. Specifically guests probing for register existence
> were affected by this.
>
> Signed-off-by: Jan Kiszka
> ---
> memory.c |2 +-
> 1
Hi Andreas,
On Tue, Jul 16, 2013 at 10:19 PM, Andreas Färber wrote:
> Am 16.07.2013 04:00, schrieb Jia Liu:
>> We should free typename here.
>>
>> Signed-off-by: Jia Liu
>> ---
>> target-openrisc/cpu.c | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/target-openrisc/cpu.c b/target-op
Hi Peter,
On Wed, Jul 17, 2013 at 12:43 AM, Peter Maydell
wrote:
> On 16 July 2013 15:31, Peter Crosthwaite wrote:
>> Hi Peter,
>>
>> On Tue, Jul 16, 2013 at 10:25 PM, Peter Maydell
>> wrote:
>>> Replace the opencoded assembly of the reg property array for the
>>> /memory node with a call to qe
Sorry about the delay,
Ack.
We arent getting the long term fix to this soon, so lets apply the patch.
Regards,
Peter
On Wed, Jul 17, 2013 at 3:56 AM, Michael Tokarev wrote:
> Ping^3 ?
>
> 28.06.2013 14:37, Michael Tokarev wrote:
>> 10.06.2013 19:40, Michael Tokarev wrote:
>>> 07.06.2013 07:51,
On 07/17/2013 09:13 AM, Alexander Graf wrote:
>
> On 17.07.2013, at 01:12, Alexander Graf wrote:
>
>>
>> On 16.07.2013, at 12:00, Alexey Kardashevskiy wrote:
>>
>>> Signed-off-by: Alexey Kardashevskiy
>>
>> Why? sPAPRTCE is a typedef struct that's defined in include/hw/ppc/spapr.h
>> which is i
On 17.07.2013, at 01:12, Alexander Graf wrote:
>
> On 16.07.2013, at 12:00, Alexey Kardashevskiy wrote:
>
>> Signed-off-by: Alexey Kardashevskiy
>
> Why? sPAPRTCE is a typedef struct that's defined in include/hw/ppc/spapr.h
> which is included from target-ppc/kvm.c. So it should work just fi
On 16.07.2013, at 12:00, Alexey Kardashevskiy wrote:
> Signed-off-by: Alexey Kardashevskiy
Why? sPAPRTCE is a typedef struct that's defined in include/hw/ppc/spapr.h
which is included from target-ppc/kvm.c. So it should work just fine.
Alex
> ---
> target-ppc/kvm.c | 4 ++--
> 1 file changed
> > Simple is good. Even for deduplication alone, I think data integrity is
> > critical - otherwise we risk stale dedup metadata pointing to clusters
> > that are unallocated or do not contain the right data. So the journal
> > will probably need to follow techniques for commits/checksums.
>
I'
Paolo,
--On 16 July 2013 18:55:15 +0200 Paolo Bonzini wrote:
What do you think? In the end I thought the schedule_bh_at stuff
was simpler.
It is simpler, but I'm not sure it is the right API. Of course, if
Kevin and Stefan says it is, I have no problem with that.
For the sake of having so
Add timed bottom halves. A timed bottom half is a bottom half that
will not execute until a given time has passed (qemu_bh_schedule_at)
or a given interval has passed (qemu_bh_schedule_in). Any qemu
clock can be used, and times are specified in nanoseconds.
Timed bottom halves can be used where ti
Laszlo Ersek writes:
> On 07/16/13 20:57, Anthony Liguori wrote:
>> Laszlo Ersek writes:
>>
>>> The g_io_channel_write_chars() documentation states,
>>>
>>> bytes_written: The number of bytes written. This can be nonzero even if
>>> the return value is not G_IO_STATUS_NORMAL.
On Tue, Jul 16, 2013 at 09:24:30PM +0200, Paolo Bonzini wrote:
> Il 16/07/2013 20:11, Eduardo Habkost ha scritto:
> > For physical bit size, what about extending it in a backwards-compatible
> > way? Something like this:
> >
> > *eax = 0x0003000; /* 48 bits virtual */
> > if (ram_size < 1T
On 07/16/13 20:57, Anthony Liguori wrote:
> Laszlo Ersek writes:
>
>> The g_io_channel_write_chars() documentation states,
>>
>> bytes_written: The number of bytes written. This can be nonzero even if
>> the return value is not G_IO_STATUS_NORMAL. [...]
>>
>> io_channel_send()
Il 16/07/2013 20:11, Eduardo Habkost ha scritto:
> For physical bit size, what about extending it in a backwards-compatible
> way? Something like this:
>
> *eax = 0x0003000; /* 48 bits virtual */
> if (ram_size < 1TB) {
> physical_size = 40; /* Keeping backwards compatibility */
>
Il 16/07/2013 20:55, Eduardo Otubo ha scritto:
Xen uses getrlimit, not sure what this one is. Perhaps glibc's wrapper
calls it?
>>>
>>> It seems to be a glibc's wrapper, yes. Removing it anyway.
>>
>> Why if Xen uses it?
>
> I'm doing virt-test runs since yesterday, it seems it doe
Laszlo Ersek writes:
> The g_io_channel_write_chars() documentation states,
>
> bytes_written: The number of bytes written. This can be nonzero even if
> the return value is not G_IO_STATUS_NORMAL. [...]
>
> io_channel_send() could lose such bytes before.
>
> Furthermore, the (
Le 16/07/2013 19:44, Peter Maydell a écrit :
For m68k, per-thread data is a purely kernel construct with no
CPU level support. Implement it via a field in the TaskState structure,
used by cpu_set_tls() and the set_thread_area/get_thread_area
syscalls. This allows us to enable compilation with NPT
Laszlo Ersek writes:
> When monitor_flush() is invoked repeatedly outside the monitor_unblocked()
> callback, for example from tlb_info() -> ... -> print_pte(), several
> watches may be added for the same event.
>
> This is no problem per se because the extra monitor_unblocked() callbacks
> are h
On 07/16/2013 02:49 PM, Paolo Bonzini wrote:
Il 16/07/2013 19:45, Eduardo Otubo ha scritto:
-{ SCMP_SYS(ugetrlimit), 241 },
Xen uses getrlimit, not sure what this one is. Perhaps glibc's wrapper
calls it?
It seems to be a glibc's wrapper, yes. Removing it anyway.
Why if Xen uses it
unicore32-linux-user provides cpu_set_tls(), so the failure
to enable target_nptl was presumably an oversight. Enable it.
Signed-off-by: Peter Maydell
Reviewed-by: Richard Henderson
---
configure |1 -
1 file changed, 1 deletion(-)
diff --git a/configure b/configure
index 84bcbdc..4a241e0
On 07/16/2013 10:45 AM, Peter Maydell wrote:
> Now all linux-user targets support building with NPTL, we can make it
> mandatory. This is a good idea because:
> * NPTL is no longer new and experimental; it is completely standard
> * in practice, linux-user without NPTL is nearly useless for
>
On 07/16/2013 10:44 AM, Peter Maydell wrote:
> From: Alexander Graf
>
> We can easily set the TLS on i386. Add code to do so.
>
> Signed-off-by: Alexander Graf
> [PMM: also remove "target_nptl=no" line from configure, for
> consistency with other patches in this series]
> Signed-off-by: Peter
On 07/16/2013 10:44 AM, Peter Maydell wrote:
> The i386 code for the get_thread_area syscall was missing a
> 'break' which meant it would have fallen through into the
> implementation of the following syscall; add it.
>
> Signed-off-by: Peter Maydell
> ---
> linux-user/syscall.c |1 +
> 1 fi
On Mon, Jul 15, 2013 at 10:47:25PM +0200, Paolo Bonzini wrote:
> Il 15/07/2013 19:49, Ian Main ha scritto:
> > OK well, I'll explain here my understanding. I apologize if I explain
> > more than needed but it might be good to get this out there anyway.
>
> No problem, it's better to be verbose th
The g_io_channel_write_chars() documentation states,
bytes_written: The number of bytes written. This can be nonzero even if
the return value is not G_IO_STATUS_NORMAL. [...]
io_channel_send() could lose such bytes before.
Furthermore, the (status == G_IO_STATUS_EOF) condition
When monitor_flush() is invoked repeatedly outside the monitor_unblocked()
callback, for example from tlb_info() -> ... -> print_pte(), several
watches may be added for the same event.
This is no problem per se because the extra monitor_unblocked() callbacks
are harmless if mon->outbuf is empty, t
When the IO thread calls monitor_flush() repeatedly & quickly in
succession, outside of callback context, many redundant G_IO_OUT watches
are installed. (One such caller is the "info tlb" / tlb_info() HMP
command which produces a lot of monitor output.)
While this redundancy is no problem in itsel
Now all linux-user targets support building with NPTL, we can make it
mandatory. This is a good idea because:
* NPTL is no longer new and experimental; it is completely standard
* in practice, linux-user without NPTL is nearly useless for
binaries built against non-ancient glibc
* it allows u
On Tue, Jul 16, 2013 at 07:46:14PM +0200, Paolo Bonzini wrote:
> Il 16/07/2013 19:38, Eduardo Habkost ha scritto:
> > On Tue, Jul 16, 2013 at 07:22:01PM +0200, Andrea Arcangeli wrote:
> >> Without this patch the guest physical bits are advertised as 40, not
> >> 44 or more depending on the hardware
Signed-off-by: Eduardo Otubo
---
qemu-seccomp.c |6 --
1 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/qemu-seccomp.c b/qemu-seccomp.c
index 1d5fd71..c44f0d8 100644
--- a/qemu-seccomp.c
+++ b/qemu-seccomp.c
@@ -108,7 +108,6 @@ static const struct QemuSeccompSyscall seccomp_
v2 update:
- set libseccomp 2.1.0 as requirement on configure script.
Since libseccomp 2.0 there's no need to check the architecture type
anymore.
Signed-off-by: Eduardo Otubo
---
configure |2 +-
qemu-seccomp.c | 13 -
2 files changed, 1 insertions(+), 14 deletions(-)
d
Il 16/07/2013 19:45, Eduardo Otubo ha scritto:
>>>
>>> -{ SCMP_SYS(ugetrlimit), 241 },
>>
>> Xen uses getrlimit, not sure what this one is. Perhaps glibc's wrapper
>> calls it?
>
> It seems to be a glibc's wrapper, yes. Removing it anyway.
Why if Xen uses it?
Paolo
Hello all,
In this small patch series I basically:
v2 update:
- set libseccomp 2.1.0 as requirement on configure script.
- removed setrlimit and added sendfile64 to the whitelist.
1) Remove the ifdef's for the (not so) new libseccomp version that does
a
best effort and translate
-On Tue, Jul 16, 2013 at 07:48:27PM +0200, Paolo Bonzini wrote:
> Il 16/07/2013 19:46, Paolo Bonzini ha scritto:
> >>> >> (see PUD with bit >=40 set)
> >> >
> >> > I am not sure I understand what caused this: if we are advertising 40
> >> > physical bits to the guest, why are we ending up with a P
The i386 code for the get_thread_area syscall was missing a
'break' which meant it would have fallen through into the
implementation of the following syscall; add it.
Signed-off-by: Peter Maydell
---
linux-user/syscall.c |1 +
1 file changed, 1 insertion(+)
diff --git a/linux-user/syscall.c
Provide the missing cpu_set_tls(), and resolve the FIXME in
cpu_clone_regs() by clearing the carry flag for the child.
This allows us to turn on building with NPTL for SPARC.
Signed-off-by: Peter Maydell
Reviewed-by: Richard Henderson
---
configure |3 ---
linux-user/spa
For architectures with no linux-user target, don't claim no NPTL
support. This has no behavioural change, but it means that we
won't accidentally add a new linux-user target without threading
support in future (because attempting to do so would be a compile
failure rather than a silent lack of supp
15.07.2013 17:17, Petar Jovanovic wrote:
> From: Petar Jovanovic
>
> sys_futex has 6 arguments, and all of these need to be copied. Fix incorrect
> declaration in the mips_syscall_args array.
Thanks, applied to the trivial patches queue.
/mjt
On 07/15/2013 05:57 PM, Paolo Bonzini wrote:
Il 15/07/2013 19:29, Eduardo Otubo ha scritto:
Signed-off-by: Eduardo Otubo
---
qemu-seccomp.c | 6 --
1 file changed, 6 deletions(-)
diff --git a/qemu-seccomp.c b/qemu-seccomp.c
index 1d5fd71..bfd372a 100644
--- a/qemu-seccomp.c
+++ b/qemu
Ping^3 ?
28.06.2013 14:37, Michael Tokarev wrote:
> 10.06.2013 19:40, Michael Tokarev wrote:
>> 07.06.2013 07:51, Peter Crosthwaite wrote:
>> []
>>> Id like to give it a few days to see what they say before we merge
>>> this. Ideally we have agreement between the two source trees, so we
>>> can co
Linux manages to have three separate orderings of the arguments to
the clone() syscall on different architectures. In the kernel these
are selected via CONFIG_CLONE_BACKWARDS and CONFIG_CLONE_BACKWARDS2.
Clean up our implementation of this to use similar #define names
rather than a TARGET_* ifdef l
For m68k, per-thread data is a purely kernel construct with no
CPU level support. Implement it via a field in the TaskState structure,
used by cpu_set_tls() and the set_thread_area/get_thread_area
syscalls. This allows us to enable compilation with NPTL.
Signed-off-by: Peter Maydell
---
configur
09.07.2013 18:07, Cole Robinson wrote:
> sasldblistusers2 doesn't have a '-a' option
Thanks, applied to the trivial pathes queue.
/mjt
From: Alexander Graf
We can easily set the TLS on i386. Add code to do so.
Signed-off-by: Alexander Graf
[PMM: also remove "target_nptl=no" line from configure, for
consistency with other patches in this series]
Signed-off-by: Peter Maydell
---
configure|1 -
linux-us
On 07/16/2013 10:44 AM, Peter Maydell wrote:
> Add x86-64 implementation of cpu_set_tls() (like the kernel, we
> just have to call do_arch_prctl() to set FS); this allows us to
> enable NPTL.
>
> Signed-off-by: Peter Maydell
> ---
> configure|1 -
> linux-user/i386/target
This patchset enables NPTL on all linux-user targets; this is a
change for i386, m68k, mipsn32, mipsn32el, mips64, mips64el,
or32, ppc64, ppc46abi32, sparc, sparc32plus, sparc64, unicore32
and x86_64 (but not plain ppc or plain mips).
Testing: I've tested with a chroot that or32 and sparc32plus wo
The target-specific headers (target_cpu.h and target_signal.h)
might need to use the target-independent structure and function
definitions of qemu.h; so include them only at the bottom of
qemu.h, not the top.
Signed-off-by: Peter Maydell
Reviewed-by: Richard Henderson
---
linux-user/qemu.h |
Most targets either (a) support threading or (b) don't have a
linux-user/bsd-user target, so make it default to 'yes', with those
targets that don't support threading having to specifically say so.
This is a mechanical no-behaviour-change commit; some of the
"no" entries it adds will be taken out
On 07/16/2013 10:28:28 AM, Fabien Chouteau wrote:
On 07/16/2013 04:06 AM, Scott Wood wrote:
> On 07/10/2013 12:10:02 PM, Fabien Chouteau wrote:
>> This implementation doesn't include ring priority, TCP/IP
Off-Load, QoS.
>>
>> Signed-off-by: Fabien Chouteau
>
> From the code comments I gather
The OpenRISC kernel ignores CLONE_SETTLS in its copy_thread()
implementation, so a cpu_set_tls() implementation is a no-op.
cpu_clone_regs() was setting the syscall return value in the
wrong register -- it is gpr[11], not gpr[2]. With these two
things fixed, we can compile with NPTL enabled.
Signe
Though threading (target_nptl) was enabled on the base PPC and MIPS
targets, it wasn't enabled for the variants mipsn32, mipsn32el,
mips64, mips64el, ppc64 or ppc46abi32 (probably an oversight).
Enable threading consistently on all these targets.
Signed-off-by: Peter Maydell
Reviewed-by: Richard
Il 16/07/2013 19:46, Paolo Bonzini ha scritto:
>>> >> (see PUD with bit >=40 set)
>> >
>> > I am not sure I understand what caused this: if we are advertising 40
>> > physical bits to the guest, why are we ending up with a PUD with
>> > bit >= 40 set?
> Because we create a guest that has bigger me
On 07/15/2013 03:55 PM, Paul Moore wrote:
On Monday, July 15, 2013 02:29:37 PM Eduardo Otubo wrote:
Since libseccomp 2.0 there's no need to check the architecture type
anymore.
Signed-off-by: Eduardo Otubo
---
qemu-seccomp.c | 13 -
1 file changed, 13 deletions(-)
Good, this
Il 16/07/2013 19:38, Eduardo Habkost ha scritto:
> On Tue, Jul 16, 2013 at 07:22:01PM +0200, Andrea Arcangeli wrote:
>> Without this patch the guest physical bits are advertised as 40, not
>> 44 or more depending on the hardware capability of the host.
>>
>> That leads to guest kernel crashes with
Add x86-64 implementation of cpu_set_tls() (like the kernel, we
just have to call do_arch_prctl() to set FS); this allows us to
enable NPTL.
Signed-off-by: Peter Maydell
---
configure|1 -
linux-user/i386/target_cpu.h |7 +++
linux-user/syscall.c |2 +-
On Tue, Jul 16, 2013 at 07:22:01PM +0200, Andrea Arcangeli wrote:
> Without this patch the guest physical bits are advertised as 40, not
> 44 or more depending on the hardware capability of the host.
>
> That leads to guest kernel crashes with injection of page faults 9
> (see oops: 0009) as bits
As we change bdrv_is_allocated to gather more information from bs and
bs->file, it will become a bit slower. It is still appropriate for online
jobs, but not for reads/writes. Call the internal function instead.
Signed-off-by: Paolo Bonzini
---
block/cow.c | 5 -
1 file changed, 4 insertio
This helps implementing is_allocated on top of get_block_status.
Signed-off-by: Paolo Bonzini
---
block.c| 5 +
qemu-img.c | 9 +
2 files changed, 6 insertions(+), 8 deletions(-)
diff --git a/block.c b/block.c
index ae87cba..0fb409b 100644
--- a/block.c
+++ b/block.c
@@ -2934,6
Only sync once per write, rather than once per sector.
Signed-off-by: Paolo Bonzini
---
block/cow.c | 19 ---
1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/block/cow.c b/block/cow.c
index fa3d41f..304e4c7 100644
--- a/block/cow.c
+++ b/block/cow.c
@@ -106,7 +106,
Il 16/07/2013 19:22, Andrea Arcangeli ha scritto:
> Without this patch the guest physical bits are advertised as 40, not
> 44 or more depending on the hardware capability of the host.
>
> That leads to guest kernel crashes with injection of page faults 9
> (see oops: 0009) as bits above 40 in the
From: "Michael R. Hines"
As requested, the protocol now includes memory unpinning support.
This has been implemented in a non-optimized manner, in such a way
that one could devise an LRU or other workload-specific information
on top of the basic mechanism to influence the way unpinning happens
du
Without this patch the guest physical bits are advertised as 40, not
44 or more depending on the hardware capability of the host.
That leads to guest kernel crashes with injection of page faults 9
(see oops: 0009) as bits above 40 in the guest pagetables are
considered reserved.
exregion-0206 [32
Signed-off-by: Paolo Bonzini
---
block/raw-posix.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/block/raw-posix.c b/block/raw-posix.c
index d011cfd..1b41ea3 100644
--- a/block/raw-posix.c
+++ b/block/raw-posix.c
@@ -1128,6 +1128,9 @@ static int64_t coroutine_fn
raw_co_get_block_status(
From: "Michael R. Hines"
This takes advantages of the previous patches:
1. use the new QEMUFileOps hook 'save_page'
2. call out to the right accessor methods to invoke
the iteration hooks defined in QEMUFileOps
Reviewed-by: Paolo Bonzini
Reviewed-by: Chegu Vinod
Tested-by: Chegu Vinod
Te
Il 16/07/2013 18:59, Hervé Poussineau ha scritto:
> Paolo Bonzini a écrit :
>> -BEGIN PGP SIGNED MESSAGE-
>> Hash: SHA1
>>
>> Il 16/07/2013 09:18, Jan Kiszka ha scritto:
>>> Sorry for sending out invitations and then being late to this party
>>> - vacation. What is the status now? Do we hav
bdrv_is_allocated can detect coroutine context and go through a fast
path, similar to other block layer functions.
Signed-off-by: Paolo Bonzini
---
block.c | 24 +++-
block/raw.c | 2 +-
block/stream.c| 4 ++--
include/block/block.h | 2 --
Alternatively, this could use a "discard zeroes data" flag returned
by bdrv_get_info.
Signed-off-by: Paolo Bonzini
---
block.c | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/block.c b/block.c
index 7ff0716..0c867b8 100644
--- a/block.c
+++ b/block.c
@@ -2977,6 +2977,7
From: "Michael R. Hines"
We were not checking for a valid 'bytes_sent' pointer before accessing it.
Reviewed-by: Eric Blake
Signed-off-by: Michael R. Hines
---
savevm.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/savevm.c b/savevm.c
index e0491e7..03fc4d9 100644
---
Paolo Bonzini a écrit :
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Il 16/07/2013 09:18, Jan Kiszka ha scritto:
Sorry for sending out invitations and then being late to this party
- vacation. What is the status now? Do we have a short-term plan to
avoid the regression or is this better solved
From: "Michael R. Hines"
This gives RDMA shared access to madvise() on the destination side
when an entire chunk is found to be zero.
Reviewed-by: Juan Quintela
Reviewed-by: Paolo Bonzini
Reviewed-by: Chegu Vinod
Tested-by: Chegu Vinod
Tested-by: Michael R. Hines
Signed-off-by: Michael R. H
On 07/16/2013 10:29 AM, Paolo Bonzini wrote:
> This series adds a subcommand to "map" that can dump file metadata.
> Metadata that is dumped includes:
>
> - whether blocks are allocated in bs->file and, if so, where
>
> - whether blocks are zero
>
> - whether data is read from bs or bs->backing_
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Il 16/07/2013 18:56, Eric Blake ha scritto:
> On 07/16/2013 10:29 AM, Paolo Bonzini wrote:
>> This series adds a subcommand to "map" that can dump file
>> metadata. Metadata that is dumped includes:
>>
>> - whether blocks are allocated in bs->file and
From: "Michael R. Hines"
This patch is in preparation for the next ones: Until now the MIG_STATE_SETUP
state was not really a 'formal' state. It has been used as a 'zero' state
and QEMU has been unconditionally transitioning into this state when
the QMP migrate command was called. In preparation
From: "Michael R. Hines"
As described in the previous patch, until now, the MIG_STATE_SETUP
state was not really a 'formal' state. It has been used as a 'zero' state
(what we're calling 'NONE' here) and QEMU has been unconditionally transitioning
into this state when the QMP migration command was
Il 16/07/2013 18:14, Alex Bligh ha scritto:
> Paolo,
>
>>> 3. aio_poll calls aio_bh_poll. If this returns true, this indicates
>>> at least one non-idle bh exists, which causes aio_poll not to
>>> block.
>>
>> No, this indicates that at least one scheduled non-idle bh exist*ed*,
>> which cause
On 07/16/2013 11:15:51 AM, Fabien Chouteau wrote:
On 07/16/2013 05:37 PM, Alexander Graf wrote:
> On 07/16/2013 05:28 PM, Fabien Chouteau wrote:
>> On 07/16/2013 04:06 AM, Scott Wood wrote:
>>> On 07/10/2013 12:10:02 PM, Fabien Chouteau wrote:
This implementation doesn't include ring priorit
From: "Michael R. Hines"
Changes since v2:
- trivial bugfix
- re-ran checkpatch
Michael R. Hines (8):
rdma: update documentation to reflect new unpin support
rdma: bugfix: ram_control_save_page()
rdma: introduce ram_handle_compressed()
rdma: core logic
rdma: send pc.ram
rdma: allow s
From: "Michael R. Hines"
Using the previous patches, we're now able to timestamp the SETUP
state. Once we have this time, let the user know about it in the
schema.
Reviewed-by: Juan Quintela
Reviewed-by: Eric Blake
Signed-off-by: Michael R. Hines
---
hmp.c |4
Now that bdrv_is_allocated detects coroutine context, the two can
use the same code.
Signed-off-by: Paolo Bonzini
---
block.c | 46 --
block/commit.c| 6 +++---
block/mirror.c| 4 ++--
block/stream.c| 4 ++--
in
qemu-img convert is assuming "that sectors which are unallocated in the input
image are present in both the output's and input's base images", but it is
only doing this if the output image is zero initialized. And checking if
the output image is zero initialized does not make much sense if the
out
Signed-off-by: Paolo Bonzini
---
block/raw-posix.c | 17 ++---
1 file changed, 10 insertions(+), 7 deletions(-)
diff --git a/block/raw-posix.c b/block/raw-posix.c
index dbc65b0..d011cfd 100644
--- a/block/raw-posix.c
+++ b/block/raw-posix.c
@@ -1089,7 +1089,7 @@ static int64_t corout
Signed-off-by: Paolo Bonzini
---
v1->v2: handle extents that are stored in bs->file
block/cow.c | 8 +++-
block/qcow.c | 9 -
block/qcow2.c| 16 ++--
block/qed.c | 35 ---
block/sheepdog.c | 2 +-
block/vdi.c
This command dumps the metadata of an entire chain, in either tabular or JSON
format.
Signed-off-by: Paolo Bonzini
---
v1->v2: mention encrypted clusters, use PRId64
qemu-img-cmds.hx | 6 ++
qemu-img.c | 185 +++
2 files change
Define the return value of get_block_status. Bits 0, 1, 2 and 9-62
are valid; bit 63 (the sign bit) is reserved for errors. Bits 3-7
are left for future extensions.
The return code is compatible with the old is_allocated API: returning
just 0 or 1 (aka BDRV_BLOCK_DATA) will not cause any behavio
For now, bdrv_get_block_status is just another name for bdrv_is_allocated.
The next patches will add more flags.
This also touches all block drivers with a mostly mechanical rename. The
sole exception is cow; because it calls cow_co_is_allocated from the read
code, we keep that function and make
Some bdrv_is_allocated callers do not expect errors, but the fallback
in qcow2.c might make other callers trip on assertion failures or
infinite loops.
Fix the callers to always look for errors.
Cc: qemu-sta...@nongnu.org
Signed-off-by: Paolo Bonzini
---
v1->v2: modify error message, add
Signed-off-by: Paolo Bonzini
---
block.c | 10 +-
1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/block.c b/block.c
index 557ce29..2d7d71f 100644
--- a/block.c
+++ b/block.c
@@ -2977,7 +2977,7 @@ static int64_t coroutine_fn
bdrv_co_get_block_status(BlockDriverState *bs,
Protocols return raw data, so you can assume the offsets to pass
through unchanged.
Signed-off-by: Paolo Bonzini
---
block.c | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/block.c b/block.c
index 0c867b8..557ce29 100644
--- a/block.c
+++ b/block.c
@@ -2991,7 +2991,11 @@
Do not do two reads for each sector; load each sector of the bitmap
and use bitmap operations to process it.
Writes are still dog slow!
Signed-off-by: Paolo Bonzini
---
v1->v2: use BDRV_SECTOR_SIZE, not 512 for bitmap array length
block/cow.c | 54 --
1 - 100 of 206 matches
Mail list logo
