Il 24/09/2012 20:14, Michael Roth ha scritto:
>>> > > I went with qUppercase because it avoids all the previous issues with
>>> > > using leading underscores, and it's reserved in terms of QEMU coding
>>> > > guidelines as far as I can tell (we generally require leading capital
>>> > > for typedefs
Il 24/09/2012 20:24, Juan Quintela ha scritto:
> Now, we give exactly the same error, just that we return the error
> directly from blk_mig_save_dirty_block() instead of storing it on
> QEMUFile and getting it through qemu_file_get_error().
Right you are. Apart from the do/while,
Reviewed-by: Pa
Il 24/09/2012 23:55, Luiz Capitulino ha scritto:
> v3
>
> - qmp_add_client(): don't use errp to test for error
>
> Luiz Capitulino (1):
> qapi: convert add_client
>
> Paolo Bonzini (2):
> pci-assign: use monitor_handle_fd_param
> monitor: add Error * argument to monitor_get_fd
>
> dump.
On 09/25/2012 07:05 AM, Michael S. Tsirkin wrote:
There is no out sg for TX, so used buf length for tx
should always be 0.
According to the spec, the len is "Total length of the descriptor chain
which was used (written to)". So I wonder if we need to pass the len
here, it looks useful for gue
On 24/09/12 19:11, Orit Wasserman wrote:
Changes from v4:
- Rename ConnectHandler to NonBlockingConnectHandler
- move function comments to functions definitions
- move connect_state allocation to outside of the loop
- fix comments text
Changes from v3:
- a
Hi Gerd,
Any suggestion?
Many thanks. -:)
On Wed, Sep 19, 2012 at 1:59 AM, Alexander Graf wrote:
> On 09/18/2012 07:36 PM, Alexander Graf wrote:
>>
>> On 09/03/2012 07:25 AM, Li Zhang wrote:
>>>
>>> When -usb option is used, global varible usb_enabled is set.
>>> And all the plaform will create
On Mon, Sep 24, 2012 at 07:34:00PM -0500, Anthony Liguori wrote:
> "Michael S. Tsirkin" writes:
>
> > Add API to copy part of iovec safely.
> >
> > Signed-off-by: Michael S. Tsirkin
> > ---
> > iov.c | 23 +++
> > iov.h | 9 +
> > 2 files changed, 32 insertions(+)
>
On Mon, Sep 24, 2012 at 07:37:40PM -0500, Anthony Liguori wrote:
> "Michael S. Tsirkin" writes:
>
> > Avoid tweaking iovec during receive. This removes
> > the need to copy the vector.
> > Note: we already have an evil cast in work_around_broken_dhclient
> > and unfortunately this adds another on
"Michael S. Tsirkin" writes:
> We know offset is 0, assert that.
>
> Signed-off-by: Michael S. Tsirkin
Reviewed-by: Anthony Liguori
Regards,
Anthony Liguori
> ---
> hw/virtio-net.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/hw/virtio-net.c b/hw/virtio-net
"Michael S. Tsirkin" writes:
> Now that we know host hdr length, we don't need to
> duplicate the logic in receive_hdr: caller can
> figure out the offset itself.
>
> Signed-off-by: Michael S. Tsirkin
Reviewed-by: Anthony Liguori
Regards,
Anthony Liguori
> ---
> hw/virtio-net.c | 18 ++
"Michael S. Tsirkin" writes:
> Avoid magling iov manually: use safe iov operations
> for processing packets incoming to guest.
> This also removes the requirement for virtio header to
> fit the first s/g entry exactly.
>
> Signed-off-by: Michael S. Tsirkin
Reviewed-by: Anthony Liguori
Regards
"Michael S. Tsirkin" writes:
> Avoid tweaking iovec during receive. This removes
> the need to copy the vector.
> Note: we already have an evil cast in work_around_broken_dhclient
> and unfortunately this adds another one.
> const on buf is ignored by this function anyway so arguably
> this is no
"Michael S. Tsirkin" writes:
> Add API to copy part of iovec safely.
>
> Signed-off-by: Michael S. Tsirkin
> ---
> iov.c | 23 +++
> iov.h | 9 +
> 2 files changed, 32 insertions(+)
>
> diff --git a/iov.c b/iov.c
> index c6a66f0..0dfcb28 100644
> --- a/iov.c
> +++ b
There's no reason to query header support at random
times: at load or feature query.
Driver also might not query functions.
Cleaner to do it at device init.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 23 ++-
1 file changed, 14 insertions(+), 9 deletions(-)
diff
Avoid mangling iovec manually: use safe iov_*
functions.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 31 +++
1 file changed, 19 insertions(+), 12 deletions(-)
diff --git a/hw/virtio-net.c b/hw/virtio-net.c
index 6e6f5f3..23feb21 100644
--- a/hw/virtio-net
Modern linux supports arbitrary header size,
which makes it possible to pass mrg buf header
to tap directly without iovec mangling.
Use this capability when it is there.
This removes the need to deal with it in
vhost-net as we do now.
Signed-off-by: Michael S. Tsirkin
---
hw/vhost_net.c | 13 -
During packet filtering, we can now use host hdr len
to offset incoming buffer unconditionally.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/hw/virtio-net.c b/hw/virtio-net.c
index d9a9f8f..c65dead 100644
--- a/hw/vi
Avoid tweaking iovec during receive. This removes
the need to copy the vector.
Note: we already have an evil cast in work_around_broken_dhclient
and unfortunately this adds another one.
const on buf is ignored by this function anyway so arguably
this is not making things much worse.
Signed-off-by:
We know offset is 0, assert that.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/hw/virtio-net.c b/hw/virtio-net.c
index 3a0d1a7..6e6f5f3 100644
--- a/hw/virtio-net.c
+++ b/hw/virtio-net.c
@@ -630,6 +630,7 @@ static
virtio net is using lots of manual mangling of iovecs;
I don't see any obvious crashes but it does look scary since guest can
manipulate iovecs. This patchset changes it to use iov_* functions
which are safer.
On a recent linux kernel, new code will actually do less
iovec processing because it is
iov_from_buf does not change iov, make it const.
Signed-off-by: Michael S. Tsirkin
---
iov.c | 2 +-
iov.h | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/iov.c b/iov.c
index 60705c7..c6a66f0 100644
--- a/iov.c
+++ b/iov.c
@@ -26,7 +26,7 @@
# include
#endif
-size_t io
Remove code duplication using guest header length that we track.
Drop specific layout requirement for rx buffers: things work
using generic iovec functions in any case.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 9 ++---
1 file changed, 2 insertions(+), 7 deletions(-)
diff --gi
Avoid magling iov manually: use safe iov operations
for processing packets incoming to guest.
This also removes the requirement for virtio header to
fit the first s/g entry exactly.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 25 ++---
1 file changed, 14 insertion
There is no out sg for TX, so used buf length for tx
should always be 0.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/virtio-net.c b/hw/virtio-net.c
index 2381ee5..d9a9f8f 100644
--- a/hw/virtio-net.c
+++ b/hw/vi
Now that we know host hdr length, we don't need to
duplicate the logic in receive_hdr: caller can
figure out the offset itself.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 18 +++---
1 file changed, 7 insertions(+), 11 deletions(-)
diff --git a/hw/virtio-net.c b/hw/virti
offset of accessed buffer is calculated using iov_length, so it
can exceed accessed len. If that happens
math in len - offset wraps around, and size becomes wrong.
As real value is 0, so this is harmless but unnecessary.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio.c | 2 +-
1 file changed, 1
Add API to copy part of iovec safely.
Signed-off-by: Michael S. Tsirkin
---
iov.c | 23 +++
iov.h | 9 +
2 files changed, 32 insertions(+)
diff --git a/iov.c b/iov.c
index c6a66f0..0dfcb28 100644
--- a/iov.c
+++ b/iov.c
@@ -228,3 +228,26 @@ void iov_hexdump(const st
Tracking these in device state instead of
re-calculating on each packet. No functional
changes.
Signed-off-by: Michael S. Tsirkin
---
hw/virtio-net.c | 29 +
1 file changed, 17 insertions(+), 12 deletions(-)
diff --git a/hw/virtio-net.c b/hw/virtio-net.c
index 649074
On 09/24/2012 08:30 PM, Kevin Wolf wrote:
> Am 24.09.2012 13:32, schrieb ching:
>> Hi all,
>>
>> My host is qemu-1.1.1 and x64 kernel 3.5.4. The guest is using aio="native"
>>
>> I am trying to use unsafe cache mode to boost i/o performance.
> aio=native requires the image to be opened with O_DIREC
Three places in the interrupt code did we not honor the mask.
Cc: Alexander Graf
Signed-off-by: Richard Henderson
---
target-s390x/helper.c | 7 ---
target-s390x/misc_helper.c | 3 ++-
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/target-s390x/helper.c b/target-s390x/h
This is already handled generically in cpu_exec.
Cc: Alexander Graf
Signed-off-by: Richard Henderson
---
target-s390x/translate.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/target-s390x/translate.c b/target-s390x/translate.c
index 6fa76a0..4cc9225 100644
--- a/target-s390x/translate.c
+
Cc: Alexander Graf
Signed-off-by: Richard Henderson
---
target-s390x/translate.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/target-s390x/translate.c b/target-s390x/translate.c
index 3214783..6fa76a0 100644
--- a/target-s390x/translate.c
+++ b/target-s390x/transl
The first patch applies to most targets, in wanting to produce
good dumps for -d op_opt without -d op.
The next three add dump_insn support to the targets lacking it.
The final three fix irritations that apply to s390x. I didn't
survey the other targets for similar mistakes.
r~
Richard Hende
Cc: Guan Xuetao
Signed-off-by: Richard Henderson
---
target-unicore32/translate.c | 4
1 file changed, 4 insertions(+)
diff --git a/target-unicore32/translate.c b/target-unicore32/translate.c
index b786a6b..36f4f2f 100644
--- a/target-unicore32/translate.c
+++ b/target-unicore32/translate.
For all targets that currently call tcg_gen_debug_insn_start,
add CPU_LOG_TB_OP_OPT to the condition that gates it.
This is useful for comparing optimization dumps, when the
pre-optimization dump is merely noise.
Signed-off-by: Richard Henderson
---
target-alpha/translate.c | 2 +-
target-
The blank lines inside the single dump make it difficult for the
eye to pick out the block. Worse, with interior newlines, but
no blank line following, the PSW line appears to belong to the
next dump block.
Cc: Alexander Graf
Signed-off-by: Richard Henderson
---
target-s390x/translate.c | 22 +
Cc: Paul Brook
Signed-off-by: Richard Henderson
---
target-m68k/translate.c | 4
1 file changed, 4 insertions(+)
diff --git a/target-m68k/translate.c b/target-m68k/translate.c
index fb707f2..451ef74 100644
--- a/target-m68k/translate.c
+++ b/target-m68k/translate.c
@@ -2953,6 +2953,10 @@ s
Also fixes a few issues while there:
1. The fd returned by monitor_get_fd() leaks in most error conditions
2. monitor_get_fd() return value is not checked. Best case we get
an error that is not correctly reported, worse case one of the
functions using the fd (with value of -1) will explo
From: Paolo Bonzini
There is no need to open-code the choice between a file descriptor
number or a named one. Just use monitor_handle_fd_param, which
also takes care of printing the error message.
Signed-off-by: Paolo Bonzini
Signed-off-by: Luiz Capitulino
---
hw/kvm/pci-assign.c | 12 +++---
From: Paolo Bonzini
Signed-off-by: Paolo Bonzini
Signed-off-by: Luiz Capitulino
---
dump.c | 3 +--
migration-fd.c | 2 +-
monitor.c | 15 +--
monitor.h | 2 +-
4 files changed, 12 insertions(+), 10 deletions(-)
diff --git a/dump.c b/dump.c
index 2bf8d8d..1a3c
v3
- qmp_add_client(): don't use errp to test for error
Luiz Capitulino (1):
qapi: convert add_client
Paolo Bonzini (2):
pci-assign: use monitor_handle_fd_param
monitor: add Error * argument to monitor_get_fd
dump.c | 3 +--
hw/kvm/pci-assign.c | 12 +++-
migration
On 09/24/2012 02:37 PM, Alex Barcelo wrote:
> just finished a git-bisect and I found this... and now I do not fully
> understand why I have the problem.
>
> To replicate the error (in a i386 machine, at least):
> $ make clean && ./configure --enable-debug && make -j && make install
> [Note: I trie
The table that was recently added for hppa is generally usable.
And with the renumbering of the TCG_COND constants it's not too
difficult to compute rather than have a table.
Signed-off-by: Richard Henderson
---
tcg/hppa/tcg-target.c | 17 ++---
tcg/tcg.h | 17 +++
just finished a git-bisect and I found this... and now I do not fully
understand why I have the problem.
To replicate the error (in a i386 machine, at least):
$ make clean && ./configure --enable-debug && make -j && make install
[Note: I tried both ppc and i386 targets, so doesn't seem machine-dep
On Mon, 24 Sep 2012 10:39:38 +0100
Mel Gorman wrote:
> On Fri, Sep 21, 2012 at 02:36:56PM -0700, Andrew Morton wrote:
>
> > Also, what has to be done to avoid the polling altogether? eg/ie, zap
> > a pageblock's PB_migrate_skip synchronously, when something was done to
> > that pageblock which
There are several cases that can be handled easier inside both
translators and code generators if we have out-of-band values
for conditions. It's easy enough to handle ALWAYS and NEVER in
the natural way inside the tcg middle-end.
Signed-off-by: Richard Henderson
---
tcg/arm/tcg-target.c | 2
Before we rearrange the TCG_COND enumeration, add a predicate for
the (single) use of comparisons vs TCGCond.
Signed-off-by: Richard Henderson
---
tcg/s390/tcg-target.c | 2 +-
tcg/tcg.h | 5 +
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/tcg/s390/tcg-target.c b/
The second patch comes from the target-s390 patch set I've been
working on. The first patch has been split out from that. The
third patch provides tcg_high_cond globally, as the 32-bit sparc
port suffers from the same bug that afflicted hppa.
Richard Henderson (3):
tcg: Add is_unsigned_cond
When movcond_i32 is available we can further reduce the generated
op count from 12 to 6, and the generated code size on i686 from
88 to 74 bytes.
Signed-off-by: Richard Henderson
---
tcg/tcg-op.h | 22 +++---
1 file changed, 15 insertions(+), 7 deletions(-)
diff --git a/tcg/tcg-
Avoiding 64-bit arithmetic (outside of the compare) reduces the
generated op count from 15 to 12, and the generated code size on
i686 from 105 to 88 bytes.
Signed-off-by: Richard Henderson
---
tcg/tcg-op.h | 20
1 file changed, 20 insertions(+)
diff --git a/tcg/tcg-op.h b/t
Changes v2->v3:
Rebase with the first 5 patches committed.
Fix 32/64-bit compile problems. Oops.
r~
Richard Henderson (2):
tcg: Streamline movcond_i64 using 32-bit arithmetic
tcg: Streamline movcond_i64 using movcond_i32
tcg/tcg-op.h | 28
1 file changed,
On Mon, 2012-09-24 at 22:01 +0200, Michael S. Tsirkin wrote:
> On Mon, Sep 24, 2012 at 12:08:16PM -0600, Alex Williamson wrote:
> > When creating structure names in CamelCase, it's easy to have
> > back-to-back capitals when using acronyms (ex. PCIINTxRoutingNotifier,
> > QEMUSGList, VFIOINTx). In
On Mon, Sep 24, 2012 at 12:08:16PM -0600, Alex Williamson wrote:
> When creating structure names in CamelCase, it's easy to have
> back-to-back capitals when using acronyms (ex. PCIINTxRoutingNotifier,
> QEMUSGList, VFIOINTx). In the worst case these can look like macros,
> but even adjoining a si
Am 24.09.2012 21:28, schrieb Eric Blake:
On 08/22/2012 01:42 PM, Stefan Weil wrote:
GLib 2.0 include files use __printf__ for the format attribute
which resolves to native format strings on w32 hosts.
QEMU wants standard format strings instead of native format
strings, so we simply change any d
On 08/22/2012 01:42 PM, Stefan Weil wrote:
> GLib 2.0 include files use __printf__ for the format attribute
> which resolves to native format strings on w32 hosts.
>
> QEMU wants standard format strings instead of native format
> strings, so we simply change any declaration with __printf__
> to us
On 24 September 2012 20:08, Stefan Weil wrote:
> It is used to avoid dynamic memory allocation for qemu_irq arrays
> with known size or single qemu_irq variables.
This patch is going to collide with Peter Crosthwaite's patch
that allows an irq array to be extended.
Also, the memory allocated wit
qi was allocated using g_malloc and never released.
It is not necessary to use dynamic memory allocation
for qi because it is copied to entries in s->timer
and not used otherwise.
Signed-off-by: Stefan Weil
---
hw/arm_timer.c |5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff -
outsignals was allocated using g_malloc and never released.
It is not necessary to use dynamic memory allocation
for that array because its entries are copied
and not used otherwise. Only 7 entries are used.
lcd_hsync can be initialized directly instead of copying the first entry
of a dynamically
It is used to avoid dynamic memory allocation for qemu_irq arrays
with known size or single qemu_irq variables.
Signed-off-by: Stefan Weil
---
hw/irq.c | 15 +--
hw/irq.h |4
2 files changed, 13 insertions(+), 6 deletions(-)
diff --git a/hw/irq.c b/hw/irq.c
index d413a0b.
There are several memory leaks in QEMU which are caused by calling
qemu_allocate_irqs.
Some of these leaks are fixed here in the 2nd and 3rd patch.
More can be fixed using the same pattern.
The first patch adds a new function which avoids unnecessary
memory allocation.
[PATCH 1/3] irq: Add new f
Implement state save/restore for the DS1338. This requires
the usual minor adjustment of types in the state struct to
get fixed-width ones with vmstate macros.
Signed-off-by: Peter Maydell
---
hw/ds1338.c | 27 ++-
1 file changed, 22 insertions(+), 5 deletions(-)
diff
Clang's static analyzer drew my attention to the mishandling of the
register pointer in ds1338_send(); one thing led to another and I fixed
a few other things while I was there.
There seems a reasonable chance that the overrun of nvram[] is
guest-exploitable, but I assume nobody treats realview or
Correct several deficiencies in the handling of the register pointer:
* it should wrap around after 0x3f, not 0xff
* guard against the caller handing us an out of range pointer
(on h/w this can never happen, because only a 7 bit value is
transferred over the I2C bus)
* there was confusion
The DS1338 datasheet documents that the current time is captured into
the secondary registers when the register pointer wraps round to zero
as well as at a START condition. Implement this.
Signed-off-by: Peter Maydell
---
hw/ds1338.c | 59 ++-
Using the virtqueue_avail_bytes() function had an unnecessarily
crippling effect on the number of bytes needed by the guest as reported
to the chardev layer in the can_read() callback.
Using the new virtqueue_get_avail_bytes() function will let us advertise
the exact number of bytes we can send to
The virtqueue_avail_bytes() function counts bytes in an int. Use an
unsigned int instead.
Signed-off-by: Amit Shah
---
hw/virtio.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/hw/virtio.c b/hw/virtio.c
index 209c763..4c9e20a3 100644
--- a/hw/virtio.c
+++ b/hw/virtio
The current virtqueue_avail_bytes() is oddly named, and checks if a
particular number of bytes are available in a vq. A better API is to
fetch the number of bytes available in the vq, and let the caller do
what's interesting with the numbers.
Introduce virtqueue_get_avail_bytes(), which returns t
This is a rebased version of the patchset sent earlier.
The current virtqueue_avail_bytes() is a weird API: it's oddly-named:
doesn't tell us what the API is going to do, and also suits just one
use-case (that in virtio-net.c).
Introduce virtqueue_get_avail_bytes(), which returns the number of
by
The 'struct tm now' field in the state structure is in fact only
ever used as a temporary (the actual RTC state is held in 'offset').
Remove it from the state structure in favour of using local variables
to avoid confusion about whether it needs to be saved on migration.
Signed-off-by: Peter Mayde
Paolo Bonzini wrote:
> Il 21/09/2012 10:47, Juan Quintela ha scritto:
>> Signed-off-by: Juan Quintela
>> ---
>> @@ -635,18 +639,15 @@ static int block_save_complete(QEMUFile *f, void
>> *opaque)
>> all async read completed */
>> assert(block_mig_state.submitted == 0);
>>
>
> Not cl
On Sat, Sep 22, 2012 at 02:33:52PM +, Blue Swirl wrote:
> On Fri, Sep 21, 2012 at 4:24 PM, Michael Roth
> wrote:
> > On Fri, Sep 21, 2012 at 05:57:42PM +0200, Paolo Bonzini wrote:
> >> Il 21/09/2012 16:07, Michael Roth ha scritto:
> >> >
> >> > QIDL_DECLARE(SerialDevice) {
> >> >
Paolo Bonzini wrote:
> Il 21/09/2012 10:47, Juan Quintela ha scritto:
>> Old code used !io_thread to know if a thread was an vcpu or not. That
>> fails when we introduce the iothread.
>>
>> Signed-off-by: Juan Quintela
>> ---
>> cpus.c | 9 +++--
>> 1 file changed, 7 insertions(+), 2 delet
When creating structure names in CamelCase, it's easy to have
back-to-back capitals when using acronyms (ex. PCIINTxRoutingNotifier,
QEMUSGList, VFIOINTx). In the worst case these can look like macros,
but even adjoining a single, all-caps acronym makes it more difficult
to interpret. For example
On 23 September 2012 07:41, Alexander Graf wrote:
> +void qemu_devtree_dumpdtb(void *fdt, int size)
> +{
> +QemuOpts *machine_opts;
> +
> +machine_opts = qemu_opts_find(qemu_find_opts("machine"), 0);
> +if (machine_opts) {
> +const char *dumpdtb = qemu_opt_get(machine_opts, "du
On Mon, Sep 24, 2012 at 05:28:35PM +0100, Peter Maydell wrote:
> In float16_to_float32, when returning an infinity, just pass zero
> as the mantissa argument to packFloat32(), rather than shifting
> a value which we know must be zero.
>
> Signed-off-by: Peter Maydell
> ---
> Spotted by the clang
On Mon, Sep 24, 2012 at 06:52:29PM +0200, Markus Armbruster wrote:
> Jason Baron writes:
>
> > On Fri, Sep 21, 2012 at 04:05:14PM +0200, Markus Armbruster wrote:
> >> Jason Baron writes:
> >>
> >> > From: Isaku Yamahata
> >> >
> >> > Introduce a helper function which initializes the ahci port
Jason Baron writes:
> On Fri, Sep 21, 2012 at 04:05:14PM +0200, Markus Armbruster wrote:
>> Jason Baron writes:
>>
>> > From: Isaku Yamahata
>> >
>> > Introduce a helper function which initializes the ahci port with
>> > ide devices.
>> > It will be used by q35 support.
>> >
>> > Cc: Alexander
Signed-off-by: Don Slutz
---
target-i386/cpu.h |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/target-i386/cpu.h b/target-i386/cpu.h
index ebb3498..254ddef 100644
--- a/target-i386/cpu.h
+++ b/target-i386/cpu.h
@@ -812,6 +812,10 @@ typedef struct CPUX86State {
uin
The dumpdtb code can be useful in more places than just for e500. Move it
to a generic place.
Signed-off-by: Alexander Graf
---
device_tree.c | 22 ++
device_tree.h |2 ++
hw/ppc/e500.c | 15 +--
3 files changed, 25 insertions(+), 14 deletions(-)
diff --g
This patch adds support for the dumpdtb machine option to all ARM boards that
have device tree support.
Signed-off-by: Alexander Graf
---
hw/arm_boot.c |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/hw/arm_boot.c b/hw/arm_boot.c
index a6e9143..0830a62 100644
--- a/hw/ar
Also known as Paravirtualization vendor.
This change is based on:
Microsoft Hypervisor CPUID Leaves:
http://msdn.microsoft.com/en-us/library/windows/hardware/ff542428%28v=vs.85%29.aspx
Linux kernel change starts with:
http://fixunix.com/kernel/538707-use-cpuid-communicate-hypervisor.html
Al
On 2012-09-24 09:28, Peter Maydell wrote:
> In float16_to_float32, when returning an infinity, just pass zero
> as the mantissa argument to packFloat32(), rather than shifting
> a value which we know must be zero.
>
> Signed-off-by: Peter Maydell
Reviewed-by: Richard Henderson
r~
In float16_to_float32, when returning an infinity, just pass zero
as the mantissa argument to packFloat32(), rather than shifting
a value which we know must be zero.
Signed-off-by: Peter Maydell
---
Spotted by the clang static analyzer. This brings this code into line with
the other float-to-floa
This was taken from:
http://article.gmane.org/gmane.comp.emulators.kvm.devel/22643
Signed-off-by: Don Slutz
---
target-i386/cpu.c | 32
1 files changed, 32 insertions(+), 0 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index 8bb20c7..b77dbf
From: Jeff Cody
This is based on Supriya Kannery's bdrv_reopen() patch series.
This provides a transactional method to reopen multiple
images files safely.
Image files are queue for reopen via bdrv_reopen_queue(), and the
reopen occurs when bdrv_reopen_multiple() is called. Changes are
staged
From: Kevin Shanahan
If readonly=on is given at device creation time, the ->readonly flag
needs to be set in the block driver state for this device so that
readonly-ness is preserved across media changes (qmp change command).
Similarly, to preserve the snapshot property requires ->open_flags to
b
From: Jeff Cody
There is currently nothing that needs to be done for VPC image
file reopen.
Signed-off-by: Jeff Cody
Signed-off-by: Kevin Wolf
---
block/vpc.c |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git a/block/vpc.c b/block/vpc.c
index c0b82c4..b6bf52f 100644
Also known as Paravirtualization vendor.
This change is based on:
Microsoft Hypervisor CPUID Leaves:
http://msdn.microsoft.com/en-us/library/windows/hardware/ff542428%28v=vs.85%29.aspx
Linux kernel change starts with:
http://fixunix.com/kernel/538707-use-cpuid-communicate-hypervisor.html
Al
From: Jeff Cody
The keep_read_only flag is no longer used, in favor of the bdrv
flag BDRV_O_ALLOW_RDWR.
Signed-off-by: Jeff Cody
Signed-off-by: Kevin Wolf
---
block.c |2 --
block_int.h |1 -
2 files changed, 0 insertions(+), 3 deletions(-)
diff --git a/block.c b/block.c
index 84
From: Jeff Cody
The aligned_buf pointer and aligned_buf size are no longer used in
raw_posix.c, so remove all references to them.
Signed-off-by: Jeff Cody
Signed-off-by: Kevin Wolf
---
block/raw-posix.c | 21 +
1 files changed, 1 insertions(+), 20 deletions(-)
diff --gi
From: Jeff Cody
Currently, bdrv_commit() reopens images r/w itself, via risky
_delete() and _open() calls. Use the new safe method for drive reopen.
Signed-off-by: Jeff Cody
Signed-off-by: Kevin Wolf
---
block.c | 48 +---
1 files changed, 5 inser
From: Jeff Cody
Rather than check for a non-NULL aligned_buf to determine if
raw_aio_submit needs to check for alignment, check for the presence
of BDRV_O_NOCACHE in the bs->open_flags.
Signed-off-by: Jeff Cody
Signed-off-by: Kevin Wolf
---
block/raw-posix.c |2 +-
1 files changed, 1 inse
From: Jeff Cody
I believe the bs->keep_read_only flag is supposed to reflect
the initial open state of the device. If the device is initially
opened R/O, then commit operations, or reopen operations changing
to R/W, are prohibited.
Currently, the keep_read_only flag is only accurate for the acti
On Fri, Sep 14, 2012 at 05:01:35PM -0600, Alex Williamson wrote:
> Same goodness as v4, plus:
>
> - Addressed comments by Blue Swirl (thanks for the review)
>(hopefully w/o breaking anything wrt slow bar endianness)
> - Fixed a couple checkpatch warnings that snuck in
>
> BTW, this works fi
From: Jeff Cody
This is derived from the Supriya Kannery's reopen patches.
This contains the raw-posix driver changes for the bdrv_reopen_*
functions. All changes are staged into a temporary scratch buffer
during the prepare() stage, and copied over to the live structure
during commit(). Upon
From: Jeff Cody
These are the stubs for the file reopen drivers for the raw format.
There is currently nothing that needs to be done by the raw driver
in reopen.
Signed-off-by: Jeff Cody
Signed-off-by: Kevin Wolf
---
block/raw.c | 10 ++
1 files changed, 10 insertions(+), 0 deletio
On Mon, Sep 24, 2012 at 05:06:11PM +0200, Paolo Bonzini wrote:
> The AIO dispatch loop will call QLIST_REMOVE and g_free even if there
> are other pending calls to qemu_aio_wait outside the current one.
>
> Signed-off-by: Paolo Bonzini
> ---
> This is on top of the existing fix that is po
Also known as Paravirtualization level.
This change is based on:
Microsoft Hypervisor CPUID Leaves:
http://msdn.microsoft.com/en-us/library/windows/hardware/ff542428%28v=vs.85%29.aspx
Linux kernel change starts with:
http://fixunix.com/kernel/538707-use-cpuid-communicate-hypervisor.html
Als
From: Jeff Cody
This patch supports reopen for VMDK image files. VMDK extents are added
to the existing reopen queue, so that the transactional model of reopen
is maintained with multiple image files.
Signed-off-by: Jeff Cody
Signed-off-by: Kevin Wolf
---
block/vmdk.c | 35
On 2012-09-22 12:52, Aurelien Jarno wrote:
> On Fri, Sep 21, 2012 at 05:18:13PM -0700, Richard Henderson wrote:
>> For tcg_gen_concat_i32_i64 we only use deposit if the host supports it.
>> For tcg_gen_concat32_i64 even if the host does not, as we get identical
>> code before and after.
>>
>> Note
1 - 100 of 228 matches
Mail list logo
