Am 02.02.2013 01:37, schrieb Andreas Färber:
Hello,
Long announced, here it is after the freeze: My 3rd attempt at CPU subclasses.
v3 is closer to v1 again, slimmed down not to touch x86_def_t much.
It was noticed that in theory a QOM class enumeration could lead to the host
CPU
Am 31.01.2013 18:57, schrieb Andreas Färber:
Hello,
This series propagates M68kCPU arguments along the call chain in preparation
for CPU_COMMON - CPUState interrupt field movements and function changes.
It is based on my current qom-cpu-next queue:
Am 01.02.2013 23:03, schrieb Paolo Bonzini:
We had two copies of a ffs function for longs with subtly different
semantics and, for the one in bitops.h, a confusing name: the result
was off-by-one compared to the library function ffsl.
Unify the functions into one, and solve the name problem
On Sat, Feb 2, 2013 at 5:02 AM, Brian Jackson i...@theiggy.com wrote:
[...]
It probably gets stripped during the install process. Check 'file'
output on both and see.
This is a stupid mistake, which the latter one is *NOT* installed by
the first one. Jackson, thanks for your suggestions ;-)
--
From: Jan Kiszka jan.kis...@siemens.com
Signed-off-by: Jan Kiszka jan.kis...@siemens.com
---
configure |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/configure b/configure
index 0657b1a..cfccec5 100755
--- a/configure
+++ b/configure
@@ -4289,6 +4289,7 @@ FILES=$FILES
Hi Christopher,
Am 02.02.2013 03:38, schrieb Christopher Friedt:
On Fri, Feb 1, 2013 at 1:01 PM, Peter Maydell peter.mayd...@linaro.org
wrote:
OSX generally works for me, with some caveats:
* current master doesn't compile because of a recent patch
related to ffsl; this should be fixed
Am 02.02.2013 11:46, schrieb Jan Kiszka:
From: Jan Kiszka jan.kis...@siemens.com
Signed-off-by: Jan Kiszka jan.kis...@siemens.com
---
configure |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/configure b/configure
index 0657b1a..cfccec5 100755
--- a/configure
Version 5 contained the halted field, that we are about to move from
CPUX86State to CPUState. To avoid inventing new VMSTATE macros for
calculating a negative offset from CPUX86State to the field in CPUState,
rather bump the minimum version from 3 to 6. We're at 12 currently.
Suggested-by: Juan
Move it to qom/cpu.c to avoid build failures depending on include order
of cpu-qom.h and exec/cpu-all.h.
Change opaques of various ..._irq_handler() functions to the
appropriate CPU type to facilitate using cpu_reset_interrupt().
Fix Coding Style issues while at it (missing braces, indentation).
This removes a global per-target function and thus takes us one step
closer to compiling multiple targets into one executable.
It will also allow to override the interrupt handling for certain CPU
families.
Signed-off-by: Andreas Färber afaer...@suse.de
---
cpu-exec.c | 36
Move it to qom/cpu.h to avoid issues with include order.
Change pc_acpi_smi_interrupt() opaque to X86CPU.
Signed-off-by: Andreas Färber afaer...@suse.de
---
cpus.c |2 +-
exec.c |2 +-
hw/alpha_typhoon.c | 10 --
hw/apic.c
Enable ARMCPUInfo to specify a custom class_init functions.
Introduce arm_v7m_class_init() and use it for cortex-m3 model.
Instead of forwarding from arm_cpu_do_interrupt() to do_interrupt_v7m(),
override CPUClass::do_interrupt with arm_v7m_cpu_do_interrupt()
in arm_v7m_class_init().
Hello,
Most of its prerequisites in v1.4.0-rc0 or in qom-cpu-next, this series
proceeds with moving fields from CPU_COMMON / CPU*State to CPUState.
For large Coding Style cleanups it seems too late now (Hard Freeze).
But the two function declarations might be considered an acceptable cleanup.
Il 11/01/2013 17:46, Paolo Bonzini ha scritto:
Trying (unsuccessfully) to break the code and reproduce the effects
of https://bugs.launchpad.net/qemu/+bug/1090558... Well, more
tests never hurt...
Paolo Bonzini (2):
rtc-test: always set register B in its entirety
rtc-test: add
On Tue, Jan 29, 2013 at 8:52 PM, Richard Henderson r...@twiddle.net wrote:
On 01/29/2013 12:06 PM, Blue Swirl wrote:
+static const Test test_u_data[] = {
+{ 1, 1, 0, 1 },
+{ 1, 1, 0, 1 },
+{ -1ull, 2, 1, -2ull },
Shouldn't '1' be '-1'? How can this test pass?
It was never implemented.
Signed-off-by: Andreas Färber afaer...@suse.de
---
target-s390x/cpu.h |1 -
1 Datei geändert, 1 Zeile entfernt(-)
diff --git a/target-s390x/cpu.h b/target-s390x/cpu.h
index 01e59b9..0070c40 100644
--- a/target-s390x/cpu.h
+++ b/target-s390x/cpu.h
@@ -315,7 +315,6
On Wed, Jan 30, 2013 at 11:05 AM, Muhammad Nouman nouman...@gmail.com wrote:
Hi! There is a special segment in Octeon processors (CVMSEG)
which maps a small portion of MIPS64 address space to
Data Cache. This region is used by the user applications
as scratch space and this space is saved by
Reindent, add missing braces and drop/adjust whitespace.
Prepares for CPUArchState-to-CPUState field movements in
cpu_cris_handle_mmu_fault(), do_interruptv10() and do_interrupt().
The remaining functions were so minor that they can be fixed in one go.
Signed-off-by: Andreas Färber
On 02/02/2013 05:13 AM, Anthony Liguori wrote:
Michael Tokarev m...@tls.msk.ru writes:
02.02.2013 00:36, Anthony Liguori wrote:
Michael Tokarev m...@tls.msk.ru writes:
02.02.2013 00:18, Michael Tokarev wrote:
Just a heads-up for now, no real diagnostics or anything like that.
Current git
It was never implemented.
Signed-off-by: Andreas Färber afaer...@suse.de
---
target-lm32/cpu.h |1 -
1 Datei geändert, 1 Zeile entfernt(-)
diff --git a/target-lm32/cpu.h b/target-lm32/cpu.h
index 4e202db..6948d0e 100644
--- a/target-lm32/cpu.h
+++ b/target-lm32/cpu.h
@@ -189,7 +189,6 @@
On Fri, Feb 1, 2013 at 1:53 PM, Stefan Hajnoczi stefa...@redhat.com wrote:
Slirp uses rfds/wfds/xfds more extensively than other QEMU components.
The rarely-used out-of-band TCP data feature is used. That means we
need the full table of select(2) to g_poll(3) events:
rfds - G_IO_IN |
On Wed, Jan 30, 2013 at 4:37 PM, Kevin Wolf kw...@redhat.com wrote:
Am 19.01.2013 17:01, schrieb Blue Swirl:
Perform input tests on random data.
Improvement to code coverage for qapi/string-input-visitor.c
is about 3 percentage points.
Signed-off-by: Blue Swirl blauwir...@gmail.com
Does
On Fri, Feb 1, 2013 at 2:28 PM, Stefan Hajnoczi stefa...@redhat.com wrote:
From: Kevin Wolf kw...@redhat.com
Return -errno instead of -1 on errors and add error checks in some
places that didn't have one. Passing things by reference requires more
correct typing, replaced a few off_ts
On 1 February 2013 22:03, Paolo Bonzini pbonz...@redhat.com wrote:
We had two copies of a ffs function for longs with subtly different
semantics and, for the one in bitops.h, a confusing name: the result
was off-by-one compared to the library function ffsl.
Unify the functions into one, and
OpenBSD appears to be trying to access cp14 crn=0 crm=1 opc1=0 opc2=0.
Implementing this as a no-op allows the image to boot.
I'm trying to track down what this is supposed to be -- it doesn't seem to be
listed in the PXA27x docs.
(Previously we were lax about decoding and tended to print
In comparison to DeviceClass::vmsd, CPU VMState is split in two,
cpu_common and cpu, and uses cpu_index as instance_id instead of -1.
Therefore add a CPU-specific CPUClass::vmsd field.
Unlike the legacy CPUArchState registration, rather register CPUState.
Signed-off-by: Juan Quintela
Ah, found the problem -- it looks like I accidentally transposed the
crm/crn values for all the perf registers in commit dc2a9045c, in the
process of converting from switch statements to the table
representation, and my test Linux image obviously doesn't prod these
registers. I'll send out a
On 02/02/2013 06:30 AM, Peter Maydell wrote:
- * Undefined if no bit exists, so code should check against 0 first.
+ * Returns -1 if no bit exists. Note that compared to the C library
+ * routine ffsl, this one returns one less.
*/
Do any of our callers actually use the -1 on 0 input
When the pxa2xx performance counter related cp14 registers were converted
from a switch-statement implementation to the new table driven cpregs
format in commit dc2a9045c, the crn and crm values for all these
registers were accidentally transposed. Fix this mistake, which was
causing OpenBSD for
A few targets already managed to implement cpu_save() and cpu_load()
without defining CPU_SAVE_VERSION that causes them to be registered.
Guard the prototypes with CPU_SAVE_VERSION to avoid this happening again
until all targets are converted to VMState or QIDL.
Signed-off-by: Andreas Färber
Expose vmstate_cpu as vmstate_x86_cpu and hook it up to CPUClass::vmsd.
Adapt opaques and VMState fields to X86CPU. Drop cpu_{save,load}().
Signed-off-by: Andreas Färber afaer...@suse.de
---
target-i386/cpu-qom.h |2 +
target-i386/cpu.c |4 +
target-i386/cpu.h |2 -
Fixed by this patch http://patchwork.ozlabs.org/patch/217673/ which
hopefully will make it into 1.4.
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1110531
Title:
regression booting pxa kernels
Since commit e67db06e9f6d7e514ee2a9b9b769ecd42977f6fb (target-or32: Add
target stubs and QOM cpu) a VMStateDescription existed, but
CPU_SAVE_VERSION was not set, so it was never registered.
Register it through CPUState.
Use a version_id of 1 and specify minimum versions as well.
Cc: Jia Liu
Hello,
In response to patches from Juan [1], this series implements a declarative way
for CPUs to register their CPU-specific VMStateDescription.
There are subtle differences between CPU and devices in how they register
their savevm handlers and VMStateDescription respectively:
* CPUs register
Expose vmstate_cpu as vmstate_lm32_cpu and hook it up to CPUClass::vmsd.
Adapt VMState fields to LM32CPU. Drop cpu_{save,load}().
Signed-off-by: Andreas Färber afaer...@suse.de
---
target-lm32/cpu-qom.h |2 ++
target-lm32/cpu.c |4
target-lm32/cpu.h |2 --
Commit b758aca1f6cdb175634812b79f5560c36c902d00 (target-alpha: Enable
the alpha-softmmu target.) introduced cpu_{save,load}() functions but
didn't define CPU_SAVE_VERSION, so they were never registered.
Drop cpu_{save,load}() and register the VMStateDescription via CPUClass.
This operates on the
This is a quick-n-dirty bandaid to let virtio-net to work on windows
again. Merely removing newly added stuff (max_virtqueue_pairs)
from struct virtio_net_config makes fixes it. Note this field
of the structure is not actually used, as far as I can see - it
is write-only, so it shouldn't affect
Am 01.02.2013 23:26, schrieb Richard Henderson:
The misnamed HOST_LONG_BITS is really HOST_POINTER_BITS. Here we're
explicitly using an unsigned long, rather than uintptr_t, so it is
more correct to select the swap size via ULONG_MAX.
Signed-off-by: Richard Henderson r...@twiddle.net
From: Andreas Färber andreas.faer...@web.de
The libqos driver for omap_i2c currently does not work on Big Endian.
Introduce helpers for reading from and writing to 16-bit armel registers.
This fixes tmp105-test failures on ppc.
Signed-off-by: Andreas Färber andreas.faer...@web.de
---
On 2 February 2013 16:45, Andreas Färber afaer...@suse.de wrote:
From: Andreas Färber andreas.faer...@web.de
The libqos driver for omap_i2c currently does not work on Big Endian.
Introduce helpers for reading from and writing to 16-bit armel registers.
This fixes tmp105-test failures on ppc.
Am 02.02.2013 17:49, schrieb Peter Maydell:
On 2 February 2013 16:45, Andreas Färber afaer...@suse.de wrote:
From: Andreas Färber andreas.faer...@web.de
The libqos driver for omap_i2c currently does not work on Big Endian.
Introduce helpers for reading from and writing to 16-bit armel
Am 02.02.2013 18:44, schrieb Peter Maydell:
On 2 February 2013 17:37, Andreas Färber afaer...@suse.de wrote:
Am 02.02.2013 17:49, schrieb Peter Maydell:
There's nothing special about the OMAP i2c device that I know of:
shouldn't the test code just be using a generic write 16 bit value
to
On 2 February 2013 17:37, Andreas Färber afaer...@suse.de wrote:
Am 02.02.2013 17:49, schrieb Peter Maydell:
There's nothing special about the OMAP i2c device that I know of:
shouldn't the test code just be using a generic write 16 bit value
to memory with appropriate endianness for target CPU
Thank you Peter for the quick response to this bug report.
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1110531
Title:
regression booting pxa kernels
Status in QEMU:
New
Bug description:
On Sat, Feb 2, 2013 at 5:44 PM, Peter Maydell peter.mayd...@linaro.org wrote:
On 2 February 2013 17:37, Andreas Färber afaer...@suse.de wrote:
Am 02.02.2013 17:49, schrieb Peter Maydell:
There's nothing special about the OMAP i2c device that I know of:
shouldn't the test code just be using a
Am 02.02.2013 18:17, schrieb Peter Maydell:
Add explicit braces round an empty for-loop body; this fits
QEMU style and is easier to read than an inconspicuous semicolon
at the end of the line. It also silences a clang warning:
disas/i386.c:4723:49: warning: for loop has empty body
On Sat, 2013-02-02 at 20:42 +0800, Jason Wang wrote:
On 02/02/2013 05:13 AM, Anthony Liguori wrote:
Michael Tokarev m...@tls.msk.ru writes:
02.02.2013 00:36, Anthony Liguori wrote:
Michael Tokarev m...@tls.msk.ru writes:
02.02.2013 00:18, Michael Tokarev wrote:
Just a heads-up for
On Sat, Feb 2, 2013 at 8:18 PM, Peter Maydell peter.mayd...@linaro.org wrote:
On 2 February 2013 18:26, Blue Swirl blauwir...@gmail.com wrote:
On Sat, Feb 2, 2013 at 5:44 PM, Peter Maydell peter.mayd...@linaro.org
wrote:
On 2 February 2013 17:37, Andreas Färber afaer...@suse.de wrote:
Am
On 2 February 2013 18:26, Blue Swirl blauwir...@gmail.com wrote:
On Sat, Feb 2, 2013 at 5:44 PM, Peter Maydell peter.mayd...@linaro.org
wrote:
On 2 February 2013 17:37, Andreas Färber afaer...@suse.de wrote:
Am 02.02.2013 17:49, schrieb Peter Maydell:
libqtest.h has no generic endian-aware
On 2 February 2013 12:40, Blue Swirl blauwir...@gmail.com wrote:
On Wed, Jan 30, 2013 at 4:37 PM, Kevin Wolf kw...@redhat.com wrote:
Am 19.01.2013 17:01, schrieb Blue Swirl:
Perform input tests on random data.
Improvement to code coverage for qapi/string-input-visitor.c
is about 3 percentage
It's OK and expected for visitors to return errors when presented with
the fuzz test's random data. This means the test harness needs to
handle them; check for and free any error after each visitor call,
and only free the string returned by visit_type_str if visit_type_str
succeeded.
This fixes a
Am 02.02.2013 22:19, schrieb Peter Maydell:
It's OK and expected for visitors to return errors when presented with
the fuzz test's random data. This means the test harness needs to
handle them; check for and free any error after each visitor call,
and only free the string returned by
From: Dillon Amburgey dill...@dillona.com
Signed-off-by: Dillon Amburgey dill...@dillona.com
---
linux-user/syscall.c | 22 --
1 files changed, 12 insertions(+), 10 deletions(-)
diff --git a/linux-user/syscall.c b/linux-user/syscall.c
index a148d9f..7344052 100644
---
virtio-net multiqueue added a new field to the config layout. The config
layout is rounded up to the nearest power of 2 and this field caused
field causes the size to go from 8 bytes to 16 bytes.
For reasons unknown, the increase in config size causes the Windows
virtio-net driver to fail. This
The windows PV drivers seem to get very upset if the config size changes.
Even when we disable features, we don't change the config size today which
is a bug.
This patch is a pretty rough way to solve this problem and it only handles
the multiqueue case specifically.
Cc: Michael Tokarev
Signed-off-by: Anthony Liguori aligu...@us.ibm.com
---
hw/s390x/s390-virtio-bus.c | 3 ++-
hw/s390x/virtio-ccw.c | 3 ++-
hw/virtio-pci.c| 3 ++-
hw/virtio.h| 3 ++-
4 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/hw/s390x/s390-virtio-bus.c
Vadim Rozenfeld vroze...@redhat.com writes:
On Sat, 2013-02-02 at 20:42 +0800, Jason Wang wrote:
Have a look at this issue. It was caused by multiqueue patch who adds a
new field to virtio_net_cfg. Not sure multiqueue is the root cause since
I also find even w/o multiqueue, adding any new
Michael Tokarev m...@tls.msk.ru writes:
This is a quick-n-dirty bandaid to let virtio-net to work on windows
again. Merely removing newly added stuff (max_virtqueue_pairs)
from struct virtio_net_config makes fixes it. Note this field
of the structure is not actually used, as far as I can
On 2 February 2013 21:37, Andreas Färber afaer...@suse.de wrote:
Am 02.02.2013 22:19, schrieb Peter Maydell:
It's OK and expected for visitors to return errors when presented with
the fuzz test's random data. This means the test harness needs to
handle them; check for and free any error after
On 2 February 2013 23:04, dill...@dillona.com wrote:
From: Dillon Amburgey dill...@dillona.com
Signed-off-by: Dillon Amburgey dill...@dillona.com
Reviewed-by: Peter Maydell peter.mayd...@linaro.org
-- PMM
Hi Edgar,
Here is one more version of this patch set. I've addressed as many
comments as I could. There is more work to be done, particularly in
moving to the common GPIO api, but that work can be done as a follow on
patch series.
Cheers,
g.
This is a trivial patch to harmonize the coding style on
hw/etraxfs_eth.c. This is in preparation to split off the bitbang mdio
code into a separate file.
Cc: Peter Maydell peter.mayd...@linaro.org
Cc: Paul Brook p...@codesourcery.com
Cc: Edgar E. Iglesias edgar.igles...@gmail.com
Cc: Anthony
Trivial patch to add #defines for defined PHY register address and bit fields
Cc: Peter Maydell peter.mayd...@linaro.org
Cc: Paul Brook p...@codesourcery.com
Cc: Edgar E. Iglesias edgar.igles...@gmail.com
Cc: Anthony Liguori aligu...@us.ibm.com
Cc: Andreas Färber afaer...@suse.de
Signed-off-by:
On 2 February 2013 23:40, Grant Likely grant.lik...@secretlab.ca wrote:
static const VMStateDescription vmstate_smc91c111 = {
@@ -71,6 +76,8 @@ static const VMStateDescription vmstate_smc91c111 = {
VMSTATE_BUFFER_UNSAFE(data, smc91c111_state, 0, NUM_PACKETS * 2048),
There really isn't anything tdk-specific about tdk_init() other than the
phy id registers. The function should instead be generalized for any
phy, at least as far as the ID registers are concerned. For the most
part the read/write behaviour should be very similar across PHYs.
This patch renames
Am 03.02.2013 00:06, schrieb Anthony Liguori:
Signed-off-by: Anthony Liguori aligu...@us.ibm.com
---
hw/s390x/s390-virtio-bus.c | 3 ++-
hw/s390x/virtio-ccw.c | 3 ++-
hw/virtio-pci.c| 3 ++-
hw/virtio.h| 3 ++-
4 files changed, 8 insertions(+), 4
The RST and ANEG_RST bits are commands, not settings. An operating
system will get confused (or at least u-boot does) if those bits remain
set after writing to them. Therefore, mask them out on write.
Similarly, no bits in the ID1, ID2, and remote capability registers are
writeable; so mask them
The smc91c111 device has bitbanged MDIO access, but the model doesn't
yet implement it. This patch uses the generalized bitbang MDIO support
pulled out of etraxfs Ethernet driver.
The MDIO state machine is driven by changes in state to the clock
control bit in the management register. The PHY
The MDIO state machine has a moderate amount of duplicate code in the
state processing that can be consolidated. This patch does so and
reorganizes it a bit so that far less code is required. Most of the
states simply stream a fixed number of bits in as a single integer and
can be handled by a
The etraxfs and Xilinx axienet Ethernet models implement quite a nice
MDIO core that supports both bitbanging and direct register access. This
change factors the common code out into a separate file. There are no
functional changes here, just movement of code.
The etraxfs and axienet are slightly
The MDIO model needs to have VMState support before it can be used by
devices that support VMState. This patch adds VMState macros for both
qemu_mdio and qemu_phy.
Cc: Peter Maydell peter.mayd...@linaro.org
Cc: Paul Brook p...@codesourcery.com
Cc: Edgar E. Iglesias edgar.igles...@gmail.com
Cc:
Il 31/01/2013 09:33, Orit Wasserman ha scritto:
If my above assumption that the guest reads unmapped memory as zeroes is
right, this mapping
is not necessary in the case of a zero dup page.
We just have to make sure that we are still in the very first round when
deciding not to sent
Il 02/02/2013 16:11, Eric Blake ha scritto:
On 02/02/2013 06:30 AM, Peter Maydell wrote:
- * Undefined if no bit exists, so code should check against 0 first.
+ * Returns -1 if no bit exists. Note that compared to the C library
+ * routine ffsl, this one returns one less.
*/
Do any of
Am 01.02.2013 18:03, schrieb Paolo Bonzini:
Il 31/01/2013 09:33, Orit Wasserman ha scritto:
If my above assumption that the guest reads unmapped memory as zeroes is
right, this mapping
is not necessary in the case of a zero dup page.
We just have to make sure that we are still in the very
74 matches
Mail list logo