Public bug reported:
Looking at the OpenBSD buildbot logs I noticed a warning that appears to be a
bug in the code.
OpenBSD has a 32-bit time_t on all archs at the moment (32-bit and 64-bit).
CCi386-softmmu/monitor.o
/buildbot-qemu/default_openbsd_current/build/monitor.c: In function
'exp
QEMU has no separate amd64 target; it is all handled by target-i386.
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/902413
Title:
qemu-i386-user on ARM host: wine hangs/spins when trying to run
an
Understood, thanks a lot for the pointers. From a quick skim it doesn't
look like I'm directly running into any of these ARM host issues (yet).
I'm hopeful that the i386 target will get increasing attention in the
future as ARM devices get more widespread after x86 was the standard for
so long. Out
Using {GET,MAKE}_TCGV_*() on parameters requires properly passing
arguments into the macro, so use strongly-typed functions while at it.
Signed-off-by: Andreas Färber
---
v1 -> v2:
* Use target_long for immediate arguments, not tcg_target_long
tcg/tcg-op.h | 1021
Am 10.12.2011 17:35, schrieb Peter Maydell:
> Make tcg_const_ptr() include a cast so that you can pass it a
> pointer. This allows us to drop the casts we had in all the places
> that use this macro.
>
> Signed-off-by: Peter Maydell
Acked-by: Andreas Färber
At first the use of tcg_target_long
On Fri, Dec 9, 2011 at 16:17, Paul Brook wrote:
>> A group of us are starting to work on sandboxing QEMU device emulation
>> code. We're just getting started investigating various approaches, and
>> want to engage the community to gather input.
>>
>> Following are the design points that we are cu
On Sat, Dec 10, 2011 at 16:29, Jan Kiszka wrote:
> On 2011-12-10 17:26, Blue Swirl wrote:
>> On Sat, Dec 10, 2011 at 16:03, Jan Kiszka wrote:
>>> On 2011-12-10 16:54, Blue Swirl wrote:
On Sat, Dec 10, 2011 at 15:51, Jan Kiszka wrote:
> On 2011-12-10 16:49, Blue Swirl wrote:
>>>
Remove target dependencies and compile Cirrus VGA in hwlib.
Signed-off-by: Blue Swirl
---
Makefile.objs |1 +
Makefile.target |1 -
hw/cirrus_vga.c |3 ---
3 files changed, 1 insertions(+), 4 deletions(-)
diff --git a/Makefile.objs b/Makefile.objs
index 3a699ee..932c68a 100644
---
Signed-off-by: Blue Swirl
---
hw/sga.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/sga.c b/hw/sga.c
index 7ef750a..ea11937 100644
--- a/hw/sga.c
+++ b/hw/sga.c
@@ -35,7 +35,7 @@ typedef struct ISAGAState {
ISADevice dev;
} ISASGAState;
-static int isa_cir
Instead of each target knowing or guessing the guest page size,
iterate through the dirty ranges.
Signed-off-by: Blue Swirl
---
cpu-all.h | 30 ++
hw/tcx.c | 54 ++
hw/vga.c | 16 +++-
memory.c |
Signed-off-by: Blue Swirl
---
Makefile.target | 13 +++---
default-configs/alpha-softmmu.mak|2 +
default-configs/arm-softmmu.mak |1 +
default-configs/i386-softmmu.mak |2 +
default-configs/mips-softmmu.mak |2 +
default-configs/mips64-
Instead of each target knowing or guessing the guest page size,
just pass the desired size of dirtied memory area. This should also
improve performance due to memset() optimizations.
Signed-off-by: Blue Swirl
---
arch_init.c |3 ++-
cpu-all.h |9 +++--
hw/cirrus_vga.c | 1
Improve VGA selection logic, push check for device availabilty to vl.c.
Create the devices at board level unconditionally.
Remove now unused pci_try_create*() functions.
Make PCI VGA devices optional.
Signed-off-by: Blue Swirl
---
hw/alpha_pci.c| 11 +++--
hw/boards.h |1 -
I'm not happy with the iterative dirty range finding API (fully
implemented on TCX here). It is not very useful from the device point
of view since memory address to coordinate conversion may be
nontrivial. Any suggestions?
Blue Swirl (6):
vga: improve VGA logic
memory: change dirty setting AP
Make tcg_const_ptr() include a cast so that you can pass it a
pointer. This allows us to drop the casts we had in all the places
that use this macro.
Signed-off-by: Peter Maydell
---
Minor cleanup; I have a patchset cooking that wants to use
tcg_const_ptr(), which is why I noticed this infelicity
On 2011-12-10 17:26, Blue Swirl wrote:
> On Sat, Dec 10, 2011 at 16:03, Jan Kiszka wrote:
>> On 2011-12-10 16:54, Blue Swirl wrote:
>>> On Sat, Dec 10, 2011 at 15:51, Jan Kiszka wrote:
On 2011-12-10 16:49, Blue Swirl wrote:
>>
>> +ISADevice *pit_init(int base, qemu_irq irq)
>
>>>
On Sat, Dec 10, 2011 at 16:03, Jan Kiszka wrote:
> On 2011-12-10 16:54, Blue Swirl wrote:
>> On Sat, Dec 10, 2011 at 15:51, Jan Kiszka wrote:
>>> On 2011-12-10 16:49, Blue Swirl wrote:
>
> +ISADevice *pit_init(int base, qemu_irq irq)
Please retain this function in pc.h, or even
On Sat, Dec 10, 2011 at 15:58, Jan Kiszka wrote:
> On 2011-12-10 16:40, Blue Swirl wrote:
>> On Fri, Dec 9, 2011 at 07:52, Jan Kiszka wrote:
>>> On 2011-12-09 08:45, Jan Kiszka wrote:
On 2011-12-08 22:16, Blue Swirl wrote:
> On Thu, Dec 8, 2011 at 11:52, Jan Kiszka wrote:
>> This in
On 2011-12-10 16:54, Blue Swirl wrote:
> On Sat, Dec 10, 2011 at 15:51, Jan Kiszka wrote:
>> On 2011-12-10 16:49, Blue Swirl wrote:
+ISADevice *pit_init(int base, qemu_irq irq)
>>>
>>> Please retain this function in pc.h, or even better, introduce i8254.h.
>>
>> No concerns about i8254.h
On 2011-12-10 16:40, Blue Swirl wrote:
> On Fri, Dec 9, 2011 at 07:52, Jan Kiszka wrote:
>> On 2011-12-09 08:45, Jan Kiszka wrote:
>>> On 2011-12-08 22:16, Blue Swirl wrote:
On Thu, Dec 8, 2011 at 11:52, Jan Kiszka wrote:
> This introduces the alternative APIC backend which makes use of
On Sat, Dec 10, 2011 at 15:51, Jan Kiszka wrote:
> On 2011-12-10 16:49, Blue Swirl wrote:
>>>
>>> +ISADevice *pit_init(int base, qemu_irq irq)
>>
>> Please retain this function in pc.h, or even better, introduce i8254.h.
>
> No concerns about i8254.h, but this function does not qualify for static
On Sat, Dec 10, 2011 at 12:28, Jan Kiszka wrote:
> This is a small preparatory series to allow the introduction of the KVM
> in-kernel PIT. Of course, it is also a fix for the various bugs in the
> related PIT/HPET code. See patches for details.
>
> Jan Kiszka (2):
> hpet: Save/restore cached RTC
On 2011-12-10 16:49, Blue Swirl wrote:
>>
>> +ISADevice *pit_init(int base, qemu_irq irq)
>
> Please retain this function in pc.h, or even better, introduce i8254.h.
No concerns about i8254.h, but this function does not qualify for static
inline.
>
>> +{
>> +ISADevice *dev;
>> +
>> +dev
On Sat, Dec 10, 2011 at 12:28, Jan Kiszka wrote:
> From: Jan Kiszka
>
> When the HPET enters legacy mode, the IRQ output of the PIT is
> suppressed and replaced by the HPET timer 0. But the current code to
> emulate this was broken in many ways. It reset the PIT state after
> re-enabling, it work
On Fri, Dec 9, 2011 at 07:52, Jan Kiszka wrote:
> On 2011-12-09 08:45, Jan Kiszka wrote:
>> On 2011-12-08 22:16, Blue Swirl wrote:
>>> On Thu, Dec 8, 2011 at 11:52, Jan Kiszka wrote:
This introduces the alternative APIC backend which makes use of KVM's
in-kernel device model. External N
malc wrote:
> > Inefficient in what way? The generated code only grows by a single
> > unconditional branch.
>
> The generated code grows by a memory write
Yes, an additional store instruction is generated, but the number of
store instructions *executed* does not change. The original code
alrea
For ARM hosts (mostly being worked on):
https://bugs.launchpad.net/qemu/+bug/893208
https://bugs.launchpad.net/qemu/+bug/883136
https://bugs.launchpad.net/qemu/+bug/883133
https://bugs.launchpad.net/qemu/+bug/870990
For x86 multithreaded (mostly *not* being worked on):
https://bugs.launchpad.net/q
** Also affects: qemu
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/870990
Title:
compile failure on ARMv7 hosts when compiled for thumb if --enable-
d
** Also affects: qemu
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/893208
Title:
qemu on ARM hosts can't boot i386 image
Status in QEMU:
New
Status i
Use new helper function to avoid new temporaries and moves.
Signed-off-by: Andreas Färber
---
target-i386/translate.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/target-i386/translate.c b/target-i386/translate.c
index c929ef9..17465e9 100644
--- a/target-i386/tran
gen_op_mov_reg_v() wants a TCGv, so use cpu_T[0] like pextrd does.
Signed-off-by: Andreas Färber
---
target-i386/translate.c | 17 +
1 files changed, 9 insertions(+), 8 deletions(-)
diff --git a/target-i386/translate.c b/target-i386/translate.c
index 6d09496..5ce8d3e 100644
--
Encapsule the conversion from/to i64.
This avoids a temporary and a tcg_gen_trunc_i64_tl()/ext_tl_i64().
Signed-off-by: Andreas Färber
---
tcg/tcg-op.h | 12
1 files changed, 12 insertions(+), 0 deletions(-)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index d065e74..07c6d3e 100644
-
Use TCGv (tl) versions instead.
Signed-off-by: Andreas Färber
---
target-i386/translate.c | 130 +++---
1 files changed, 65 insertions(+), 65 deletions(-)
diff --git a/target-i386/translate.c b/target-i386/translate.c
index 1ef8d16..0ffc450 100644
--- a/
Thanks for your quick reply, Peter.
Are there more specific bug entries tracking both the general problems
you're talking about that I could monitor for progress, or any pointers
on the direction to go to improve the situation?
--
You received this bug notification because you are a member of qe
Insert call to tcg_gen_ext_tl_i64().
Signed-off-by: Andreas Färber
---
target-i386/translate.c |6 +-
1 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/target-i386/translate.c b/target-i386/translate.c
index 0ffc450..c929ef9 100644
--- a/target-i386/translate.c
+++ b/target-
gen_op_mov_v_reg() wants a TCGv, so use cpu_tmp0 like pinsrd does.
Signed-off-by: Andreas Färber
---
target-i386/translate.c | 17 +
1 files changed, 9 insertions(+), 8 deletions(-)
diff --git a/target-i386/translate.c b/target-i386/translate.c
index 17465e9..6d09496 100644
--
Hello,
This series makes target-i386 compile with DEBUG_TCGV_TL.
Patches 1-2 could be cherry-picked already.
Patch 1 is fairly trivial.
Patch 2 is the only patch in the series that needs an additional temporary.
Patch 3 introduces two new functions to avoid that elsewhere. Needs TCGv series.
Wi
On Sat, 10 Dec 2011, Andreas Gustafsson wrote:
> malc wrote:
> > On Sat, 10 Dec 2011, Andreas Gustafsson wrote:
> >
> > > When the i386 cmpxchg instruction is executed with a memory operand
> > > and the comparison result is "unequal", do the memory write before
> > > changing the accumulator ins
From: Jan Kiszka
When the HPET enters legacy mode, the IRQ output of the PIT is
suppressed and replaced by the HPET timer 0. But the current code to
emulate this was broken in many ways. It reset the PIT state after
re-enabling, it worked against a stale static PIT structure, and it did
not prope
This is a small preparatory series to allow the introduction of the KVM
in-kernel PIT. Of course, it is also a fix for the various bugs in the
related PIT/HPET code. See patches for details.
Jan Kiszka (2):
hpet: Save/restore cached RTC IRQ level
i8254: Rework & fix interaction with HPET in le
From: Jan Kiszka
In legacy mode, the HPET suppresses the RTC interrupt delivery via IRQ
8 but keeps track of the RTC output level and applies it when legacy
mode is turned off again. This value has to be preserved across save/
restore as it cannot be reconstructed otherwise.
Signed-off-by: Jan K
malc wrote:
> On Sat, 10 Dec 2011, Andreas Gustafsson wrote:
>
> > When the i386 cmpxchg instruction is executed with a memory operand
> > and the comparison result is "unequal", do the memory write before
> > changing the accumulator instead of the other way around, because
> > otherwise the new
Am 10.12.2011 11:07, schrieb Peter Maydell:
> On 10 December 2011 09:02, Andreas Färber wrote:
>> Working on 16- and 20-bit targets using TARGET_LONG_BITS == 32, I found it
>> too
>> easy to accidentally use, e.g., tcg_temp_free() in place of
>> tcg_temp_free_i32().
>> In case of 78k0/rl78 it ma
On 2011-12-09 22:54, Anthony PERARD wrote:
> During the initialisation of the machine at restore time, the access to the
> VRAM will fail because QEMU does not know yet the right guest address to map,
> so the vram_ptr is NULL.
>
> So this patch avoid using a NULL pointer during initialisation, an
On Sat, 10 Dec 2011, Andreas Gustafsson wrote:
> When the i386 cmpxchg instruction is executed with a memory operand
> and the comparison result is "unequal", do the memory write before
> changing the accumulator instead of the other way around, because
> otherwise the new accumulator value will i
Multithreaded programs don't work (reliably) in x86 user emulation mode. This
is a known (longstanding) bug.
ARM hosts are also currently known to have problems (as stated in the qemu 1.0
release notes).
--
You received this bug notification because you are a member of qemu-
devel-ml, which is
On 10 December 2011 09:02, Andreas Färber wrote:
> Working on 16- and 20-bit targets using TARGET_LONG_BITS == 32, I found it too
> easy to accidentally use, e.g., tcg_temp_free() in place of
> tcg_temp_free_i32().
> In case of 78k0/rl78 it may not practically matter yet, but it complicates
> goi
When the i386 cmpxchg instruction is executed with a memory operand
and the comparison result is "unequal", do the memory write before
changing the accumulator instead of the other way around, because
otherwise the new accumulator value will incorrectly be used in the
comparison when the instructio
Am 10.12.2011 01:49, schrieb Andreas Färber:
Am 10.12.2011 00:19, schrieb Stefan Weil:
This file can be used with codespell to suppress wrong
misspelling reports.
Signed-off-by: Stefan Weil
---
scripts/codespell.exclude | 215
+
1 files changed, 215
Switch them to use the new {MAKE,GET}_TCGV_TL() macros.
Signed-off-by: Andreas Färber
---
tcg/tcg-op.h |7 +++
1 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index e5c2e37..a6c3d5f 100644
--- a/tcg/tcg-op.h
+++ b/tcg/tcg-op.h
@@ -2129,16 +2129,
Using {GET,MAKE}_TCGV_*() on parameters requires properly passing
arguments into the macro, so use strongly-typed functions while at it.
Signed-off-by: Andreas Färber
---
tcg/tcg-op.h | 1021 +-
1 files changed, 866 insertions(+), 155 delet
Hello,
Working on 16- and 20-bit targets using TARGET_LONG_BITS == 32, I found it too
easy to accidentally use, e.g., tcg_temp_free() in place of tcg_temp_free_i32().
In case of 78k0/rl78 it may not practically matter yet, but it complicates
going from a 32-bit target to 64 bits, as in the case of
It's easy to omit _i32 somewhere if working on one 32-bit target,
despite DEBUG_TCGV, because TCGv is simply aliased to TCGv_i32/i64.
If DEBUG_TCGV_TL is defined, use a new struct TCGv with distinguished
accessors to catch mixups.
This cannot be done unconditionally for DEBUG_TCGV because some ta
Signed-off-by: Andreas Färber
---
tcg/tcg-op.h |4
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index 82e04e7..cbccaef 100644
--- a/tcg/tcg-op.h
+++ b/tcg/tcg-op.h
@@ -2127,6 +2127,8 @@ static inline void tcg_gen_deposit_i64(TCGv_i64 ret,
TC
54 matches
Mail list logo
