On 09/21/2012 09:43 AM, Aurelien Jarno wrote:
This patch series first does a bit of cleanup and bug fixing on the
MIPS TCG backend. Then some optimizations are added, mostly by
implementing new TCG operations. The last patch depends on the movcond
series by Richard Henderson.
This has been
On Fri, Sep 21, 2012 at 04:05:14PM +0200, Markus Armbruster wrote:
Jason Baron jba...@redhat.com writes:
From: Isaku Yamahata yamah...@valinux.co.jp
Introduce a helper function which initializes the ahci port with ide
devices.
It will be used by q35 support.
Cc: Alexander Graf
Now that it's possible to detect copies, we can optimize the case
the op r, a, a = movi r, 0. This helps in the computation of
overflow flags when one of the two args is 0.
Reviewed-by: Richard Henderson r...@twiddle.net
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
tcg/optimize.c |
The copy propagation pass tries to keep track what is a copy of what
and what has copy of what, and in addition it keep a circular list of
of all the copies. Unfortunately this doesn't fully work: a mov from
a temp which has a state COPY changed it into a state HAS_COPY.
Later when this temp is
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
tcg/optimize.c | 20
1 file changed, 20 insertions(+)
diff --git a/tcg/optimize.c b/tcg/optimize.c
index c8ae50b..35532a1 100644
--- a/tcg/optimize.c
+++ b/tcg/optimize.c
@@ -668,6 +668,26 @@ static TCGArg
When both argument of brcond/movcond/setcond are the same or when one
of the two values is a constant equal to zero, it's possible to do
further optimizations.
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
tcg/optimize.c | 127 +---
1
It is possible to due copy propagation for all operations, even the one
that have side effects or clobber arguments (it only concerns input
arguments). That said, the call operation should be handled differently
due to the variable number of arguments.
Reviewed-by: Richard Henderson
Now that we can easily detect all copies, we can optimize the
op r, a, a = mov r, a case a bit more.
Reviewed-by: Richard Henderson r...@twiddle.net
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
tcg/optimize.c |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
This patch series rework the copy propagation in order to generate
better code. The first two patches are cleanup and bug fixes, the third
Ping. There don't seem to be any objections to this. Thanks,
Alex
On Fri, 2012-09-14 at 17:04 -0600, Alex Williamson wrote:
On Fri, 2012-09-14 at 17:01 -0600, Alex Williamson wrote:
Same goodness as v4, plus:
- Addressed comments by Blue Swirl (thanks for the review)
(hopefully
Use the deposit op instead of and hardcoded bit field insertion. It
allows the host to emit the corresponding instruction if available.
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
target-arm/translate.c | 20 ++--
1 file changed, 6 insertions(+), 14 deletions(-)
diff --git a/tcg/ppc/tcg-target.c b/tcg/ppc/tcg-target.c
index 26c4b33..0fb6fc7 100644
--- a/tcg/ppc/tcg-target.c
+++ b/tcg/ppc/tcg-target.c
@@ -390,6 +390,7 @@ static int tcg_target_const_match(tcg_target_long val,
#define ORCXO31(412)
#define EQVXO31(284)
#define NAND XO31(476)
The copy propagation doesn't check the types of the temps during copy
propagation. However TCG is using the mov_i32 for the i64 to i32
conversion and thus the two are not equivalent.
With this patch tcg_opt_gen_mov() doesn't consider two temps of
different type as copies anymore.
So far it seems
This patch series optimizes the ARM target by:
- using globals instead of ld/st function
- using TCG code instead of helpers
- marking some helpers const and pure
--
Changes v1 - v2
- updated patch 3 to use movcond instead of setcond. Also converted sar.
- updated patch 4 following changes
Am 17.09.2012 17:50, schrieb Andreas Färber:
Am 17.09.2012 17:28, schrieb Richard Henderson:
Commit 6375e09e changed the type of TranslationBlock.tb_next,
but failed to change the type of TCGContext.tb_next.
Signed-off-by: Richard Hendersonr...@twiddle.net
Reviewed-by: Andreas
On 09/21/12 10:18, Eduardo Habkost wrote:
On Thu, Sep 20, 2012 at 04:06:27PM -0400, Don Slutz wrote:
From http://lkml.indiana.edu/hypermail/linux/kernel/1205.0/00100.html
EAX should be KVM_CPUID_FEATURES (0x4001) not 0.
Added hypervisor-vendor=kvm0 to get the older CPUID result. kvm1
Signed-off-by: Victor Toso vic...@itoso.org
---
hw/virtio-balloon.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/hw/virtio-balloon.c b/hw/virtio-balloon.c
index dd1a650..9e0cd88 100644
--- a/hw/virtio-balloon.c
+++ b/hw/virtio-balloon.c
@@ -193,12 +193,15 @@ static
Cc: Peter Maydell peter.mayd...@linaro.org
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
target-arm/helper.h | 19 ++-
1 file changed, 10 insertions(+), 9 deletions(-)
diff --git a/target-arm/helper.h b/target-arm/helper.h
index 794e2b1..8b9adf1 100644
---
TCG_TEMP_ANY has no different meaning than TCG_TEMP_UNDEF, so use
the later instead.
Reviewed-by: Richard Henderson r...@twiddle.net
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
tcg/optimize.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git
Now that the movcond TCG op is available, it's possible to replace
shl and shr helpers by TCG code. The code generated by TCG is slightly
longer than the code generated by GCC for the helper but is still worth
it as this avoid all the consequences of using an helper: globals saved
back to memory,
Commit 25c4d9cc changed all TCGOpcode enums to be available, so we don't
need to #ifdef #endif the one that are available only on some targets.
This makes the code easier to read.
Reviewed-by: Richard Henderson r...@twiddle.net
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
tcg/tcg.c |
Use globals for CC flags instead of loading/storing them each they are
accessed. This allows some optimizations to be performed by the TCG
optimization passes.
Cc: Peter Maydell peter.mayd...@linaro.org
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
target-arm/translate.c | 127
Now that the setcond TCG op is available, it's possible to replace
add_cc and sub_cc helpers by TCG code. The code generated by TCG is
actually very close to the one generated by GCC for the helper, and
this avoid all the consequences of using an helper: globals saved back
to memory, no possible
The op a, a, b form is better handled on non-RISC host than the op
a, b, a form, so swap the arguments to this form when possible, and
when b is not a constant.
This reduces the number of generated instructions by a tiny bit.
Signed-off-by: Aurelien Jarno aurel...@aurel32.net
---
tcg/optimize.c
On Fri, Sep 21, 2012 at 04:26:58PM -0400, Don Slutz wrote:
On 09/21/12 10:18, Eduardo Habkost wrote:
On Thu, Sep 20, 2012 at 04:06:27PM -0400, Don Slutz wrote:
From http://lkml.indiana.edu/hypermail/linux/kernel/1205.0/00100.html
EAX should be KVM_CPUID_FEATURES (0x4001) not 0.
Added
On 09/21/2012 09:34 AM, Paolo Bonzini wrote:
Il 21/09/2012 16:08, Juan Quintela ha scritto:
We call buffered_put_buffer with iothread held, and buffered_flush() does
synchronous writes. We only want to do the synchronous writes outside.
This means that the buffer can grow to up to
On Fri, Sep 21, 2012 at 11:18:39AM -0700, Richard Henderson wrote:
On 09/21/2012 09:43 AM, Aurelien Jarno wrote:
Don't use the global pointer in TCG, in case helpers try access global
variables.
Err.. isn't the GP computed at the entry point of functions that need to use
it?
I read that
On Fri, Sep 21, 2012 at 10:13:39AM -0700, Richard Henderson wrote:
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 r...@twiddle.net
---
tcg/tcg-op.h
On Fri, Sep 21, 2012 at 10:13:40AM -0700, Richard Henderson wrote:
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 r...@twiddle.net
---
tcg/tcg-op.h | 22
On 09/21/12 16:49, Eduardo Habkost wrote:
On Fri, Sep 21, 2012 at 04:26:58PM -0400, Don Slutz wrote:
On 09/21/12 10:18, Eduardo Habkost wrote:
On Thu, Sep 20, 2012 at 04:06:27PM -0400, Don Slutz wrote:
From http://lkml.indiana.edu/hypermail/linux/kernel/1205.0/00100.html
EAX should be
On Fri, Sep 21, 2012 at 11:46:15AM +0100, Mel Gorman wrote:
This reverts
mm-compaction-check-lock-contention-first-before-taking-lock.patch as it
is replaced by a later patch in the series.
Signed-off-by: Mel Gorman mgor...@suse.de
Acked-by: Rafael Aquini aqu...@redhat.com
On Fri, Sep 21, 2012 at 11:46:16AM +0100, Mel Gorman wrote:
This reverts
mm-compaction-abort-compaction-loop-if-lock-is-contended-or-run-too-long-fix
as it is replaced by a later patch in the series.
Signed-off-by: Mel Gorman mgor...@suse.de
Acked-by: Rafael Aquini aqu...@redhat.com
On Fri, Sep 21, 2012 at 11:46:19AM +0100, Mel Gorman wrote:
Compactions migrate scanner acquires the zone-lru_lock when scanning a range
of pages looking for LRU pages to acquire. It does this even if there are
no LRU pages in the range. If multiple processes are compacting then this
can cause
On Fri, Sep 21, 2012 at 11:46:21AM +0100, Mel Gorman wrote:
This reverts commit 7db8889a (mm: have order 0 compaction start off
where it left) and commit de74f1cc (mm: have order 0 compaction start
near a pageblock with free pages). These patches were a good idea and
tests confirmed that
On Fri, 21 Sep 2012 11:46:18 +0100
Mel Gorman mgor...@suse.de wrote:
Changelog since V2
o Fix BUG_ON triggered due to pages left on cc.migratepages
o Make compact_zone_order() require non-NULL arg `contended'
Changelog since V1
o only abort the compaction if lock is contended or run too
On Fri, Sep 21, 2012 at 11:46:18AM +0100, Mel Gorman wrote:
From: Shaohua Li s...@fusionio.com
Changelog since V2
o Fix BUG_ON triggered due to pages left on cc.migratepages
o Make compact_zone_order() require non-NULL arg `contended'
Changelog since V1
o only abort the compaction if
On Fri, Sep 21, 2012 at 11:46:23AM +0100, Mel Gorman wrote:
This is almost entirely based on Rik's previous patches and discussions
with him about how this might be implemented.
Order 0 compaction stops when enough free pages of the correct page
order have been coalesced. When doing
On Fri, Sep 21, 2012 at 05:28:27PM -0400, Don Slutz wrote:
On 09/21/12 16:49, Eduardo Habkost wrote:
On Fri, Sep 21, 2012 at 04:26:58PM -0400, Don Slutz wrote:
On 09/21/12 10:18, Eduardo Habkost wrote:
On Thu, Sep 20, 2012 at 04:06:27PM -0400, Don Slutz wrote:
From
On 09/21/2012 05:17 AM, Vasilis Liaskovitis wrote:
Guest can respond to ACPI hotplug events e.g. with _EJ or _OST method.
This patch implements a tail queue to store guest notifications for memory
hot-add and hot-remove requests.
Guest responses for memory hotplug command on a per-dimm basis
On Fri, Sep 21, 2012 at 11:46:17AM +0100, Mel Gorman wrote:
This reverts
mm-compaction-abort-compaction-loop-if-lock-is-contended-or-run-too-long.patch
as it is replaced by a later patch in the series.
Signed-off-by: Mel Gorman mgor...@suse.de
Acked-by: Rafael Aquini aqu...@redhat.com
On Fri, Sep 21, 2012 at 11:46:22AM +0100, Mel Gorman wrote:
When compaction was implemented it was known that scanning could potentially
be excessive. The ideal was that a counter be maintained for each pageblock
but maintaining this information would incur a severe penalty due to a
shared
On Fri, Sep 21, 2012 at 11:46:20AM +0100, Mel Gorman wrote:
Compactions free scanner acquires the zone-lock when checking for PageBuddy
pages and isolating them. It does this even if there are no PageBuddy pages
in the range.
This patch defers acquiring the zone lock for as long as possible.
On 09/21/2012 01:10 PM, malc wrote:
+if (dest == v2) {
+label_ptr = s-code_ptr;
+tcg_out32 (s, tcg_to_bc[tcg_invert_cond (cond)]);
+tcg_out_mov (s, TCG_TYPE_I32, dest, v1);
+reloc_pc14 (label_ptr, (tcg_target_long) s-code_ptr);
+
On 09/21/2012 02:23 PM, Aurelien Jarno wrote:
Now I wonder if it wouldn't be better to write brcond2 as setcond2 +
brcond. And even setcond2 as a pair of setcond in TCG, which would allow
some optimizations in case both high parts are zero.
I think brcond2 vs setcond2 is a choice that has to
On Fri, 21 Sep 2012, Richard Henderson wrote:
On 09/21/2012 01:10 PM, malc wrote:
+if (dest == v2) {
+label_ptr = s-code_ptr;
+tcg_out32 (s, tcg_to_bc[tcg_invert_cond (cond)]);
+tcg_out_mov (s, TCG_TYPE_I32, dest, v1);
+reloc_pc14
On 09/21/2012 05:17 AM, Vasilis Liaskovitis wrote:
Returns total physical memory available to guest in bytes, including
hotplugged
memory. Note that the number reported here may be different from what the
guest
sees e.g. if the guest has not logically onlined hotplugged memory.
This
On 09/21/2012 12:33 PM, Aurelien Jarno wrote:
+static void gen_sar(TCGv dest, TCGv t0, TCGv t1)
+{
+TCGv tmp1, tmp2, tmp3;
+tmp1 = tcg_temp_new_i32();
+tcg_gen_andi_i32(tmp1, t1, 0xff);
+tmp2 = tcg_const_i32(0x1f);
+tmp3 = tcg_const_i32(0);
+
On 09/21/2012 08:07 AM, Michael Roth wrote:
Signed-off-by: Michael Roth mdr...@linux.vnet.ibm.com
---
docs/qidl.txt | 347
+
1 file changed, 347 insertions(+)
create mode 100644 docs/qidl.txt
+Specifying What/How State Gets Saved
On 09/21/2012 12:43 PM, Aurelien Jarno wrote:
+/* Return 2 if the condition can't be simplified, and the result
+ of the condition (0 or 1) if it can */
Ok as-is. But I'd bike-shed the fail return to -1.
r~
On 09/21/2012 08:07 AM, Michael Roth wrote:
Adds an abstract Lexer class to handle tokenizer via a
peek/pop/peekline/popline interface, along with an implementation for C
based on the lexer from qc.git
Signed-off-by: Michael Roth mdr...@linux.vnet.ibm.com
---
scripts/lexer.py | 306
On 09/21/2012 08:07 AM, Michael Roth wrote:
This introduces the QIDL parser to process QIDL annotations in C files.
This code is mostly a straight import from qc.git, with some reworking
to handle the declaration/annotation format and lexer we're using for
QEMU.
Signed-off-by: Michael Roth
On 09/21/2012 12:43 PM, Aurelien Jarno wrote:
+tmp = ((1ull args[4]) - 1);
+tmp = (temps[args[1]].val ~(tmp args[3]))
+ | ((temps[args[2]].val tmp) args[3]);
+tcg_opt_gen_movi(gen_args, args[0], tmp);
We do have a
On 09/21/2012 12:43 PM, Aurelien Jarno wrote:
Changes v1 - v2
- fixed conflict in patch 3 following movcond introduction
- movcond optimization added in patch 7
- added patch 10
Aurelien Jarno (10):
tcg/optimize: remove TCG_TEMP_ANY
tcg/optimize: check types in copy propagation
On Fri, Sep 21, 2012 at 05:18:09PM -0600, Eric Blake wrote:
On 09/21/2012 08:07 AM, Michael Roth wrote:
Adds an abstract Lexer class to handle tokenizer via a
peek/pop/peekline/popline interface, along with an implementation for C
based on the lexer from qc.git
Signed-off-by: Michael
Also known as Paravirtualization CPUIDs.
This is primarily done so that the guest will think it is running
under vmware when hypervisor-vendor=vmware is specified as a
property of a cpu.
This depends on:
http://lists.gnu.org/archive/html/qemu-devel/2012-09/msg01400.html
As far as I know it is
Fix duplicate name (kvmclock = kvm_clock2) also.
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c | 12
1 files changed, 8 insertions(+), 4 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index 0313cf5..25ca986 100644
--- a/target-i386/cpu.c
+++
The check using INT_MAX (2147483647) is wrong in this case.
Signed-off-by: Fred Oliveira folive...@cloudswitch.com
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
This was taken from:
http://article.gmane.org/gmane.comp.emulators.kvm.devel/22643
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c | 32
1 files changed, 32 insertions(+), 0 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/kvm.c | 19 +++
1 files changed, 19 insertions(+), 0 deletions(-)
diff --git a/target-i386/kvm.c b/target-i386/kvm.c
index f8a5177..ff82034 100644
--- a/target-i386/kvm.c
+++ b/target-i386/kvm.c
@@ -454,6 +454,25 @@
The subject of the andi and assertion patches has come up on this
list earlier this week, between Max Filippov, malc and myself.
The posibility of using deposit to implement concat occurred to
me while working on the MIPS FPU conversion patch.
r~
Richard Henderson (8):
tcg: Adjust
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
Checking that we don't try for idx != [01] is trivial. Checking
that we don't issue more than one of any index requires a tad
more data and some ifdefs protecting that new variable.
Signed-off-by: Richard Henderson r...@twiddle.net
Cc: Max Filippov jcmvb...@gmail.com
---
tcg/tcg-op.h | 11
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
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c | 66 +
1 files changed, 66 insertions(+), 0 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index 478b671..2423090 100644
--- a/target-i386/cpu.c
+++
These are modeled after x86_cpuid_set_vendor and x86_cpuid_get_vendor.
Since kvm's vendor is shorter, the test for correct size is removed and zero
padding is added.
Set Microsoft's Vendor now that we can. Value defined in:
Signed-off-by: Don Slutz d...@cloudswitch.com
---
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
These are modeled after x86_cpuid_get_xlevel and x86_cpuid_set_xlevel.
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c |8
target-i386/cpu.h |2 ++
2 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index
These are modeled after x86_cpuid_get_xlevel and x86_cpuid_set_xlevel.
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c | 29 +
1 files changed, 29 insertions(+), 0 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index
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
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
Also known as Paravirtualization vendor.
This is EBX, ECX, EDX data for 0x4000.
QEMU knows this is KVM_CPUID_SIGNATURE (0x4000).
This is based on:
Microsoft Hypervisor CPUID Leaves:
http://msdn.microsoft.com/en-us/library/windows/hardware/ff542428%28v=vs.85%29.aspx
Linux kernel
Also known as Paravirtualization level or maximim cpuid function present in
this leaf.
This is just the EAX value for 0x4000.
QEMU knows this is KVM_CPUID_SIGNATURE (0x4000).
This is based on:
Microsoft Hypervisor CPUID Leaves:
Like the C assert macro, except only enabled for CONFIG_DEBUG_TCG,
and without having to set _NDEBUG and disable all other asserts at
the same time.
The use of __builtin_unreachable (when available) gives the compiler
the same information, which may (or may not) help it optimize better.
Given these are constants, checking once here means everything
after can assume they're correct.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/tcg-op.h | 8
1 file changed, 8 insertions(+)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index d2fb283..ecb1ac3 100644
---
Note that xori_i64 failed to perform even the minimal
optimizations promised by the README.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/tcg-op.h | 19 +++
1 file changed, 15 insertions(+), 4 deletions(-)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c | 57 +++-
target-i386/cpu.h | 19 +
2 files changed, 74 insertions(+), 2 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index bfe7b27..478b671
Signed-off-by: Don Slutz d...@cloudswitch.com
---
target-i386/cpu.c | 11 +++
1 files changed, 11 insertions(+), 0 deletions(-)
diff --git a/target-i386/cpu.c b/target-i386/cpu.c
index e8a21b5..12bd5d4 100644
--- a/target-i386/cpu.c
+++ b/target-i386/cpu.c
@@ -1999,6 +1999,17 @@ void
Note that andi_i64 failed to perform even the minimal
optimizations promised by the README.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/tcg-op.h | 67 ++--
1 file changed, 56 insertions(+), 11 deletions(-)
diff --git
Copy the same optimizations from ori_i32.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/tcg-op.h | 19 +--
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git a/tcg/tcg-op.h b/tcg/tcg-op.h
index c8633ff..fd16499 100644
--- a/tcg/tcg-op.h
+++ b/tcg/tcg-op.h
@@
The README file documented the operand ordering of the tcg_gen_*
functions. Since we're documenting opcodes here, use the true
operand ordering.
Signed-off-by: Richard Henderson r...@twiddle.net
Cc: malc av1...@comtv.ru
---
tcg/README | 10 +-
1 file changed, 5 insertions(+), 5
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/hppa/tcg-target.c | 21 +
tcg/hppa/tcg-target.h | 2 +-
2 files changed, 22 insertions(+), 1 deletion(-)
diff --git a/tcg/hppa/tcg-target.c b/tcg/hppa/tcg-target.c
index 5385d45..793b90d 100644
---
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 that this relies on the ANDI - EXTU patch for the identity claim.
Signed-off-by: Richard Henderson r...@twiddle.net
---
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 13 +
1 file changed, 13 insertions(+)
diff --git a/tcg/sparc/tcg-target.c b/tcg/sparc/tcg-target.c
index d401f8e..03c385a 100644
--- a/tcg/sparc/tcg-target.c
+++ b/tcg/sparc/tcg-target.c
@@ -78,12 +78,25
The xtensa-test image generates a sra_i32 with count 0x40.
Whether this is accident of tcg constant propagation or
originating directly from the instruction stream is immaterial.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 18 --
1 file changed,
The address we pick in sparc64.ld is also 0x6000, so doing a fixed map
on top of that is guaranteed to blow up. Choosing 0x4000 is exactly
right for the max of code_gen_buffer_size set below.
No need to ever use MAP_FIXED. While getting our desired address helps
optimize the generated
Changes since v2:
* Patch 4 split out from patch 5 (afaerber feedback)
* TB chaining preserves wrt retranslation
* Last patch for branch retranslation is new
The patch set is rebased on cfb75cb9807463ebe18b127096b48b5d0db1ce03
and is available at
git://repo.or.cz/qemu/rth.git tcg-sparc
Signed-off-by: Richard Henderson r...@twiddle.net
---
exec-all.h | 9 ++---
tcg/sparc/tcg-target.c | 21 ++---
2 files changed, 24 insertions(+), 6 deletions(-)
diff --git a/exec-all.h b/exec-all.h
index dba9609..6516da0 100644
--- a/exec-all.h
+++ b/exec-all.h
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 19 +++
1 file changed, 11 insertions(+), 8 deletions(-)
diff --git a/tcg/sparc/tcg-target.c b/tcg/sparc/tcg-target.c
index 1db0c9d..876da4f 100644
--- a/tcg/sparc/tcg-target.c
+++
And change from %i4/%i5 to %g1/%o7 to remove a v8plus fixme.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 115 +
1 file changed, 59 insertions(+), 56 deletions(-)
diff --git a/tcg/sparc/tcg-target.c
Current code doesn't actually work in 32-bit mode at all. Since
no one really noticed, drop the complication of v7 and v8 cpus.
Eliminate the --sparc_cpu configure option and standardize macro
testing on TCG_TARGET_REG_BITS / HOST_LONG_BITS
Signed-off-by: Richard Henderson r...@twiddle.net
---
At the same time, split out the tlb load logic to a new function.
Fixes the cases of two data registers and two address registers.
Fixes the signature of, and adds missing, qemu_ld/st opcodes.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 777
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tcg/sparc/tcg-target.c b/tcg/sparc/tcg-target.c
index 608fc46..0a19313 100644
--- a/tcg/sparc/tcg-target.c
+++ b/tcg/sparc/tcg-target.c
@@ -236,7 +236,7
Signed-off-by: Richard Henderson r...@twiddle.net
---
configure | 2 ++
tcg/sparc/tcg-target.c | 26 +++---
tcg/sparc/tcg-target.h | 2 ++
3 files changed, 27 insertions(+), 3 deletions(-)
diff --git a/configure b/configure
index df888f2..0dc4170 100755
---
We can now move the TCG variable from %g[56] to a call-preserved
windowed register.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 3 ++-
tcg/sparc/tcg-target.h | 8 +---
2 files changed, 3 insertions(+), 8 deletions(-)
diff --git a/tcg/sparc/tcg-target.c
Don't use -ffixed-gN. Don't link statically. Don't save/restore
AREG0 around calls. Don't allocate space on the stack for AREG0 save.
Signed-off-by: Richard Henderson r...@twiddle.net
---
configure | 12 ---
tcg/sparc/tcg-target.c | 55
When host and target have differing alignment rules, using a cast
and direct memory operation can result in SIGBUS. Use memcpy instead,
which the compiler will happily optimize when alignment is satisfied.
Signed-off-by: Richard Henderson r...@twiddle.net
Reviewed-by: Peter Maydell
Not actually implemented, but at least we avoid the tcg assert at startup.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/sparc/tcg-target.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/tcg/sparc/tcg-target.c b/tcg/sparc/tcg-target.c
index baed3b4..608fc46 100644
---
On 09/21/12 17:53, Eduardo Habkost wrote:
On Fri, Sep 21, 2012 at 05:28:27PM -0400, Don Slutz wrote:
On 09/21/12 16:49, Eduardo Habkost wrote:
On Fri, Sep 21, 2012 at 04:26:58PM -0400, Don Slutz wrote:
On 09/21/12 10:18, Eduardo Habkost wrote:
On Thu, Sep 20, 2012 at 04:06:27PM -0400, Don
On Fri, 21 Sep 2012 11:46:22 +0100
Mel Gorman mgor...@suse.de wrote:
When compaction was implemented it was known that scanning could potentially
be excessive. The ideal was that a counter be maintained for each pageblock
but maintaining this information would incur a severe penalty due to a
On Fri, 21 Sep 2012 11:46:20 +0100
Mel Gorman mgor...@suse.de wrote:
Compactions free scanner acquires the zone-lock when checking for PageBuddy
pages and isolating them. It does this even if there are no PageBuddy pages
in the range.
This patch defers acquiring the zone lock for as long as
301 - 400 of 401 matches
Mail list logo