Aliasing was forcing s-code_ptr to be re-read after the store.
Keep the pointer in a local variable to help the compiler.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/tcg.c | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/tcg/tcg.c b/tcg/tcg.c
index
V2:
* Feedback from Aurelien
* Added the first patch.
* ARM patches fixed.
Sorry about not merging back from the tree on my arm machine
before posting the patch set last time.
r~
Richard Henderson (9):
tcg: Tidy generated code for tcg_outN
tcg-i386: Add and use tcg_out64
Allow the code that tcg generates to be less obtuse, passing in
the return address directly instead of computing it in the helper.
Maintain the old entrance point unchanged as an alternate entry point.
Delete the helper_st*_cmmu prototypes; the implementations did not exist.
Signed-off-by:
Use a 7 byte lea before the ultimate 10 byte movq.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/i386/tcg-target.c | 25 -
1 file changed, 20 insertions(+), 5 deletions(-)
diff --git a/tcg/i386/tcg-target.c b/tcg/i386/tcg-target.c
index 841bd75..8226171
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/arm/tcg-target.c | 48 +++-
1 file changed, 43 insertions(+), 5 deletions(-)
diff --git a/tcg/arm/tcg-target.c b/tcg/arm/tcg-target.c
index 6c4854d..6cfceff 100644
--- a/tcg/arm/tcg-target.c
+++
Avoid a loop in the tlb_fill path; the fill will either succeed or
generate an exception.
Inline the slow_ld/st function; it was a complete copy of the main
helper except for the actual cross-page unaligned code, and the
compiler was inlining it anyway.
Add unlikely markers optimizing for the
No point in splitting the write into 32-bit pieces.
Reviewed-by: Aurelien Jarno aurel...@aurel32.net
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/i386/tcg-target.c | 3 +--
tcg/tcg.c | 7 +++
2 files changed, 8 insertions(+), 2 deletions(-)
diff --git
Use existing stack space for arguments; don't push/pop.
Use less ifdefs and more C ifs.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/i386/tcg-target.c | 165 ++
1 file changed, 74 insertions(+), 91 deletions(-)
diff --git
Discontinue the jump-around-jump-to-jump scheme, trading it for a single
immediate move instruction. The two extra jumps always consume 7 bytes,
whereas the immediate move is either 5 or 7 bytes depending on where the
code_gen_buffer gets located.
Signed-off-by: Richard Henderson
Instead of using a branch-call-branch sequence, arrange for a
call-branch sequence, using the ARM's conditional call insn.
This reduces the size of the slow-path within the TB, and makes
the GETPC_EXT implementation identical for TCG and not-TCG.
Signed-off-by: Richard Henderson r...@twiddle.net
No changes since the patch set was first posted, but rebased against head.
r~
The following changes since commit f202039811d8746b0586d2fd5f61de6c8cf68056:
Open up 1.7 development branch (2013-08-15 15:41:13 -0500)
are available in the git repository at:
The destination of the call_pal, and the cpu state, is very predictable;
there's no need for exiting the cpu loop.
Signed-off-by: Richard Henderson r...@twiddle.net
---
target-alpha/helper.h | 1 +
target-alpha/sys_helper.c | 12
target-alpha/translate.c | 25
With appropriate flushing when the PALBR changes, the target of
a CALL_PAL is so predictable we can chain to it.
Signed-off-by: Richard Henderson r...@twiddle.net
---
target-alpha/helper.h | 1 +
target-alpha/sys_helper.c | 5 +
target-alpha/translate.c | 20
3
This allows significantly more threading, and occasionally larger TBs,
when processing code for the kernel and PALcode.
Signed-off-by: Richard Henderson r...@twiddle.net
---
target-alpha/translate.c | 37 +
1 file changed, 29 insertions(+), 8 deletions(-)
Signed-off-by: Richard Henderson r...@twiddle.net
---
hw/alpha/typhoon.c | 202 +
1 file changed, 174 insertions(+), 28 deletions(-)
diff --git a/hw/alpha/typhoon.c b/hw/alpha/typhoon.c
index b7fb044..2450045 100644
--- a/hw/alpha/typhoon.c
+++
Il 16/08/2013 11:45, Alexey Kardashevskiy ha scritto:
Did this patch go anywhere? I could not find it in
git://github.com/bonzini/qemu.git scsi-next or anywhere else.
Should I repost it as qemu 1.6 was released and we started new cycle?
I am asking because I have 2 more patches about
Il 16/08/2013 14:41, ashish yadav ha scritto:
Hi All,
I am using CentOS 6.3 .
I want to use qemu-1.4.0 with it ( due to dpdk_ovs requirement) .
If qemu-1.4.0 is supported with Centos 6.3 ( kernel version 2.6.32-279) ?
It works, you just have to compile it yourself.
Paolo
Ping.
r~
On 08/05/2013 11:28 AM, Richard Henderson wrote:
About half of these patches are focused on reducing the number of
full 64-bit constants that need to be generated for addresses:
E.g. patch 5, looking through the function descriptor. If the
program is built --disable-pie, the
Ping.
r~
On 07/24/2013 12:50 PM, Richard Henderson wrote:
It has been pointed out on LKML that the alpha umount syscall numbers
are named wrong, and a patch to rectify that has been posted for 3.11.
Glibc works around this by treating NR_umount as NR_umount2 if
NR_oldumount exists.
Il 02/08/2013 09:05, Peter Lieven ha scritto:
can you give an update what are to current plans/schedule to merge this
series? I have
a few patches in the queue that in their current version depend on this
series being merged.
It should go in soon, perhaps a couple of weeks.
Paolo
V2 adds the --adjust-vma idea from Max Filippov.
r~
Richard Henderson (3):
disas: Implement fallback to dump object code as hex
disas: Add disas-objdump.pl
disas-objdump: Pass --adjust-vma to objdump
disas.c | 47 ++-
scripts/disas-objdump.pl | 99
The OBJD-[HT] tags will be used by a script to run the hex blob
through objdump --disassemble.
Signed-off-by: Richard Henderson r...@twiddle.net
---
disas.c | 47 +--
1 file changed, 37 insertions(+), 10 deletions(-)
diff --git a/disas.c b/disas.c
The script massages the output produced for architectures that are
not supported internally by qemu though an external objdump program
for disassembly.
Signed-off-by: Richard Henderson r...@twiddle.net
---
scripts/disas-objdump.pl | 87
1 file
Il 16/08/2013 16:41, Mike Day ha scritto:
This patch set merges Paolo's conversion of address spaces to enable
RCU. There is one more patchset coming for TLB access that I'm
debugging right now. After I submit the last one I'll start working on
enabling RCU more widely - and the series will
This gives the dumped blob its correct address during disassembly,
which makes pc-relative insns much easier to interpret.
Signed-off-by: Richard Henderson r...@twiddle.net
---
scripts/disas-objdump.pl | 18 +++---
1 file changed, 15 insertions(+), 3 deletions(-)
diff --git
Il 16/08/2013 23:58, Alex Williamson ha scritto:
Since commit 23326164 we align access sizes to match the alignment of
the address, but we don't align the access size itself. This means we
let illegal access sizes (ex. 3) slip through if the address is
sufficiently aligned (ex. 4). This
The QEMU mascot which was already used for the NSIS installer
is now used for all QEMU executables.
Signed-off-by: Stefan Weil s...@weilnetz.de
---
version.rc |2 ++
1 file changed, 2 insertions(+)
diff --git a/version.rc b/version.rc
index 82e10ec..a50d62f 100644
--- a/version.rc
+++
On 08/16/13 23:58, Alex Williamson wrote:
Since commit 23326164 we align access sizes to match the alignment of
the address, but we don't align the access size itself. This means we
let illegal access sizes (ex. 3) slip through if the address is
sufficiently aligned (ex. 4). This results in
Hi,
The fundamental reason is traditional qemu-dm and upstream qemu using a
different vgabios-cirrus.bin,
qemu-dm is using xen-4.1.2/tools/firmware/vgabios/
VGABIOS-lgpl-latest.cirrus.bin,
but upstream qemu is using qemu-1.2.2/pc-bios/vgabios-cirrus.bin
the pivotal patch is :
# HG changeset
(side point)
On 08/17/13 10:23, Laszlo Ersek wrote:
if (l = access_size_max) {
return access_size_max;
}
return 1 max(3, lmb(l));
lol, of course this should have been min()...
Alex's patch is OK of course.
Laszlo
On 08/16/13 13:13, arm...@redhat.com wrote:
static void pc_init_isa(QEMUMachineInitArgs *args)
{
-ram_addr_t ram_size = args-ram_size;
-const char *cpu_model = args-cpu_model;
-const char *kernel_filename = args-kernel_filename;
-const char *kernel_cmdline =
On 08/16/13 13:13, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
Don't explode when the variable is used just a few times, and never
changed.
Signed-off-by: Markus Armbruster arm...@redhat.com
---
hw/i386/pc_q35.c | 22 +-
1 file changed, 9
comments below
On 08/16/13 13:13, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
Don't explode QEMUMachineInitArgs before passing it to
sun4m_hw_init(), sun4uv_init().
Signed-off-by: Markus Armbruster arm...@redhat.com
---
hw/sparc/sun4m.c | 113
On 08/16/13 13:13, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
Don't explode when the variable is used just once, and never changed.
Signed-off-by: Markus Armbruster arm...@redhat.com
Acked-by: Alexander Graf ag...@suse.de
---
hw/ppc/e500plat.c | 18
comments below
On 08/16/13 13:13, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
Pass on the generic arguments unadulterated, and the machine-specific
ones as separate argument.
Signed-off-by: Markus Armbruster arm...@redhat.com
Acked-by: Alexander Graf ag...@suse.de
On 08/16/13 13:13, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
We set default boot order cad in every single machine definition
except pseries and moxiesim, even though very few boards actually
care for boot order, and cad makes sense for even fewer.
Machines that
Eric,
On 08/16/13 15:18, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
This gets rid of one of the last get_param_value() users, makes
multiple -smbios work sanely, cleans up the gross side effect in
qemu_uuid_parse(), and more. Topped off with a little feature in the
On 08/16/2013 07:18 AM, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
Currently, -smbios type=T,NAME=VAL,... adds one field (T,NAME) with
value VAL to fw_cfg for each unique NAME. If NAME occurs multiple
times, the last one's VAL is used (before the QemuOpts conversion,
On 08/17/2013 06:08 AM, Laszlo Ersek wrote:
Eric,
On 08/16/13 15:18, arm...@redhat.com wrote:
From: Markus Armbruster arm...@redhat.com
This gets rid of one of the last get_param_value() users, makes
multiple -smbios work sanely, cleans up the gross side effect in
qemu_uuid_parse(), and
Am 16.08.2013 15:18, schrieb arm...@redhat.com:
From: Markus Armbruster arm...@redhat.com
I'd like to access QEMUMachine from a QEMUMachine init() method, which
is currently not possible. Instead of passing it as an argument, I
simply set current_machine earlier.
We had such a patch for
On Sat, 2013-08-17 at 10:23 +0200, Laszlo Ersek wrote:
On 08/16/13 23:58, Alex Williamson wrote:
Since commit 23326164 we align access sizes to match the alignment of
the address, but we don't align the access size itself. This means we
let illegal access sizes (ex. 3) slip through if the
On Sat, 2013-08-17 at 08:33 +0200, Paolo Bonzini wrote:
Il 16/08/2013 23:58, Alex Williamson ha scritto:
Since commit 23326164 we align access sizes to match the alignment of
the address, but we don't align the access size itself. This means we
let illegal access sizes (ex. 3) slip through
Il 17/08/2013 10:23, Laszlo Ersek ha scritto:
What Paolo posted seems almost identical.
clz32(l): leading zeros in l
qemu_fls(l) == 32 - clz32(l): position of leftmost bit set, 1-based
qemu_fls(l) - 1: position of leftmost bit set, 0-based
Not sure if the
Allow host detection on linux systems without glibc 2.16 or later.
Signed-off-by: Richard Henderson r...@twiddle.net
---
include/elf.h | 34 ++
tcg/ppc64/tcg-target.c | 11 ++-
2 files changed, 36 insertions(+), 9 deletions(-)
diff --git
Abstract away dependence on a system implementation of getauxval.
Signed-off-by: Richard Henderson r...@twiddle.net
---
include/qemu/osdep.h | 20
linux-user/main.c| 1 +
util/Makefile.objs | 1 +
util/getauxval.c | 91
The getauxval routine was added to glibc for 2.16. In order to
better support qemu on systems prior to 2.16, add a qemu_getauxval
entry point, and add the relevant defines to our own elf.h.
r~
Richard Henderson (6):
osdep: Create qemu_getauxval and qemu_init_auxval
tcg-ppc64: Use
No need to set up a SIGILL signal handler for detection anymore.
Remove a ton of sanity checks that must be true, given that we're
requiring a 64-bit build (the note about 31-bit KVM is satisfied
by configuring with TCI).
Signed-off-by: Richard Henderson r...@twiddle.net
---
include/elf.h
Allow host detection on linux systems without glibc 2.16 or later.
Signed-off-by: Richard Henderson r...@twiddle.net
---
include/elf.h| 22 ++
tcg/arm/tcg-target.c | 15 ++-
2 files changed, 28 insertions(+), 9 deletions(-)
diff --git a/include/elf.h
Allow host detection on non-linux hosts.
Signed-off-by: Richard Henderson r...@twiddle.net
---
util/getauxval.c | 56 ++--
1 file changed, 54 insertions(+), 2 deletions(-)
diff --git a/util/getauxval.c b/util/getauxval.c
index 55b639c..2e93a14
With this we no longer pass down envp, and thus all systems can have
the same void prototype. So also eliminate a useless thunk.
Signed-off-by: Richard Henderson r...@twiddle.net
---
include/qemu/cache-utils.h | 4 ++--
linux-user/main.c | 2 +-
util/cache-utils.c | 51
With the optimization in tcg_liveness_analysis,
we can avoid the MFLO when it is unused.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/mips/tcg-target.c | 10 ++
tcg/mips/tcg-target.h | 4 ++--
2 files changed, 12 insertions(+), 2 deletions(-)
diff --git
Use them in places where mulu2 and muls2 are used.
Optimize mulx2 with dead low part to mulxh.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/aarch64/tcg-target.h | 4
tcg/arm/tcg-target.h | 2 ++
tcg/hppa/tcg-target.h| 2 ++
tcg/i386/tcg-target.h| 4
We have -- or will have -- several targets which have a native
multiply-highpart instruction: ppc*, ia64, aarch64, alpha.
If we leave only the mul[us]2 opcode with which to expose this,
we have to handle the register allocation bits in the backends.
Better, IMO, to expose the two parts at the TCG
Using these instead of mulu2 and muls2 lets us avoid having to argument
overlap analysis in the backend. Normal register allocation will DTRT.
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/ppc64/tcg-target.c | 32 +++-
tcg/ppc64/tcg-target.h | 8
Signed-off-by: Richard Henderson r...@twiddle.net
---
tcg/optimize.c | 23 +++
1 file changed, 23 insertions(+)
diff --git a/tcg/optimize.c b/tcg/optimize.c
index e8dedf3..b29bf25 100644
--- a/tcg/optimize.c
+++ b/tcg/optimize.c
@@ -304,6 +304,25 @@ static TCGArg
Hi Alli have written a java program, keep sending the gdb single step
command to qemu, i can execute ~60 single steps per seconds. But in bochs i can
execute ~130. How can I make qemu run faster in single step?
Thanksfrom Peter
The socket is lagging, i keep sending gdb command to qemu and read the result,
for every 60 commands, i got 4 seconds lagging, it hangs on the socket:read()
in my side for 4 seconds.
The qemu is using g_io_channel_write_chars() function to send back the data
through socket, anyway to optimize
On 08/16/2013 11:17:06 AM, Herbei Dacian wrote:
my system should run in far less memory. something like 2-4MB.
but first I need to have a system running so that I can monitor with
qemu the addresses accessed for read execute and write by the code
run by the emulator.
if I reach that is a
58 matches
Mail list logo
