On 03/07/2014 03:32 AM, Peter Maydell wrote:
This is v4 of the AArch64 system emulation patches, and it's
an important milestone -- this is enough to boot a Linux kernel.
Does this boot an aarch64 kernel with tcg on x86 host?
--
Thanks,
Xuebing Wang
Probably no one uses disable-kvm for x86_64.
Also move 3 kvm-i386 function declarations from sysemu/kvm.h
=> target-i386/kvm_i386.h
- kvm_pc_gsi_handler()
- kvm_pc_setup_irq_routing()
- kvm_i8259_init()
Signed-off-by: Xuebing Wang
---
hw/i386/kvm/i8259.c|2 +-
hw/i386/kvm/ioapi
The build issue was caused by the patch of removing
"#if defined CONFIG_KVM || !defined NEED_CPU_H"
Signed-off-by: Xuebing Wang
---
kvm-stub.c |1 +
1 file changed, 1 insertion(+)
diff --git a/kvm-stub.c b/kvm-stub.c
index e979f76..f859249 100644
--- a/kvm-stub.c
+++ b/kvm-stub.
Size of text section of ppc64-softmmu/qemu-system-ppc64 is reduced by
152 bytes.
Signed-off-by: Xuebing Wang
---
include/hw/ppc/openpic.h |1 -
target-ppc/Makefile.objs |1 -
target-ppc/kvm-stub.c| 18 --
target-ppc/kvm_ppc.h |5 +
4 files changed, 5
As explained in previous patch, kvm functions won't be optimized out for
non-CONFIG_KVM.
Signed-off-by: Xuebing Wang
---
target-i386/kvm-stub.c |6 --
1 file changed, 6 deletions(-)
diff --git a/target-i386/kvm-stub.c b/target-i386/kvm-stub.c
index 2b9e801..0ef642d 100644
--- a/t
- (b) is identical to (a)
- result of after patch is identical to before patch
Signed-off-by: Xuebing Wang
---
include/sysemu/kvm.h | 13 -
1 file changed, 13 deletions(-)
diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h
index a02d67c..1829206 100644
--- a/include/
To eliminate kvm-stub for target-i386.
Size of text section of x86_64-linux-user/qemu-x86_64 is reduced by 104 bytes
by inline these 2 functions.
Signed-off-by: Xuebing Wang
---
include/sysemu/kvm.h |3 ---
target-i386/Makefile.objs |1 -
target-i386/cpu.c |2
Hi Community,
I am not sure if there is value for this patchset.
After the first pach:
Size of x86_64-softmmu/qemu-system-x86_64 is unchanged.
Size of sh4-softmmu/qemu-system-sh4 increases by about 2.8KB.
Xuebing Wang (7):
kvm: remove the hack "#if defined CONFIG_KVM || !defined NEED_
The build issue was caused by the patch of removing
"#if defined CONFIG_KVM || !defined NEED_CPU_H"
Signed-off-by: Xuebing Wang
---
target-ppc/kvm_ppc.h | 12
1 file changed, 12 insertions(+)
diff --git a/target-ppc/kvm_ppc.h b/target-ppc/kvm_ppc.h
index 5f78e4b..c532
Hi Peter,
Thanks. You are correct, I don't know what I was thinking. :-)
On 03/04/2014 08:34 PM, Peter Maydell wrote:
On 4 March 2014 12:09, Xuebing wang wrote:
target-*/gdbstub.c implementers only need to know gdbstub hooks (thus
gdbstub API), they don't care "cpu.h", a
#ifdef CONFIG_KVM
#include
#include
@@ -169,6 +178,7 @@ int kvm_init_vcpu(CPUState *cpu);
int kvm_cpu_exec(CPUState *cpu);
#ifdef NEED_CPU_H
+#include "cpu.h" /* target-xxx/cpu.h, required for target_ulong */
void kvm_setup_guest_memory(void *start, size_t size);
void kvm_flush_coales
index 367eda1..f0157e3 100644
--- a/include/qom/cpu.h
+++ b/include/qom/cpu.h
@@ -510,18 +510,4 @@ void qemu_init_vcpu(CPUState *cpu);
*/
void cpu_single_step(CPUState *cpu, int enabled);
-#ifdef CONFIG_SOFTMMU
-extern const struct VMStateDescription vmstate_cpu_common;
-#else
-#define vmsta
diff --git a/include/exec/gdbstub.h b/include/exec/gdbstub.h
index a608a26..14addcb 100644
--- a/include/exec/gdbstub.h
+++ b/include/exec/gdbstub.h
@@ -11,6 +11,8 @@
#define GDB_WATCHPOINT_ACCESS4
#ifdef NEED_CPU_H
+#include "cpu.h" /* target-xxx/cpu.h, required for target_ulong,
+
diff --git a/include/disas/disas.h b/include/disas/disas.h
index c13ca9a..e5cdfd7 100644
--- a/include/disas/disas.h
+++ b/include/disas/disas.h
@@ -1,9 +1,9 @@
#ifndef _QEMU_DISAS_H
#define _QEMU_DISAS_H
-#include "qemu-common.h"
-
#ifdef NEED_CPU_H
+#include "cpu.h" /* target-xxx/cpu.h, re
Hi Paolo, thanks for reviewing.
On 03/04/2014 06:19 PM, Paolo Bonzini wrote:
Hi,
in general I agree with this patch. I have a few comments, and I
suggest that you split it in multiple patches so that it's easier to
get it in when each part is ready.
I spent some time trying to reduce it.
On 03/04/2014 05:42 PM, Stefan Hajnoczi wrote:
On Tue, Mar 04, 2014 at 10:47:21AM +0800, Xuebing Wang wrote:
Signed-off-by: Xuebing Wang
---
docs/api-hierarchy.txt | 93
1 file changed, 93 insertions(+)
create mode 100644 docs/api
Hi Andreas, thank you very much for your reply.
Would you please help review/correct doc/api-hierarchy too?
On 03/04/2014 11:45 AM, Andreas Färber wrote:
Hi Xuebing,
Am 04.03.2014 03:47, schrieb Xuebing Wang:
Q2) Does it make sense to remove NEED_CPU_H from qemu-common.h?
IMO not in this
Use vaddr (instead of target_ulong) for virtual address. See document HACKING
for more information about vaddr and target_ulong.
Signed-off-by: Xuebing Wang
---
include/sysemu/memory_mapping.h |3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/include/sysemu
include/exec/exec-all.h should ONLY be included in target-xxx/*
- 'git grep -nw exec-all.h' confirms this
Signed-off-by: Xuebing Wang
---
cputlb.c |1 -
tcg/README |2 +-
tci.c |2 +-
3 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/cputlb.c b/cput
Note: there is a FIXME to be addressed in this patch.
For every appearance of NEED_CPU_H, there must be '#include "cpu.h"' to
include "target-xxx/cpu.h", because the code below NEED_CPU_H depends on
architecture-specific information.
Signed-off-by: Xuebing Wang
This patch does below:
- Move the declaration of 2 translate functions from translate-all.h into
include/exec/translate.h
- remove file translate-all.h
Signed-off-by: Xuebing Wang
---
exec.c |2 --
include/exec/translate.h |4
translate-all.c
In preparation for removing include/exec/address-spaces.h
Signed-off-by: Xuebing Wang
---
include/exec/address-spaces.h | 29 -
include/exec/memory.h | 19 +++
include/hw/sh4/sh_intc.h |1 +
3 files changed, 16 insertions(+), 33
Remove '#include "exec/address-spaces.h"' is done by:
git grep -w "exec\/address-spaces.h" | cut -d: -f1 |
xargs sed -i '/exec\/address-spaces.h/d'
Signed-off-by: Xuebing Wang
---
arch_init.c |1 -
cputlb.c
Signed-off-by: Xuebing Wang
---
include/exec/exec-all.h | 302 +
include/exec/translate.h | 306 ++
2 files changed, 307 insertions(+), 301 deletions(-)
create mode 100644 include/exec/translate.h
diff
h] .
Note: remove unnecessary inclusion of "cpu-all.h" too. "cpu-all.h" is
architecture-specific, should ONLY be included from target-xxx/*
- 'git grep -nw "cpu-all.h"' confirms this
Signed-off-by: Xuebing Wang
---
arch_init.c |1 -
cpu-exec
ormation. Here are
a few examples:
- CPUArchState
- TARGET_LONG_BITS
- TARGET_PAGE_BITS
- TARGET_PHYS_ADDR_SPACE_BITS, TARGET_VIRT_ADDR_SPACE_BITS
And constructs that are derived from above, examples are:
- target_long, target_ulong
Examples of non-architecture-dependent are:
- hwaddr
- v
Signed-off-by: Xuebing Wang
---
docs/api-hierarchy.txt | 93
1 file changed, 93 insertions(+)
create mode 100644 docs/api-hierarchy.txt
diff --git a/docs/api-hierarchy.txt b/docs/api-hierarchy.txt
new file mode 100644
index 000..9982f51
ude/qemu/*?
Q5) Should we move exec/def-helper.h => tcg/def-helper.h and
exec/spinlock.h => include/qemu/spinlock.h, to reflect the definition of
exec is: memory API, translation API and cpu API?
After this patchset, cpu-all.h and exec-all.h should ONLY be included from
target-xxx/*.
T
-aarch64 needs to be reliable.
-- Thanks,
Xuebing Wang
Signed-off-by: Xuebing Wang
Reviewed-By: Alex Bligh
---
qemu-timer.c | 97 ++
1 file changed, 50 insertions(+), 47 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index c9801da..e592c14 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
Signed-off-by: Xuebing Wang
Reviewed-By: Alex Bligh
---
qemu-timer.c | 86 ++
1 file changed, 45 insertions(+), 41 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index 471150c..6f13a76 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
Signed-off-by: Xuebing Wang
Reviewed-By: Alex Bligh
---
qemu-timer.c | 160 ++
1 file changed, 82 insertions(+), 78 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index 2db87ba..471150c 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
Signed-off-by: Xuebing Wang
Reviewed-By: Alex Bligh
---
qemu-timer.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index 6f13a76..c9801da 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
@@ -548,6 +548,9 @@ int qemu_poll_ns(GPollFD
Signed-off-by: Xuebing Wang
Reviewed-By: Alex Bligh
---
qemu-timer.c |9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index e592c14..21aff82 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
@@ -22,13 +22,10 @@
* THE SOFTWARE
already information in the header file include/qemu/timer.h
-- "As above" part of the information is wrong
/*
* As above, but return -1 for no deadline, and do not cap to 2^32
* as we know the result is always positive.
*/
Signed-off-by: Xuebing Wang
Reviewed-By: Alex Bligh
--
This patch does below:
- Make QEMUTimerList functions private (remove from APIs)
- Comment out function QEMUTimerList timerlist_get_clock() in order
to fix compile error
Signed-off-by: Xuebing Wang
Reviewed-By: Alex Bligh
---
include/qemu/timer.h | 94
nnecessary #include and use minimal required #include
--------
Xuebing Wang (7):
timer: move QEMUTimerList functions together
timer: make QEMUTimerList functions private (remove from APIs)
timer: move QEMUClockType related
tream quality. I expect we'll have
this done within a month or two, but it won't make it into the QEMU
2.0 release (slightly too late).
Hi Peter,
Would you please point to me where can I get this work-in-progress code?
--
Thanks,
Xuebing Wang
o the qemu seems use 32-bit ARM cpu model whereas I have
need to use 64-bit ARM cpu model as I have specified by "-cpu host"
qemu parameter in Foundationv8 model.
-Jani
2014-02-27 14:21 GMT+01:00 Xuebing wang <mailto:xbi...@gmail.com>>:
Hi Jani,
Is below video s
emory address 0xffc000636498: Unknown error -1"
In addition stepi or any other run related command does not work.
You would need to implement breakpoint support for AArch64 in
both KVM and QEMU, I think. That feature is currently not implemented.
thanks
-- PMM
--
Thanks,
Xuebing Wang
opc & ~MO_SIGN]' ...)
tcg_out_calli spots that the displacement is too big for a call insn
and emits the
0x2aaade72d12c: mov$0x2afd98602c10,%r10
0x2aaade72d136: callq *%r10 // Call helper function
thanks
-- PMM
--
Thanks,
Xuebing Wang
e is to make the design easier to understand for new engineers.
Thanks again.
On 02/26/2014 06:34 PM, Paolo Bonzini wrote:
Il 26/02/2014 10:55, Peter Maydell ha scritto:
On 26 February 2014 09:25, Xuebing Wang wrote:
This patch does below:
- Move the declaration of 2 translate functions
xargs sed -i 's/\/exec\/exec-all.h/g'
Note:
1) "exact whole word match" is considered.
2) We may move translate related from include/exec/exec-all.h into
include/exec/translate.h later.
Signed-off-by: Xuebing Wang
---
exec.c |2 +-
include/exec/e
This patch does below:
- Make QEMUTimerList functions private (remove from APIs)
- Comment out function QEMUTimerList timerlist_get_clock() in order
to fix compile error
Signed-off-by: Xuebing Wang
---
include/qemu/timer.h | 94
qemu
Signed-off-by: Xuebing Wang
---
qemu-timer.c | 86 ++
1 file changed, 45 insertions(+), 41 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index 471150c..6f13a76 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
@@ -456,6 +456,51
Signed-off-by: Xuebing Wang
---
qemu-timer.c |9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index e592c14..21aff82 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
@@ -22,13 +22,10 @@
* THE SOFTWARE.
*/
-#include "sysemu/sys
Signed-off-by: Xuebing Wang
---
qemu-timer.c | 97 ++
1 file changed, 50 insertions(+), 47 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index c9801da..e592c14 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
@@ -501,53 +501,6
Signed-off-by: Xuebing Wang
---
qemu-timer.c | 13 -
1 file changed, 8 insertions(+), 5 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index 6f13a76..c9801da 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
@@ -548,6 +548,9 @@ int qemu_poll_ns(GPollFD *fds, guint nfds, int64_t
gether and make them private
- put QEMUClockType related functions together
- move QEMUTimerListGroup function to be below QEMUClockType
- put QEMUTimer related functions together
- put general utility functions together
- clean unnecessary #include and use minimal required #include
Xuebing Wang (7)
Signed-off-by: Xuebing Wang
---
qemu-timer.c | 160 ++
1 file changed, 82 insertions(+), 78 deletions(-)
diff --git a/qemu-timer.c b/qemu-timer.c
index 2db87ba..471150c 100644
--- a/qemu-timer.c
+++ b/qemu-timer.c
@@ -314,6 +314,10
already information in the header file include/qemu/timer.h
-- "As above" part of the information is wrong
/*
* As above, but return -1 for no deadline, and do not cap to 2^32
* as we know the result is always positive.
*/
Signed-off-by: Xuebing Wang
---
include/qemu/time
?
- qemu_clock_warp
- timer_get and timer_put
- cpu_get_ticks, cpu_enable_ticks, cpu_disable_ticks
Q4) Do we need this API qemu_clock_get_main_loop_timerlist?
--
Thanks,
Xuebing Wang
52 matches
Mail list logo
