diff mbox series

[v5] i386/cpu_dump: support AVX512 ZMM regs dump

Message ID 1618538904-93433-1-git-send-email-robert.hu@linux.intel.com (mailing list archive)
State New, archived
Headers show
Series [v5] i386/cpu_dump: support AVX512 ZMM regs dump | expand

Commit Message

Robert Hoo April 16, 2021, 2:08 a.m. UTC
Since commit fa4518741e (target-i386: Rename struct XMMReg to ZMMReg),
CPUX86State.xmm_regs[] has already been extended to 512bit to support
AVX512.
Also, other qemu level supports for AVX512 registers are there for
years.
But in x86_cpu_dump_state(), still only dump XMM registers no matter
YMM/ZMM is enabled.
This patch is to complement this, let it dump XMM/YMM/ZMM accordingly.

Signed-off-by: Robert Hoo <robert.hu@linux.intel.com>
---
Changelog:
v5: fix a minor issue. rebase to latest master.
v4: stringent AVX512 case and AVX case judgement criteria
v3: fix some coding style issue.
v2: dump XMM/YMM/ZMM according to XSAVE state-components enablement.
 target/i386/cpu-dump.c | 62 ++++++++++++++++++++++++++++++++++++++------------
 1 file changed, 47 insertions(+), 15 deletions(-)

Comments

Richard Henderson April 16, 2021, 5:27 p.m. UTC | #1
On 4/15/21 7:08 PM, Robert Hoo wrote:
> Since commit fa4518741e (target-i386: Rename struct XMMReg to ZMMReg),
> CPUX86State.xmm_regs[] has already been extended to 512bit to support
> AVX512.
> Also, other qemu level supports for AVX512 registers are there for
> years.
> But in x86_cpu_dump_state(), still only dump XMM registers no matter
> YMM/ZMM is enabled.
> This patch is to complement this, let it dump XMM/YMM/ZMM accordingly.
> 
> Signed-off-by: Robert Hoo<robert.hu@linux.intel.com>
> ---
> Changelog:
> v5: fix a minor issue. rebase to latest master.
> v4: stringent AVX512 case and AVX case judgement criteria
> v3: fix some coding style issue.
> v2: dump XMM/YMM/ZMM according to XSAVE state-components enablement.
>   target/i386/cpu-dump.c | 62 ++++++++++++++++++++++++++++++++++++++------------
>   1 file changed, 47 insertions(+), 15 deletions(-)

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~
Eduardo Habkost April 19, 2021, 8:18 p.m. UTC | #2
On Fri, Apr 16, 2021 at 10:08:24AM +0800, Robert Hoo wrote:
> Since commit fa4518741e (target-i386: Rename struct XMMReg to ZMMReg),
> CPUX86State.xmm_regs[] has already been extended to 512bit to support
> AVX512.
> Also, other qemu level supports for AVX512 registers are there for
> years.
> But in x86_cpu_dump_state(), still only dump XMM registers no matter
> YMM/ZMM is enabled.
> This patch is to complement this, let it dump XMM/YMM/ZMM accordingly.
> 
> Signed-off-by: Robert Hoo <robert.hu@linux.intel.com>

Queued for 6.1, thanks!
Eduardo Habkost April 19, 2021, 9:59 p.m. UTC | #3
On Mon, Apr 19, 2021 at 04:18:25PM -0400, Eduardo Habkost wrote:
> On Fri, Apr 16, 2021 at 10:08:24AM +0800, Robert Hoo wrote:
> > Since commit fa4518741e (target-i386: Rename struct XMMReg to ZMMReg),
> > CPUX86State.xmm_regs[] has already been extended to 512bit to support
> > AVX512.
> > Also, other qemu level supports for AVX512 registers are there for
> > years.
> > But in x86_cpu_dump_state(), still only dump XMM registers no matter
> > YMM/ZMM is enabled.
> > This patch is to complement this, let it dump XMM/YMM/ZMM accordingly.
> > 
> > Signed-off-by: Robert Hoo <robert.hu@linux.intel.com>
> 
> Queued for 6.1, thanks!

Dequeuing, as it causes build failures on multiple configs:

https://gitlab.com/ehabkost/qemu/-/pipelines/288890306

Example:


../target/i386/cpu-dump.c: In function 'x86_cpu_dump_state':
../target/i386/cpu-dump.c:511:50: error: format '%lx' expects argument of type 'long unsigned int', but argument 4 has type 'uint64_t' {aka 'long long unsigned int'} [-Werror=format=]
                 qemu_fprintf(f, "Opmask%02d=%016lx%s", i, env->opmask_regs[i],
                                             ~~~~~^        ~~~~~~~~~~~~~~~~~~~
                                             %016llx
../target/i386/cpu-dump.c:517:47: error: format '%lx' expects argument of type 'long unsigned int', but argument 4 has type 'uint64_t' {aka 'long long unsigned int'} [-Werror=format=]
                 qemu_fprintf(f, "ZMM%02d=%016lx %016lx %016lx %016lx %016lx "
                                          ~~~~~^
                                          %016llx
Robert Hoo May 6, 2021, 8:56 a.m. UTC | #4
On Mon, 2021-04-19 at 17:59 -0400, Eduardo Habkost wrote:
> On Mon, Apr 19, 2021 at 04:18:25PM -0400, Eduardo Habkost wrote:
> > On Fri, Apr 16, 2021 at 10:08:24AM +0800, Robert Hoo wrote:
> > > Since commit fa4518741e (target-i386: Rename struct XMMReg to
> > > ZMMReg),
> > > CPUX86State.xmm_regs[] has already been extended to 512bit to
> > > support
> > > AVX512.
> > > Also, other qemu level supports for AVX512 registers are there
> > > for
> > > years.
> > > But in x86_cpu_dump_state(), still only dump XMM registers no
> > > matter
> > > YMM/ZMM is enabled.
> > > This patch is to complement this, let it dump XMM/YMM/ZMM
> > > accordingly.
> > > 
> > > Signed-off-by: Robert Hoo <robert.hu@linux.intel.com>
> > 
> > Queued for 6.1, thanks!
> 
> Dequeuing, as it causes build failures on multiple configs:
> 
> https://gitlab.com/ehabkost/qemu/-/pipelines/288890306
> 
> Example:
> 
> 
> ../target/i386/cpu-dump.c: In function 'x86_cpu_dump_state':
> ../target/i386/cpu-dump.c:511:50: error: format '%lx' expects
> argument of type 'long unsigned int', but argument 4 has type
> 'uint64_t' {aka 'long long unsigned int'} [-Werror=format=]
>                  qemu_fprintf(f, "Opmask%02d=%016lx%s", i, env-
> >opmask_regs[i],
>                                              ~~~~~^        ~~~~~~~~~~
> ~~~~~~~~~
>                                              %016llx
> ../target/i386/cpu-dump.c:517:47: error: format '%lx' expects
> argument of type 'long unsigned int', but argument 4 has type
> 'uint64_t' {aka 'long long unsigned int'} [-Werror=format=]
>                  qemu_fprintf(f, "ZMM%02d=%016lx %016lx %016lx %016lx
> %016lx "
>                                           ~~~~~^
>                                           %016llx
> 
Hi Eduardo,

I've sent v6 
https://patchwork.kernel.org/project/qemu-devel/patch/1618986232-73826-1-git-send-email-robert.hu@linux.intel.com/
, would you retest and queue it?
Thanks.

>
diff mbox series

Patch

diff --git a/target/i386/cpu-dump.c b/target/i386/cpu-dump.c
index aac21f1..ece64e1 100644
--- a/target/i386/cpu-dump.c
+++ b/target/i386/cpu-dump.c
@@ -478,6 +478,11 @@  void x86_cpu_dump_state(CPUState *cs, FILE *f, int flags)
     qemu_fprintf(f, "EFER=%016" PRIx64 "\n", env->efer);
     if (flags & CPU_DUMP_FPU) {
         int fptag;
+        const uint64_t avx512_mask = XSTATE_OPMASK_MASK | \
+                                     XSTATE_ZMM_Hi256_MASK | \
+                                     XSTATE_Hi16_ZMM_MASK | \
+                                     XSTATE_YMM_MASK | XSTATE_SSE_MASK,
+                       avx_mask = XSTATE_YMM_MASK | XSTATE_SSE_MASK;
         fptag = 0;
         for(i = 0; i < 8; i++) {
             fptag |= ((!env->fptags[i]) << i);
@@ -499,21 +504,48 @@  void x86_cpu_dump_state(CPUState *cs, FILE *f, int flags)
             else
                 qemu_fprintf(f, " ");
         }
-        if (env->hflags & HF_CS64_MASK)
-            nb = 16;
-        else
-            nb = 8;
-        for(i=0;i<nb;i++) {
-            qemu_fprintf(f, "XMM%02d=%08x%08x%08x%08x",
-                         i,
-                         env->xmm_regs[i].ZMM_L(3),
-                         env->xmm_regs[i].ZMM_L(2),
-                         env->xmm_regs[i].ZMM_L(1),
-                         env->xmm_regs[i].ZMM_L(0));
-            if ((i & 1) == 1)
-                qemu_fprintf(f, "\n");
-            else
-                qemu_fprintf(f, " ");
+
+        if ((env->xcr0 & avx512_mask) == avx512_mask) {
+            /* XSAVE enabled AVX512 */
+            for (i = 0; i < NB_OPMASK_REGS; i++) {
+                qemu_fprintf(f, "Opmask%02d=%016lx%s", i, env->opmask_regs[i],
+                    ((i & 3) == 3) ? "\n" : " ");
+            }
+
+            nb = (env->hflags & HF_CS64_MASK) ? 32 : 8;
+            for (i = 0; i < nb; i++) {
+                qemu_fprintf(f, "ZMM%02d=%016lx %016lx %016lx %016lx %016lx "
+                                "%016lx %016lx %016lx\n",
+                             i,
+                             env->xmm_regs[i].ZMM_Q(7),
+                             env->xmm_regs[i].ZMM_Q(6),
+                             env->xmm_regs[i].ZMM_Q(5),
+                             env->xmm_regs[i].ZMM_Q(4),
+                             env->xmm_regs[i].ZMM_Q(3),
+                             env->xmm_regs[i].ZMM_Q(2),
+                             env->xmm_regs[i].ZMM_Q(1),
+                             env->xmm_regs[i].ZMM_Q(0));
+            }
+        } else if ((env->xcr0 & avx_mask)  == avx_mask) {
+            /* XSAVE enabled AVX */
+            nb = env->hflags & HF_CS64_MASK ? 16 : 8;
+            for (i = 0; i < nb; i++) {
+                qemu_fprintf(f, "YMM%02d=%016lx %016lx %016lx %016lx\n",
+                             i,
+                             env->xmm_regs[i].ZMM_Q(3),
+                             env->xmm_regs[i].ZMM_Q(2),
+                             env->xmm_regs[i].ZMM_Q(1),
+                             env->xmm_regs[i].ZMM_Q(0));
+            }
+        } else { /* SSE and below cases */
+            nb = env->hflags & HF_CS64_MASK ? 16 : 8;
+            for (i = 0; i < nb; i++) {
+                qemu_fprintf(f, "XMM%02d=%016lx %016lx%s",
+                             i,
+                             env->xmm_regs[i].ZMM_Q(1),
+                             env->xmm_regs[i].ZMM_Q(0),
+                             (i & 1) ? "\n" : " ");
+            }
         }
     }
     if (flags & CPU_DUMP_CODE) {