On Wed, Jun 08, 2016 at 17:04:19 +0200, Peter Krempa wrote:
On Wed, Jun 08, 2016 at 14:41:30 +0200, Jiri Denemark wrote:
> CPUID instruction normally takes its parameter from EAX, but sometimes
> ECX is used as an additional parameter. This patch prepares the x86 CPU
> driver code for the new 'ecx_in' CPUID parameter.
>
> Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com>
> ---
> src/cpu/cpu_x86.c | 52 +++++++++++++---------
> src/cpu/cpu_x86_data.h | 1 +
> src/qemu/qemu_monitor_json.c | 4 ++
> .../qemumonitorjson-getcpu-ecx.data | 10 ++---
> .../qemumonitorjson-getcpu-full.data | 6 +--
> .../qemumonitorjson-getcpu-host.data | 8 ++--
> 6 files changed, 49 insertions(+), 32 deletions(-)
>
> diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c
> index 82921db..ae809de 100644
> --- a/src/cpu/cpu_x86.c
> +++ b/src/cpu/cpu_x86.c
[...]
> @@ -1860,7 +1869,8 @@ cpuidCall(virCPUx86CPUID *cpuid)
> "=b" (cpuid->ebx),
> "=c" (cpuid->ecx),
> "=d" (cpuid->edx)
> - : "a" (cpuid->eax_in));
> + : "a" (cpuid->eax_in),
> + "c" (cpuid->ecx_in));
> # else
> /* we need to avoid direct use of ebx for CPUID output as it is used
> * for global offset table on i386 with -fPIC
> @@ -1876,7 +1886,8 @@ cpuidCall(virCPUx86CPUID *cpuid)
> "=r" (cpuid->ebx),
> "=c" (cpuid->ecx),
> "=d" (cpuid->edx)
> - : "a" (cpuid->eax_in)
> + : "a" (cpuid->eax_in),
> + "c" (cpuid->ecx_in)
I think the removal of clearing of the ECX register from the next patch
in the asm code above actually belongs here.
Good catch, apparently I was not careful enough when splitting the
changes in two patches.
Jirka