We do not need or want to be allocating page sized quanta.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
Cc: Stefan Weil <sw@weilnetz.de>
---
 util/oslib-win32.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/util/oslib-win32.c b/util/oslib-win32.c
index c654dafd93..8d838bf342 100644
--- a/util/oslib-win32.c
+++ b/util/oslib-win32.c
@@ -56,10 +56,8 @@ void *qemu_try_memalign(size_t alignment, size_t size)
 {
     void *ptr;
 
-    if (!size) {
-        abort();
-    }
-    ptr = VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
+    assert(size != 0);
+    ptr = _aligned_malloc(alignment, size);
     trace_qemu_memalign(alignment, size, ptr);
     return ptr;
 }
@@ -93,9 +91,7 @@ void *qemu_anon_ram_alloc(size_t size, uint64_t *align, bool shared)
 void qemu_vfree(void *ptr)
 {
     trace_qemu_vfree(ptr);
-    if (ptr) {
-        VirtualFree(ptr, 0, MEM_RELEASE);
-    }
+    _aligned_free(ptr);
 }
 
 void qemu_anon_ram_free(void *ptr, size_t size)
-- 
2.25.1

Am 16.09.20 um 02:46 schrieb Richard Henderson:

> We do not need or want to be allocating page sized quanta.
>
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> ---
> Cc: Stefan Weil <sw@weilnetz.de>
> ---
>  util/oslib-win32.c | 10 +++-------
>  1 file changed, 3 insertions(+), 7 deletions(-)
>
> diff --git a/util/oslib-win32.c b/util/oslib-win32.c
> index c654dafd93..8d838bf342 100644
> --- a/util/oslib-win32.c
> +++ b/util/oslib-win32.c
> @@ -56,10 +56,8 @@ void *qemu_try_memalign(size_t alignment, size_t size)
>  {
>      void *ptr;
>  
> -    if (!size) {
> -        abort();
> -    }
> -    ptr = VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
> +    assert(size != 0);
> +    ptr = _aligned_malloc(alignment, size);
>      trace_qemu_memalign(alignment, size, ptr);
>      return ptr;
>  }
> @@ -93,9 +91,7 @@ void *qemu_anon_ram_alloc(size_t size, uint64_t *align, bool shared)
>  void qemu_vfree(void *ptr)
>  {
>      trace_qemu_vfree(ptr);
> -    if (ptr) {
> -        VirtualFree(ptr, 0, MEM_RELEASE);
> -    }
> +    _aligned_free(ptr);
>  }
>  
>  void qemu_anon_ram_free(void *ptr, size_t size)


According to the documentation, malloc.h should be included for
_aligned_malloc. See
https://docs.microsoft.com/de-de/cpp/c-runtime-library/reference/aligned-malloc?view=vs-2019

I'd also use g_assert instead of assert to make sure that it is not
removed by NDEBUG.

Thanks,

Stefan

On Tue, Sep 15, 2020 at 05:46:33PM -0700, Richard Henderson wrote:
> We do not need or want to be allocating page sized quanta.
> 
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Would it be safe to keep patches 2-6 applied in case this one
gets dropped or reverted for any reason?

I consider patches 2-6 bug fixes (that I'd like to merge soon),
and this one as an optimization (that could be merged later).

> ---
> Cc: Stefan Weil <sw@weilnetz.de>
> ---
>  util/oslib-win32.c | 10 +++-------
>  1 file changed, 3 insertions(+), 7 deletions(-)
> 
> diff --git a/util/oslib-win32.c b/util/oslib-win32.c
> index c654dafd93..8d838bf342 100644
> --- a/util/oslib-win32.c
> +++ b/util/oslib-win32.c
> @@ -56,10 +56,8 @@ void *qemu_try_memalign(size_t alignment, size_t size)
>  {
>      void *ptr;
>  
> -    if (!size) {
> -        abort();
> -    }
> -    ptr = VirtualAlloc(NULL, size, MEM_COMMIT, PAGE_READWRITE);
> +    assert(size != 0);
> +    ptr = _aligned_malloc(alignment, size);
>      trace_qemu_memalign(alignment, size, ptr);
>      return ptr;
>  }
> @@ -93,9 +91,7 @@ void *qemu_anon_ram_alloc(size_t size, uint64_t *align, bool shared)
>  void qemu_vfree(void *ptr)
>  {
>      trace_qemu_vfree(ptr);
> -    if (ptr) {
> -        VirtualFree(ptr, 0, MEM_RELEASE);
> -    }
> +    _aligned_free(ptr);
>  }
>  
>  void qemu_anon_ram_free(void *ptr, size_t size)
> -- 
> 2.25.1
> 
> 

-- 
Eduardo

On 9/17/20 11:43 AM, Eduardo Habkost wrote:
> On Tue, Sep 15, 2020 at 05:46:33PM -0700, Richard Henderson wrote:
>> We do not need or want to be allocating page sized quanta.
>>
>> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
> 
> Would it be safe to keep patches 2-6 applied in case this one
> gets dropped or reverted for any reason?

Yes.

Only objects using QEMU_ALIGNED() should have an __alignof__ larger than
qemu_max_align_t.

Therefore, I believe that the CPUArchState objects to be the only ones that
will go through qemu_memalign.  So we won't end up with all QOM objects being
page allocated on Windows.


r~