On 22 January 2018 at 18:02, Chuck Atkins wrote:
> Hi Emil,
>
>> Fixes: a4be2bcee2f ("swr: allow a single swr architecture to be builtin")
>
>
> It doesn't fix anything that was broken from that commit. The issues with
> error handling were already present before then,
Chuck,
Change verified on my end; check-in coming coming up soon.
George
On Jan 22, 2018, at 2:11 PM, Kyriazis, George
> wrote:
On Jan 22, 2018, at 12:02 PM, Chuck Atkins
>
On Jan 22, 2018, at 12:02 PM, Chuck Atkins
> wrote:
Hi Emil,
Fixes: a4be2bcee2f ("swr: allow a single swr architecture to be builtin")
It doesn't fix anything that was broken from that commit. The issues with
error handling were
Hi Emil,
Fixes: a4be2bcee2f ("swr: allow a single swr architecture to be builtin")
>
It doesn't fix anything that was broken from that commit. The issues with
error handling were already present before then, it's just that the changes
in a4be2bcee2f were substantial so this commit works off the
On 22 January 2018 at 17:03, Chuck Atkins wrote:
> Hi Emil,
>
>> Please include your follow-up reply/context as commit message.
>
>
> Will do.
>
>
>> > +// cleanup for failed screen creation
>> > +PUBLIC void swr_destroy_screen_internal(struct swr_screen **screen);
>>
>>
This makes the following changes to address cleanup issues:
- Error conditions now return NULL instead of calling exit()
- swr_creen is now freed upon error, rather than leak.
- Library handle from dlopen is now closed upon swr_screen destruction
v2: Added additional context in commit msg and
This makes the followingt changes to address cleanup issues:
- Error conditions now return null instead of calling exit()
- swr_creen is now freed upon error, rather than leak.
- Library handle from dlopen is now closed upon swr_screen destruction
v2: Added additional context in commit msg and
Hi Emil,
Please include your follow-up reply/context as commit message.
>
Will do.
> +// cleanup for failed screen creation
> > +PUBLIC void swr_destroy_screen_internal(struct swr_screen **screen);
>
> I'm fairly sure you don't need to make this function public. It's used
> within the same
On 22 January 2018 at 15:12, Chuck Atkins wrote:
> Signed-off-by: Chuck Atkins
Please include your follow-up reply/context as commit message.
> // arch-specific dll entry point
> PUBLIC struct pipe_screen *swr_create_screen_internal(struct
For context, without this, the library handle from dlopen never get's
closed, even under successful operation, and the swr_screen created never
get's freed on error. Also error conditions resulted in exit() rather than
NULL return.
- Chuck
On Mon, Jan 22, 2018 at 10:12 AM, Chuck Atkins
Signed-off-by: Chuck Atkins
---
src/gallium/drivers/swr/swr_loader.cpp | 100 +
src/gallium/drivers/swr/swr_public.h | 6 +-
src/gallium/drivers/swr/swr_screen.cpp | 26 +++--
src/gallium/drivers/swr/swr_screen.h | 3 +
4
Signed-off-by: Chuck Atkins
---
src/gallium/drivers/swr/swr_loader.cpp | 100 +
src/gallium/drivers/swr/swr_public.h | 6 +-
src/gallium/drivers/swr/swr_screen.cpp | 26 +++--
src/gallium/drivers/swr/swr_screen.h | 3 +
4
12 matches
Mail list logo