[Mesa-dev] [PATCH 7/8] mesa: fix assertion in _mesa_drawbuffers()

[Roland Scheidegger]

Am 11.08.2014 00:51, schrieb Brian Paul:
> On 08/08/2014 07:43 PM, Roland Scheidegger wrote:
>> Am 08.08.2014 23:20, schrieb Brian Paul:
>>> Fixes failed assertion when _mesa_update_draw_buffers() was called
>>> with GL_DRAW_BUFFER == GL_FRONT_AND_BACK.  The piglit gl30basic hit
>>> this.
>>>
>>> Cc: "10.2" <mesa-stable at lists.freedesktop.org>
>>> ---
>>>   src/mesa/main/buffers.c |    5 +++--
>>>   1 file changed, 3 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/src/mesa/main/buffers.c b/src/mesa/main/buffers.c
>>> index b13a7af..6b4fac9 100644
>>> --- a/src/mesa/main/buffers.c
>>> +++ b/src/mesa/main/buffers.c
>>> @@ -494,10 +494,11 @@ _mesa_drawbuffers(struct gl_context *ctx,
>>> GLuint n, const GLenum *buffers,
>>>      }
>>>
>>>      /*
>>> -    * If n==1, destMask[0] may have up to four bits set.
>>> +    * destMask[0] may have up to four bits set
>>> +    * (ex: glDrawBuffer(GL_FRONT_AND_BACK)).
>>>       * Otherwise, destMask[x] can only have one bit set.
>>>       */
>>> -   if (n == 1) {
>>> +   if (_mesa_bitcount(destMask[0]) > 1) {
>>>         GLuint count = 0, destMask0 = destMask[0];
>>>         while (destMask0) {
>>>            GLint bufIndex = ffs(destMask0) - 1;
>>>
>>
>> Hmm I don't understand how that could fail. Either you have a winsys
>> fbo, in which case n has to be 1, or you have a user fbo, in which case
>> you can only have the single color_buffer_x bits.
>> What am I missing here?
> 
> It happened when called from _mesa_update_draw_buffers() where n =
> MaxDrawBuffers (8) and the user had previously called
> glDrawBuffer(GL_FRONT_AND_BACK).  This could only happen during a
> MakeCurrent() operation.
> 
> -Brian
> 

Ahh I see now. MakeCurrent() will, unlike the other callers, cause all
(max) draw buffers to be updated, without a destMask, thus n = 8 but
only the first buffer really containing anything useful (the rest
shouldn't be set and their calculated masks all zero). Looks good then.
I wasn't able to trigger it for some reason, though.

Reviewed-by: Roland Scheidegger <sroland at vmware.com>