Hi Thomasz,
I have extended the MFC-FIMC testcase posted by Kamil sometime ago to
sanity test the UMM patches. This test is multi-threaded(further
explanation for developers who may not have seen it yet), where thread
one parses the encoded stream and feeds into the codec IP driver(aka
MFC driver). Second thread will dequeue the buffer from MFC driver
(DMA_BUF export) and queues it into a CSC IP(aka FIMC) driver(DMA_BUF
import). Third thread dequeues the frame from FIMC driver and either
pushes it into LCD driver for display or dumps to a flat file for analysis.
MFC driver exports the fd's and FIMC driver imports and attaches it.
During FIMC QBUF (thats when the attach and map happens), it is observed
that in the function vb2_dc_map_dmabuf() call to
vb2_dc_get_contiguous_size() fails. This is because contig_size <
buf->size. contig_size is calculated from the SGT which we would have
constructed in the function vb2_dc_pages_to_sgt().
Let me know if you need more details.
Regards,
Subash
On 05/07/2012 08:41 PM, Tomasz Stanislawski wrote:
Hi Subash,
Could you provide a detailed description of a test case
that causes a failure of vb2_dc_pages_to_sgt?
Regards,
Tomasz Stanislawski
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
