apelete | Hi larsc, you there ? | 16:02 |
---|---|---|
apelete | was wondering in which function should data->bytes_xfered be updated (after dma transfer finished) | 16:03 |
apelete | tried updating it in dma_complete callback, but it does not always work | 16:04 |
apelete | sometimes the pointer to mmc_data is null from within dma_complete(), which causes a kernel crash | 16:05 |
apelete | see here: http://paste.debian.net/105107/ | 16:06 |
apelete | better here: http://paste.debian.net/105108/ | 16:08 |
DocScrutinizer05 | o.O | 16:08 |
apelete | larsc: maybe there's a better place than dma_complete() to update data->bytes_xfered ? | 16:09 |
apelete | I was thinking about mmc_irq(), don't know if it's called at the end of each dma transfer | 16:10 |
DocScrutinizer05 | dma_complete sounds about logically correct place. You however may want to make sure to use the correct data source, since maybe dma itself changes pointer to mmc_data | 16:13 |
larsc | apelete: doesn't trally matter, as long as it is updated before mmc_request_done() | 16:14 |
DocScrutinizer05 | the correct data source is where dma returns number of bytes transferred | 16:15 |
DocScrutinizer05 | particularly the DMA hw might abort prematurely for whatever rwason, and in this case you want to read out whatever hw register to learn about number of bytes actually transferred | 16:18 |
DocScrutinizer05 | since this approach is also applicable for DMA corectly comleting its job, it's the way to go | 16:19 |
wpwrak | bwahaha. and so the blitzkrieg begins :) | 22:04 |
DocScrutinizer05 | huh? | 22:28 |
DocScrutinizer05 | wpwrak: please elaborate! | 22:29 |
DocScrutinizer05 | you're scaring me | 22:33 |
wpwrak | argentina vs. bosnia. first goal in the 3rd minute, amidst great confusion on the side of bosnia (it was an own goal) | 22:49 |
DocScrutinizer05 | haha | 22:50 |
wpwrak | for the rest of the first half, it was fairly unremarkable. of course, that was to be expected. with one goal in its favour, argentina feels it can rest on its laurels. | 22:51 |
wpwrak | but .. so far, so good | 22:52 |
wpwrak | in any case, it would be a miracle if argentina didn't get past the first around. in its group, there are bosnia, iran, and nigeria. of the three, bosnia is already the strongest opponent. | 22:53 |
wpwrak | but it was surprising to hear all the noise as i was still fetching some tv snack. i thought they were particularly enthusiastic to see their team on the field. didn't expect the first goal quite that soon ;-) | 22:56 |
wpwrak | 2nd half begins | 23:00 |
wpwrak | bosnia got a honorable defeat | 23:53 |
--- Mon Jun 16 2014 | 00:00 |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!