Hi Guennadi, Em Thu, 6 Aug 2009 01:28:17 +0200 (CEST) Guennadi Liakhovetski <g.liakhovetski@xxxxxx> escreveu: > Hi Mauro, > > I posted this fix and haven't got any objections (apart from unused > variables, which are removed in this version), but - please have a look at > it before pulling. In mx3_camera.c::mx3_videobuf_queue(), which is the > driver's .buf_queue() method I temporarily drop the spinlock and reenable > interrupts to call a possible sleeping function, and then re-lock. Is this > acceptable? If not - the IPU IDMAC DMA driver would have to be changed. If > it is - please, pull. This depends on how your driver works. The spinlock is important to put a memory barrier protecting videobuf queue data that can be used inside the IRQ routine and via videobuf ioctls. provided that the routines you're calling at videobuf_queue don't touch at the videobuf queue, I don't see any trouble with this strategy. Your code seems to work. > Also notice, this is a fix for 2.6.31, and I forgot > to put the "Priority" tag in it, sorry. Ok, I'll re-tag it on my queue. Cheers, Mauro -- 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
