On 09.09.2015 18:15, Robert Baldyga wrote: > On 09/09/2015 01:47 AM, Krzysztof Kozlowski wrote: >> On 08.09.2015 18:38, Robert Baldyga wrote: >>> Hello, >>> >>> This patch set contains three patches: two minor fixes and one quite >>> important bugfix enabling UART status handling in DMA mode, which was >>> missing so far. It enables, among others, 'break' contition handling, >>> which is necessary if we want to use Magic SysRq. So this patch fixes >>> Magic SysRq handling for serial consoles using UART in DMA mode, >> >> Is it even worth using UART for serial console in DMA mode? How many >> benefits it brings? Anyone measured real data? >> >> The patch which enabled DMA for all serial ports on Exynos4 uncovered >> some issues. That is one benefit... but was it worth it? > > Well, UART in DMA mode should work at least as good as in PIO, so there > is nothing wrong in having serial console configured in DMA. Sure, it > rather barely reduces number of interrupts while used as typical serial > console, but in some cases (eg. using zmodem) it can be more profitable. That's some reason... and mentioned earlier testing of DMA which uncovered some interesting issues. :) > > To be honest, I have no strong feeling about that. I've decided to > utilize new possibility opened by introducing DMA mode in serial driver, > so I've send patch enabling it for each port on Exynos4. But if it > turned out to be unneeded I will not insist on it. I was just thinking loud... I don't have any data neither so let's go ahead with serial DMA. Best regards, Krzysztof -- To unsubscribe from this list: send the line "unsubscribe linux-serial" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
