On Fri, Apr 11, 2014 at 12:47 AM, Andrew Ruder andy@aeruder.net wrote:
On Fri, Apr 11, 2014 at 12:33:45AM +0530, Jagan Teki wrote:
It would be great if you mentioned issue scenario for status poll case drivers/mtd/spi/sf_ops.c: ret = spi_xfer(spi, 8, NULL, &status, 0);
OK - means issue only with soft_spi.c is it?
Yes, and a couple other drivers.
Can you share the issue log or typical use case scenario w.r.t soft_spi.c, I need to understand how this got resolved with your change.
Yes, you actually posted one such case that will not work "correctly" with soft_spi. In the line of code you posted above, soft_spi will actually perform a read from address 0x0. In most cases, the read side isn't a huge deal, but on the write side it can cause all kinds of surprises.
I understand you assigned '0' when dout is NULL and you took the buf only when din is !NULL.
Yes, just handling NULL case to be how most drivers are handling it and how apparently most users of the spi modules (like mtd/spi/sf_ops) are clearly expecting it to work.
I saw some check-patch issues, can you fix - plan to apply.
thanks!