+ Avri, Clément (due to recent related discussions) On 21 November 2018 at 00:08, Wolfram Sang <wsa+rene...@sang-engineering.com> wrote: > On Renesas R-Car SDHI hardware, we sometimes had timeouts accessing the RPMB. > This is because AutoCMD23/12 features needs a properly filled sbc to work > correctly. But RPMB sends an individual CMD23. I could have fixed the driver > but after some research concluded that fixing the core seems the proper thing > to do. I also added some sanity checking while here. Please let me know what > you think. > > Tested on a R-Car M3-N. No timeouts showed up anymore. I'll try to improve the > testing, though, to check if I can make the occasional timeouts from before > more reproducible. And then confirm that this patchset improves the situation > :) > > Wolfram Sang (3): > mmc: core: validate user input for RPMB block count > mmc: core: use mrq->sbc when sending CMD23 for RPMB > mmc: core: remove obsolete mmc_set_blockcount() function > > drivers/mmc/core/block.c | 14 +++++++++----- > drivers/mmc/core/core.c | 14 -------------- > drivers/mmc/core/core.h | 2 -- > 3 files changed, 9 insertions(+), 21 deletions(-) > > -- > 2.11.0 >
These changes makes perfect sense to me! I give it a day or two to allow people to comment/test, then I will queue them up. Perhaps we should add a suggested by tag from Clément for patch2, as he kind of suggested this change already [1]. Kind regards Uffe [1] https://patchwork.kernel.org/patch/10645847/
