On 19/01/2019 20:47, Marc Gonzalez wrote: > On 19/01/2019 10:56, Christoph Hellwig wrote: > >> On Jan 18, 2019 at 10:48:15AM -0700, Jens Axboe wrote: >> >>> It's UFS that totally buggy, if you look at its queuecommand, it does: >>> >>> if (!down_read_trylock(&hba->clk_scaling_lock)) >>> return SCSI_MLQUEUE_HOST_BUSY; >>> >>> UFS either needs to get fixed up, or we'll want a way to do something like >>> the below. >> >> I think the right answer is to just revert the offending patch instead >> of papering over it in the SCSI code. > > [ Adjusting recipients list ] > > Full thread, for new recipients: > https://www.spinics.net/lists/linux-fsdevel/msg138601.html > > Christoph, do you mean a3cd5ec55f6c7 ? And by the way, can someone explain these two oddities: 1) the system only hangs if we try to read more than 3.8 GB (starting point is irrelevant) 2) the system does not hang if we use iflag=direct Regards.
