On Wed, 1 Mar 2017 16:50:57 +0100 Christoph Hellwig <hch@xxxxxx> wrote: > On Tue, Feb 28, 2017 at 10:48:45PM -0800, Stephen Hemminger wrote: > > Let me know, I can run another test and dump more data. > > Could it be that we keep the old sense buffer values around because > my commit change the way how sense buffers are handled. A while ago > I suggested this patch to fix it, so maybe you can try it: > > http://git.infradead.org/users/hch/block.git/commitdiff/148cff67b401e2229c076c0ea418712654be77e4 It appears that is already in the code I am testing in linux-next... bool scsi_normalize_sense(const u8 *sense_buffer, int sb_len, struct scsi_sense_hdr *sshdr) { memset(sshdr, 0, sizeof(struct scsi_sense_hdr)); if (!sense_buffer || !sb_len) return false; sshdr->response_code = (sense_buffer[0] & 0x7f); if (!scsi_sense_valid(sshdr))
