On Monday, September 18, 2006 8:56 PM, James Bottomley wrote: go. > > Erm, but surely, since reset->dv is fairly standard, this must need a > generic fix. > > Can you check this out. I think it does the right thing gating > spi_dv_device() with a new flag. The pending flag is more used to > manage scheduled domain validations. The original idea was > that resets > would trigger scheduled dv that would use this flag correctly (and not > be in the direct path of DV). For some reason the mptspi seems to > manage to entangle DV with the reset inline. > Will the dv thread that issed the host reset, wait for the lld host reset to complete itself before going forward? If not, dv will repeat itself. This issue was reported by the folks at OSDL. I've asked Mark Haverkamp to try this patch, and waiting for results. I provided a different patch to Mark last Friday, that does basically the same as your patch does, except from the lld. That patch failed. He has a full rack of disks, and the failing disk was at ID8. By the time dv completed on ID=0-7, the dv_pending flag was cleared on I8, and started DV again, that again would issue host_reset. Eric - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
