For those new to this bug report, the interesting part starts here: https://bugzilla.kernel.org/show_bug.cgi?id=89511#c19 In short, these are USB drive adapters that doesn't support the SYNCHRONIZE CACHE command but claim to be write-back. (In at least one case, it initially claims to be write-through but then changes to write-back following a reset, or something like that.) On Thu, 4 Jun 2015 bugzilla-daemon@xxxxxxxxxxxxxxxxxxx wrote: > --- Comment #50 from Markus Rathgeb <maggu2810@xxxxxxxxx> --- > I ran in the same trouble with an old USB case for 2.5'' IDE hard discs. > > The drive cache_type is set to "write back" if it plugged in. > The drive is working using if I set the cache type to "temporary write > through". > > So, I think this is the exact same problem discussed here. > > At the moment I tested the workaround using v3.18.2 (gentoo-sources). > Without the workaround the problem also exists using the Kernel v4.0.4 (used > gentoo-sources, but I think that should not be related to the problem at all). > > Is there anything I / we can do, to solve this problem at all in the kernel > space (e.g. if SYNCHRONIZE CACHE command is not working, fallback to write > through)? This is up to the SCSI developers. The real question is what should happen when a drive really does have a write cache but doesn't recognize the SYNCHRONIZE CACHE command. I don't know what the kernel can do in this case -- maybe just log a one-time error message and turn off sdkp->WCE? Alan Stern -- 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
