On Sun, 7 Oct 2018, Klaus Kusche wrote:

> Hello,
> 
> On 03/10/2018 16:02, Alan Stern wrote:
> > Well, what happens if you add only US_FL_BROKEN_FUA without
> > US_FL_IGNORE_UAS?  Does it work?
> 
> Tried all four combinations with gentoo 4.18.12.
> Just US_FL_IGNORE_UAS fails with the error message in the original bug report.
> 
> The other three combinations (both flags removed, both flags set,
> just US_FL_BROKEN_FUA set) work, at least for a quick test with ext4.
> 
> My guess:
> Without US_FL_IGNORE_UAS (i.e. when using UAS),
> the FUA command triggering the error is not used at all,
> so US_FL_BROKEN_FUA doesn't make a difference.

That may or may not be correct.  A usbmon trace would tell us.

> > For that matter, does your kernel config include the uas driver?
> 
> Yes, it does.
> 
> 
> I think we have to wait for sel...@redhat.com .
> He could tell why he added US_FL_IGNORE_UAS,
> if it's really needed, and why it was not included in mainline.

Good point.

> However, as US_FL_BROKEN_FUA is set for JMS567,
> it is perhaps a good idea to set it for JMS566, too,
> independent of US_FL_IGNORE_UAS (which is not set for JMS567).

Agreed.  Especially since we know that FUA does trigger a problem when
UAS isn't used.

Alan Stern

Reply via email to