Re: WDCTL_RST

[Brad Spencer <brad%anduin.eldar.org@localhost>]

Patrick Welche <prlw1%cam.ac.uk@localhost> writes:

> On Tue, Oct 16, 2018 at 03:08:18PM +0100, Patrick Welche wrote:
>> I've been getting by with the patch at the bottom of PR 53524 which
>> simply avoids the reset - hacky kludge...
>> 
>> Thanks to your email, hunting for the PR number, I stumbled on PR 52372.
>> I'll have to try Jaromir's patch tonight!
>
> and it does!
>
> http://www.netbsd.org/~jdolecek/ahci_reset_wait.diff
>
> from PR kern/52372 fixes PR kern/53524 for me.
>
>
> (Check that I booted the right kernel:
> # rep ahci_sata_phy_reset /netbsd
> Binary file /netbsd matches
> )
>
>
> Cheers,
>
> Patrick


I would like to report a negative when using the above patch.  I tried
the above diff and it really did not improve the situation for me.  I
still get the WDCTL_RST clear problems, except that it now appears to
try to probe both of my drives.  Without the patch, it would only manage
to notice one drive.  In a single case, where I powered everything down
and pretty much unplugged the system for a bit, did the test kernel
manage to probe one of the drives.  Avoiding the reset entirely, like my
PR did, allows the drives to be probed and work as expected.


-- 
Brad Spencer - brad%anduin.eldar.org@localhost - KC8VKS - http://anduin.eldar.org