pmspcv panic on boot on this box

[Warner Losh]

> On Jul 31, 2015, at 3:32 PM, Glen Barber <gjb at FreeBSD.org> wrote:
> 
> On Fri, Jul 31, 2015 at 03:11:15PM -0600, Warner Losh wrote:
>> 
>>> On Jul 31, 2015, at 2:46 PM, Larry Rosenman <ler at lerctr.org> wrote:
>>> 
>>> On 2015-07-31 15:44, Glen Barber wrote:
>>>> On Fri, Jul 31, 2015 at 03:41:37PM -0500, Larry Rosenman wrote:
>>>>> We have a winner -- rebuilt with just this patch, and GENERIC with pmspcv,
>>>>> and we boot.
>>>> Larry, thank you so much for your willingness to test this.  It is
>>>> greatly appreciated.
>>>> Glen
>>> Absolutely my pleasure.  I appreciate ALL that the re@ and developers folks do for FreeBSD.
>>> 
>>> Best OS ever :)
>> 
>> Thanks Larry. I’ll cleanup the patch (I’d never knowing commit #if 0 w/o a good reason)
>> and get it reviewed. The good news is that we can make a fairly low-risk patch for 10.2R
>> I think, subject to the blessings of our benevolent re@ overlords...
>> 
> 
> My gut instinct is to leave the driver in 10.2 as-is for now, as I'd
> like to make sure this driver does not present more pain (and certainly
> before being re-enabled in GENERIC on stable/10), since this situation
> could have been much more of a disaster.
> 
> Personally, I'd like to keep it out of GENERIC on stable/10 for at least
> a month, while things get shaken out in head.
> 
> That said, once the issues are flushed out, re@ will be happy to issue
> an EN for 10.2 to resolve the existing (and any new) issues with this.

You’re the boss. I can understand the skepticism given how badly the current
code misunderstands newbus. This instance is quite easy to fix (I should have
something in -current in a day or two). Fortunately, the series of errors would only
cause a problem if you had two Adaptec devices in your system (and not just one).

Since it is available as a module, perhaps a note explaining there was a problem
with the auto-probing code so it was omitted from GENERIC, but if you have one of
these devices put pms_load=YES in your loader.conf file. But only if you don’t have
any other devices with the Adaptec vendor ID (since if you want to load pms, it is
because you presumably have one of these cards). The path to doom is the second
trip into probe after the first trip into probe failed.

I’d be inclined to take a middle ground and fix the driver (and therefore the module)
in 10.2R, but not put it in GENERIC until 10.3, but hey that’s just me.

Interested parties should check out https://reviews.freebsd.org/D3263 for correctness
or for testing.

Warner


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 842 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.freebsd.org/pipermail/freebsd-current/attachments/20150731/9a303cac/attachment.bin>