[prev in list] [next in list] [prev in thread] [next in thread] 

List:       opensolaris-networking-discuss
Subject:    Re: [networking-discuss] dladm hangs and is untouchable
From:       Peter Memishian <peter.memishian () sun ! com>
Date:       2008-12-17 18:24:50
Message-ID: 18761.17394.499358.136370 () gargle ! gargle ! HOWL
[Download RAW message or body]


 > Cathy Zhou wrote:
 >  > It could be any of the following two bugs. If you can share the crash
 >  > dump, I can certainly take a look.
 >  >
 >  > 6668816 6738489.
 > 
 > Speaking of 6668816, it doesn't seem to be fixed yet (at least
 > according to the bugreport).  I cringe in fear every time I replace a
 > NIC with the same model, but newer PCI revision or move NICs around
 > because of this bug.
 > 
 > Could you *please* at least limit the potential damage by making
 > softmac_hold_device(): not call mod_hash_find() in an infinite loop?
 > Infinite loops in the kernel are really, really bad.  If a temporary
 > mod_hash_find() failure is rare, as the comment suggests, then
 > limiting the number of times you call mod_hash_find() seems
 > reasonable.  Especially if you insert a small sleep delay between
 > calls to mod_hash_find().

I'm uncomfortable with the precedent this sets.  At some level, the
software we build is predicated on its foundation working reliably, and
it's a slippery slope down to the point where a wide variety of primitives
are treated as suspect, with commensurate complexity and codepaths that
are rarely exercised.  6668816 is a P2 bug which can affect more than just
softmac.  It should be fixed, not routed around.

--
meem
_______________________________________________
networking-discuss mailing list
networking-discuss@opensolaris.org
[prev in list] [next in list] [prev in thread] [next in thread]