From mboxrd@z Thu Jan 1 00:00:00 1970 From: Borislav Petkov Date: Fri, 22 Aug 2008 21:49:14 +0000 Subject: Re: kernel unaligned accesses on IA64 in IDE Message-Id: <20080822214914.GA8609@gollum.tnic> List-Id: References: <20080819225606.GB22088@us.ibm.com> <20080822164538.GB9047@us.ibm.com> <9ea470500808221029l6bd79c62w4c06d948f962d95c@mail.gmail.com> <200808222036.00536.bzolnier@gmail.com> <57C9024A16AD2D4C97DC78E552063EA309EDE187@orsmsx505.amr.corp.intel.com> <200808222036.m7MKa6bT008897@agluck-lia64.sc.intel.com> <20080822205431.GA5968@gollum.tnic> <20080822213812.GC11781@us.ibm.com> In-Reply-To: <20080822213812.GC11781@us.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Nishanth Aravamudan Cc: "Luck, Tony" , Robin Holt , Bartlomiej Zolnierkiewicz , "linux-ia64@vger.kernel.org" , "linux-ide@vger.kernel.org" , FUJITA Tomonori On Fri, Aug 22, 2008 at 02:38:12PM -0700, Nishanth Aravamudan wrote: > On 22.08.2008 [22:54:31 +0200], Borislav Petkov wrote: > > On Fri, Aug 22, 2008 at 01:36:06PM -0700, Luck, Tony wrote: > > > > How about long instead of int. int leaves us with the possibility that > > > > something else will expect 8 byte alignment. > > > > > > How about this? > > > > > > Align __cmd to ward off kernel unaligned access consoles messages on > > > ia64 (and perhaps make an almost imperceptible performance improvement > > > on other architectures that can handle unaligned access, but do so > > > more slowly than aligned accesses). > > > > By the way, i've been wondering, this unaligned access breaking should have > > happened earlier, how did we miss that? Nish, Robin, can you see that in some > > older logs? > > I replied about this in a separate sub-thread, but you may not have been > on the direct Cc (though linux-ide should have been). > > This all began because I noticed one of my automated test machines had > not successfully booted a kernel in some time. I investigated via > bisection and found that CONFIG_PRINTK_TIME=y was broken on my IA64 box. > With Tony and Peter Z's help, that bug was resolved, but then with the > console working, I noticed these unaligned accesses. Sorry if it seemed > like, from my posting, they were a regression, I did not mean to give > that impression. > > They may have been there for some time, but I don't have any logs from > kernels before 2.6.27-rc3, because of the CONFIG_PRINTK_TIME=y breakage > (which was being set by default for this box). I could go back and try > older kernels, if you'd like, but it would take some time. Thanks for clarifying but it's not needed anymore since we know about the problem now. I was simply wondering why similar reports didn't appear earlier since the code that causes it is not that new - I guess there aren't that many ia64 machines using ide-cd from a recent kernel out there :) -- Regards/Gruss, Boris.