From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp.osdl.org (smtp.osdl.org [65.172.181.4]) (using TLSv1 with cipher EDH-RSA-DES-CBC3-SHA (168/168 bits)) (Client CN "smtp.osdl.org", Issuer "OSDL Hostmaster" (not verified)) by ozlabs.org (Postfix) with ESMTP id C65ED67A3E for ; Fri, 15 Sep 2006 05:14:19 +1000 (EST) Date: Thu, 14 Sep 2006 12:14:04 -0700 (PDT) From: Linus Torvalds To: Paul Mackerras Subject: Re: Please pull powerpc.git 'merge' branch In-Reply-To: <17671.64971.450232.870486@cargo.ozlabs.ibm.com> Message-ID: References: <17671.64971.450232.870486@cargo.ozlabs.ibm.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: linuxppc-dev@ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Wed, 13 Sep 2006, Paul Mackerras wrote: > > git pull \ > git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc.git merge Hmm. I'm not at all sure this pull caused the problem, and it may have started long before, but right now I have problems printing on my G5 machine. I'll get around to trying to bisect it at some point, but before I even start, I should probably check if anybody recognizes this immediately. I blamed cupsd at first, because the problem ends up being that it just hangs in running state, and I just assumed it was broken user space as usual. But straceing the dang thing shows something strange, like this: ... time(NULL) = 1024 <0.000009> select(1024, [7 8 9 10 11 12 13 14], [], NULL, {1, 0}) = 0 (Timeout) <0.000006> time(NULL) = 1024 <0.000006> select(1024, [7 8 9 10 11 12 13 14], [], NULL, {1, 0}) = 0 (Timeout) <0.000004> time(NULL) = 1024 <0.000007> select(1024, [7 8 9 10 11 12 13 14], [], NULL, {1, 0}) = 0 (Timeout) <0.000006> time(NULL) = 1024 <0.000007> select(1024, [7 8 9 10 11 12 13 14], [], NULL, {1, 0}) = 0 (Timeout) <0.000005> ... which actually looks like something is broken (it says timeout in 5 microseconds, even though the timeout structure is shown as being one second). Now, it's entirely possible that the _broken_ thing is actually strrace itself (that return value from "time()" looks a bit strange too), and that it's some tracing problem with rinning 32-bit programs on a 64-bit kernel. But I thought I'd ask if somebody recognizes this before I start wasting any more time on it.. Linus