Date: Fri, 12 Dec 2003 21:05:02 -0700 From: Scott Long <scottl@freebsd.org> To: Bruce Evans <bde@zeta.org.au> Cc: freebsd-current@freebsd.org Subject: Re: atapicam broken by ata_lowlevel.c rev.1.23 Message-ID: <3FDA8FEE.4070403@freebsd.org> In-Reply-To: <20031213121725.N2797@gamplex.bde.org> References: <20031114080021.L3913@gamplex.bde.org> <20031127144834.Y77022@gamplex.bde.org> <20031127060730.GB24980@dan.emsphone.com> <0t4qw61041.wl@nohost.unb.ca> <20031213121725.N2797@gamplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Bruce Evans wrote: > On Fri, 12 Dec 2003 bremner@unb.ca wrote: > > >>At Thu, 27 Nov 2003 00:07:30 -0600, >>Dan Nelson wrote: >> >>>In the last episode (Nov 27), Bruce Evans said: >>> >>>>[Resending due to no response after 2 weeks.] >>>> >>>>Rev.1.23 of ata-lowlevel.c broke atapicam on my BP6 system as shown >>>>by the enclosed boot -v messages (the system just hangs, apparently >>>>waiting for a disk interrupt that never arrives; there seems to be no >>>>timeout). >>> >>>If it's any consolation: "me too" :) Backing out 1.23 worked for me as >>>well. The system that I saw the hang on isn't SMP. I also >>>pre-emptively patched another SMP system before I had to drive in to >>>fix it if it hung. >> >>I just wanted to report that my boot hang problems (which were "cured" >>by reverting to revision 1.22 of ata-lowlevel.c) have gone away with >>5.2-RC1; more precisely with 5.2-CURRENT of Dec. 10. > > > This was fixed (apparently without knowing about all the reports of the > bug's realized potential) in: > > % RCS file: /home/ncvs/src/sys/dev/ata/atapi-cam.c,v > % Working file: atapi-cam.c > % head: 1.29 > % ... > % ---------------------------- > % revision 1.29 > % date: 2003/12/05 01:02:46; author: scottl; state: Exp; lines: +1 -1 > % Fix a potential problem with atapi-cam where an incorrect flag is passed > % into the ata queueing layer. > % > % Approved by: re > % ---------------------------- > > Bruce > Oh, I knew about the reports, but I couldn't find a direct link between the bug that I was fixing (merely via code inspection) and the reported problems. Since I couldn't prove the link, I didn't want to get everyone's hopes up. It's good to see that my suspicion of it fixing the problems was correct =-) Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3FDA8FEE.4070403>