Date: Tue, 3 Sep 2002 02:21:32 +0200 From: corex <corex@newtrinity.default-network.net> To: Jake Burkholder <jake@locore.ca> Cc: marius@alchemy.franken.de, freebsd-sparc@FreeBSD.ORG Subject: Re: Instability Message-ID: <20020903022132.B16493@newtrinity.default-network.net> In-Reply-To: <20020902194845.J67527@locore.ca>; from jake@locore.ca on Mon, Sep 02, 2002 at 07:48:45PM -0400 References: <Pine.GSO.4.44-Blink.0209021129360.2263-100000@thea.blinkenlights.nl> <20020902145424.I67527@locore.ca> <20020902215033.A16493@newtrinity.default-network.net> <20020902194845.J67527@locore.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Sep 02, 2002 at 07:48:45PM -0400, Jake Burkholder wrote: > > > > may not be limited to u60 hardware, haven't upgraded other ultrasparc > > hardware to that source-level, yet. most easiest way for me to trigger > > the panic is to run something under screen, e.g. a `cd / && find .` will > > do. let me know if i can/shall provide further information. > > Did you install screen from a package? I thought that the latest package > run (which was a while ago) was after I moved the signal trampoline into > libc, but I guess not. I have a fix which restores binary compatibility, > you can either wait for that and do a buildworld or rebuild all your > installed packages from ports. > > Sorry about that :-/ > yes, the installed screen is a package from ftp.freebsd.org as the port refused to build the time i set up that box iirc. it's still lying around and it's a .tgz and therefore a bit older hence it would be a .tbz i guess. however, running something under screen (no instant panic btw., e.g. find runs for a while and then the system panics midway) is just a way to force the panic. the first time it happend the day before yesterday (i think) after upgrading kernel+userland after a cvs update; after the final reboot i logged in from another machine via ssh and *boom* got that panic nearly instantly. same happend again today after upgrading to sources including gcc3.2. nothing from packges should have ran at that time (just screen and perl installed on that u60). if it survives the boot and the first login via ssh it seems to be stable unless i run screen... knowing nothing about fbsd-kernel-mutex-stuff are you sure the code is right ? i mean the PROC_LOCK() in line 380 in machdep.c is the only one not following a PROC_UNLOCK() in sendsig(); all the others are suceeding a PROC_UNLOCK()...just i wild guess... To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-sparc" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020903022132.B16493>