Date: Mon, 16 Nov 1998 09:50:48 -0500 (EST) From: Luoqi Chen <luoqi@watermarkgroup.com> To: FreeBSD-current@FreeBSD.ORG, grog@lemis.com, luoqi@watermarkgroup.com Subject: Re: Today's -CURRENT dies with "ffs_blkfree: bad size" Message-ID: <199811161450.JAA16159@lor.watermarkgroup.com>
next in thread | raw e-mail | index | archive | help
> On Sunday, 15 November 1998 at 11:11:37 -0500, Luoqi Chen wrote: > >> I've just upgraded my kernel to a sup of about 9 hours ago, and have > >> had a number of panics. The first ones happened in X, but I finally > >> got a dump. Here's a trace: > >> > >> #0 boot (howto=262400) at ../../kern/kern_shutdown.c:285 > >> 285 dumppcb.pcb_cr3 = rcr3(); > >> (kgdb) bt > >> #0 boot (howto=262400) at ../../kern/kern_shutdown.c:285 > >> #1 0xf013e1e5 in panic (fmt=0xf01cc3e1 "ffs_blkfree: bad size") > >> at ../../kern/kern_shutdown.c:453 > >> #2 0xf01cc53e in ffs_blkfree (ip=0xf1283400, bno=4, size=8192) > > ^^^^^ > > It could be caused by our delayed bmap binding. Check to see if > > bp->b_blkno == bp->b_lblkno for this block. > > bp isn't initialized at this point. It gets initialized by a call to > bread() later in the function. > I meant buffers in the list buflist->bs_children[] in ffs_reallocblks(). They were setup earlier in cluster_collectbufs() by bread(). On FreeBSD, bread() does not call VOP_BMAP() to translate logical block number into physical one if the data is in vm cache. -lq > Kirk's currently looking at this, so I'll leave him to report. > > Greg > -- > See complete headers for address, home page and phone numbers > finger grog@lemis.com for PGP public key > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199811161450.JAA16159>