Re: MFS still hosed
Can somebody please try if this fixes MFS ?
Poul-Henning
Index: ufs/mfs/mfs_vfsops.c
===
RCS file: /home/ncvs/src/sys/ufs/mfs/mfs_vfsops.c,v
retrieving revision 1.63
diff -u -r1.63 mfs_vfsops.c
--- mfs_vfsops.c1999/05/14 20:40:23 1.63
+++ mfs_vfsops.c1999/05/23 09:22:20
@@ -62,6 +62,10 @@
MALLOC_DEFINE(M_MFSNODE, MFS node, MFS vnode private part);
+/* XXX: this is bogus, should be (NUMCDEV-1) or use dynamic allocation */
+#define CDEV_MAJOR 255
+#define BDEV_MAJOR 255
+
#ifdef MFS_ROOT
static caddr_t mfs_rootbase; /* address of mini-root in kernel virtual
memory */
static u_long mfs_rootsize; /* size of mini-root in bytes */
@@ -462,8 +466,6 @@
mfs_init(vfsp)
struct vfsconf *vfsp;
{
- dev_t dev = NODEV;
- cdevsw_add(dev, mfs_cdevsw, NULL);
- cdevsw_add_generic(255, major(dev), mfs_cdevsw);
+ cdevsw_add_generic(BDEV_MAJOR, CDEV_MAJOR, mfs_cdevsw);
return (0);
}
--
Poul-Henning Kamp FreeBSD coreteam member
p...@freebsd.org Real hackers run -current on their laptop.
FreeBSD -- It will take a long time before progress goes too far!
To Unsubscribe: send mail to majord...@freebsd.org
with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Sun, 23 May 1999 11:32:27 +0200, Poul-Henning Kamp wrote: Can somebody please try if this fixes MFS ? Do you want your patch applied alongside Luoqi Chen's patch to kern_conf.c from last week? I've been able to mount_mfs without problems since then. Ciao, Sheldon/ To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
In message 4478.927463...@axl.noc.iafrica.com, Sheldon Hearn writes: On Sun, 23 May 1999 11:32:27 +0200, Poul-Henning Kamp wrote: Can somebody please try if this fixes MFS ? Do you want your patch applied alongside Luoqi Chen's patch to kern_conf.c from last week? I've been able to mount_mfs without problems since then. Yes. The problem here is to get MFS-rootfs to work. -- Poul-Henning Kamp FreeBSD coreteam member p...@freebsd.org Real hackers run -current on their laptop. FreeBSD -- It will take a long time before progress goes too far! To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Wed, 19 May 1999 12:04:54 +1000, Bruce Evans wrote: dumping to dev (0, 131089), offset 524288 dump device bad The dev_t changes obfuscated it by printing it in %d format instead of as part of the dev number in [0x]%x format. So you reckon that whatever problem is making it imp[ossible for me to take dumps, it was present before the dev_t changes and I should be looking elsewhere? Ciao, Sheldon. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
dumping to dev (0, 131089), offset 524288 dump device bad The dev_t changes obfuscated it by printing it in %d format instead of as part of the dev number in [0x]%x format. So you reckon that whatever problem is making it imp[ossible for me to take dumps, it was present before the dev_t changes and I should be looking elsewhere? dump device bad is printed for d_dump() returning ENXIO, which is fairly unambiguous. The new wd driver's d_dump would return ENODEV, so you must be using the old wd driver. The old wd_driver's d_dump only returns ENXIO when the drive doesn't exist or has never been opened or is not labeled. Bruce To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Wed, 19 May 1999 17:40:50 +1000, Bruce Evans wrote: so you must be using the old wd driver. The old wd_driver's d_dump only returns ENXIO when the drive doesn't exist or has never been opened or is not labeled. I'm using the old wd driver (controller wdc in CURRENT). The disk is labeled and I assume it's opened by swapon when it's configured as a swap device. So I'm at a loss as to why I get ENXIO (I did look at that part of the source, but thought it best to avoid talking in source terms, since I don't fully understand the meanings of the errors). Ciao, Sheldon. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Mon, 17 May 1999 13:02:29 -0400, Luoqi Chen wrote: Are you sure you have the latest -current? I committed a fix Friday night. Hi Luoqi, I remade world and kernel (config -g -r) with yesterday's HEAD and still get a panic mounting MFS. My kernel config includes MFS but not MFS_ROOT. Shout if you want a trace. I haven't provided one here, since I'll need to copy it down by hand because I'm using the ATA* driver that's still under development and doesn't support crashdumps. Ciao, Sheldon. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Tue, 18 May 1999 08:53:49 +0200, Sheldon Hearn wrote: I remade world and kernel (config -g -r) with yesterday's HEAD and still get a panic mounting MFS. My kernel config includes MFS but not MFS_ROOT. Shout if you want a trace. I haven't provided one here, since I'll need to copy it down by hand because I'm using the ATA* driver that's still under development and doesn't support crashdumps. I think I know why a lot of people _aren't_ seeing this panic: it doesn't show up if you #undef DEVT_FASCIST in kern/kern_conf.c . I found this by accident because I can't get a crash dump using the wd driver. I get dumping to dev (0, 131089), offset 524288 dump device bad Suspecting this whole major/minor/dev_t business, I #undef'd DEVT_FASCIST and all of a sudden I can't reproduce the panic to test dumps. ;-) I'm not whining, I just want to make sure that it's a known issue, not something that's completely resolved. Ciao, Sheldon. To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
I think I know why a lot of people _aren't_ seeing this panic: it doesn't show up if you #undef DEVT_FASCIST in kern/kern_conf.c . #undef DEVT_FASICIST weakens tthe error detection. I found this by accident because I can't get a crash dump using the wd driver. I get dumping to dev (0, 131089), offset 524288 dump device bad 131089 is the ordinary minor number 0x20011 (slice 2, unit 2, partition 1) (for wd2s1b). The dev_t changes obfuscated it by printing it in %d format instead of as part of the dev number in [0x]%x format. They should have used %#x format. Bruce To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
With todays -current, mounting /tmp using swap /tmp mfs rw,nosuid,nodev,-s=32768 0 0 yields a Are you sure you have the latest -current? I committed a fix Friday night. -lq To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Mon, May 17, 1999 at 01:02:29PM -0400, Luoqi Chen wrote: Are you sure you have the latest -current? I committed a fix Friday night. I'm quite positive, because I experimented with a fresh -current/kernel yesterday. Changing NUMCDEV from 256 to 255 in kern_conf.c fixes the problem though. I'm not sure which fix you're alluding to... -- Jos Backus _/ _/_/_/ Reliability means never _/ _/ _/ having to say you're sorry. _/ _/_/_/ -- D. J. Bernstein _/ _/ _/_/ jos.bac...@nl.origin-it.com _/_/ _/_/_/ use Std::Disclaimer; To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Mon, May 17, 1999 at 03:37:36PM -0400, Luoqi Chen wrote: If you don't have `options MFS' in your config file, you will need to recompile the mfs kld module. I know, I did that this evening (did a make world/kernel rebuild around 8pm CEST): -r-xr-xr-x 1 root wheel 12874 May 17 21:20 mfs.ko* Still crashes without the NUMCDEV adjustment though... -- Jos Backus _/ _/_/_/ Reliability means never _/ _/ _/ having to say you're sorry. _/ _/_/_/ -- D. J. Bernstein _/ _/ _/_/ jos.bac...@nl.origin-it.com _/_/ _/_/_/ use Std::Disclaimer; To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
MFS still hosed
With todays -current, mounting /tmp using swap /tmp mfs rw,nosuid,nodev,-s=32768 0 0 yields a Fatal trap 12: page fault while in kernel mode fault virtual address= 0x9d203590 fault code = supervisor read, page not present instruction pointer = 0x8:0xc016f30c stack pointer= 0x10:0xc5aa2d70 frame pointer= 0x10:0xc5aa2d9c code segment = base 0x0, limit 0x, type 0x1b = DPL 0, pres 1, def32 1, gran 1 processor eflags = interrupt enabled, resume, IOPL = 0 current process = 362 (mount_mfs) Stopped at checkalias+0x13c: movl cdevsw(%eax),%eax %eax contains 0xdcfba6d0 This line in checkalias() in vfs_subr.c seems to be the culprit: 1488: 56 pushl %esi 1489: e8 fc ff ff ff call 148a checkalias+0x12e 148e: 8b 04 85 00 00 movl 0x0(,%eax,4),%eax 1493: 00 00 1495: c1 e0 02shll $0x2,%eax = 1498: 8b 80 00 00 00 movl 0x0(%eax),%eax 149d: 00 149e: 89 45 e4movl %eax,0xffe4(%ebp) (I'm not sure yet which source line this section corresponds to :-) Anyone else seen this? -- Jos Backus _/ _/_/_/ Reliability means never _/ _/ _/ having to say you're sorry. _/ _/_/_/ -- D. J. Bernstein _/ _/ _/_/ jos.bac...@nl.origin-it.com _/_/ _/_/_/ use Std::Disclaimer; To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Sun, 16 May 1999, Jos Backus wrote: With todays -current, mounting /tmp using swap /tmp mfs rw,nosuid,nodev,-s=32768 0 0 yields a Fatal trap 12: page fault while in kernel mode fault virtual address= 0x9d203590 I've been getting this for a week now. :( Luoqui Chen suggested bumping NUMCDEV in src/sys/kern/kern_conf.c to 255 to get around it, and a couple of people had success, including myself. I've filed PR kern/11737 with the fix. Doug White Internet: dwh...@resnet.uoregon.edu| FreeBSD: The Power to Serve http://gladstone.uoregon.edu/~dwhite| www.freebsd.org To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
Re: MFS still hosed
On Sun, May 16, 1999 at 01:36:44PM -0700, Doug White wrote: I've been getting this for a week now. :( Luoqui Chen suggested bumping NUMCDEV in src/sys/kern/kern_conf.c to 255 to get around it, and a couple of people had success, including myself. I'll do that then. Thanks! Groetjes, -- Jos Backus _/ _/_/_/ Reliability means never _/ _/ _/ having to say you're sorry. _/ _/_/_/ -- D. J. Bernstein _/ _/ _/_/ jos.bac...@nl.origin-it.com _/_/ _/_/_/ use Std::Disclaimer; To Unsubscribe: send mail to majord...@freebsd.org with unsubscribe freebsd-current in the body of the message
