ps segfaults, kernel oopses

wapcaplet · 02-01-2004, 03:16 PM

I gotta turn off that cron job that emerge-updates my system every day... this is the second bizarre annoyance it has caused.

OK, so I just upgraded my kernel (Gentoo) to 2.4.22-gentoo-r5. Everything seems peachy, except for these at boot time:

Code:

Feb  1 14:11:37 malcolm <1>Unable to handle kernel NULL pointer dereference at virtual address 00000108
Feb  1 14:11:37 malcolm printing eip:
Feb  1 14:11:37 malcolm c01ef1b3
Feb  1 14:11:37 malcolm *pde = 00000000
Feb  1 14:11:37 malcolm Oops: 0000
Feb  1 14:11:37 malcolm CPU:    0
Feb  1 14:11:37 malcolm EIP:    0010:[<c01ef1b3>]    Tainted: P 
Feb  1 14:11:37 malcolm EFLAGS: 00210246
Feb  1 14:11:37 malcolm eax: 00000108   ebx: 00000000   ecx: 0002133b   edx: 00000000
Feb  1 14:11:37 malcolm esi: 000000c8   edi: 000002a7   ebp: 000003ff   esp: d09f9f04
Feb  1 14:11:37 malcolm ds: 0018   es: 0018   ss: 0018
Feb  1 14:11:37 malcolm Process ps (pid: 31235, stackpage=d09f9000)
(...)

Looks like they're being caused by 'ps' - in fact, every time I run 'ps' I get a segfault, and another of these kernel oopses gets appended to my system log. Tried reinstalling the procps package; no luck. Last time something like this happened, it turned out I had compiled something with the '+static' USE flag that I shouldn't have; I checked, and nothing I have installed was compiled with this flag.

Any ideas? Aside from not being able to use 'ps', this isn't too severe - just annoying. System runs fine otherwise.

wapcaplet · 02-01-2004, 03:25 PM

P.S. (no pun intended) - 'top', 'w', and 'uptime' also segfault, presumably because they call on 'ps' in some way. 'free', 'vmstat', and others that are part of the procps package seem OK.

Hmm, does 'ps' need some feature in the kernel that I may have disabled? I didn't really change anything significant in there.

wapcaplet · 02-01-2004, 04:04 PM

Nevermind. Known problem. Looks like all I gotta do is remerge gentoo-sources.