[uml-user] WARNING: at mm/page_alloc.c:2386 __alloc_pages_nodemask+0x153/0x750()

[Toralf F. <tor...@gm...>]

While trying trinity under a UML I often run into this situation :


2013-03-12T21:54:41.934+01:00 trinity kernel: ------------[ cut here ]------------
2013-03-12T21:54:41.934+01:00 trinity kernel: WARNING: at mm/page_alloc.c:2386 __alloc_pages_nodemask+0x153/0x750()
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fcfc:  [<083426b8>] dump_stack+0x22/0x24
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fd14:  [<0807d11a>] warn_slowpath_common+0x5a/0x80
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fd3c:  [<0807d1e3>] warn_slowpath_null+0x23/0x30
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fd4c:  [<080d2bb3>] __alloc_pages_nodemask+0x153/0x750
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fdd8:  [<080d31d8>] __get_free_pages+0x28/0x50
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fdf0:  [<080fbe1f>] __kmalloc_track_caller+0x3f/0x180
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fe18:  [<080de616>] memdup_user+0x26/0x70
2013-03-12T21:54:41.934+01:00 trinity kernel: 3837fe34:  [<080de81e>] strndup_user+0x3e/0x60
2013-03-12T21:54:41.935+01:00 trinity kernel: 3837fe50:  [<0823c493>] sys_request_key+0x53/0x170
2013-03-12T21:54:41.935+01:00 trinity kernel: 3837feac:  [<08062a92>] handle_syscall+0x82/0xb0
2013-03-12T21:54:41.935+01:00 trinity kernel: 3837fef4:  [<08074e7d>] userspace+0x46d/0x590
2013-03-12T21:54:41.935+01:00 trinity kernel: 3837ffec:  [<0805f7cc>] fork_handler+0x6c/0x70
2013-03-12T21:54:41.935+01:00 trinity kernel: 3837fffc:  [<00000000>] 0x0
2013-03-12T21:54:41.935+01:00 trinity kernel:
2013-03-12T21:54:41.935+01:00 trinity kernel: ---[ end trace 2e631e8b4588be93 ]---
2013-03-12T21:54:41.935+01:00 trinity kernel: VFS: Warning: trinity-child0 using old stat() call. Recompile your binary.
2013-03-12T21:54:41.935+01:00 trinity kernel: VFS: Warning: trinity-child0 using old stat() call. Recompile your binary.


>From the author of trinity I was told:

	Looks like the "tried to kmalloc more than what kmalloc can do" bug.
	That shows up from time to time in various places.

	Not sure how this can happen from this trace, because it looks like
	we clamp everywhere with PAGE_SIZE.  Possibly a UML bug, dunno.


The host kernel is 3.7.10 at a 32bit stable Gentoo, the UML guest kernels are 3.8.2 and 3.9-rc2--112-g7c6baa3
from the later the config is attached.

-- 
MfG/Sincerely
Toralf Förster
pgp finger print: 7B1A 07F4 EC82 0F90 D4C2 8936 872A E508 7DB6 9DA3