Subject: Re: Please audit pool use in your code!
To: David Brownlee <abs@NetBSD.org>
From: Matthew Orgass <darkstar@city-net.com>
List: tech-kern
Date: 07/31/2006 01:46:57
On 2006-07-28 abs@NetBSD.org wrote:

>  	Just had this on a box which had until this point been running
>  	NetBSD-3 without any problems for many months. Only change was
>  	it now had a USB printer attached (previously no UDB devices).
>
>  	It _did_ fill up the disk, though its done that before
>  	occasionally without problems:
>
>  	Is there a PR open to which it would be useful appending this?
>
> uid 2015, pid 15953, command smbd, on /home: file system full
> uvm_fault(0xbe1a17ec, 0, 0, 1) -> 0xe
> fatal page fault in supervisor mode
> trap type 6 code 0 eip a053e92b cs 8 eflags 10246 cr2 48 ilevel 5
> panic: trap
> Begin traceback...
> trap() at netbsd:trap+0x171
> --- trap (number 6) ---
> usb_transfer_complete(a24c7b00,16,a06c317f,0,0) at
> netbsd:usb_transfer_complete+
> 0x16b
> uhci_abort_xfer(a24c7b00,6,bd99fcec,a053e7a9,a24c7b00) at
> netbsd:uhci_abort_xfer
> +0x1a4
> uhci_device_bulk_abort(a24c7b00,0,bd99fd0c,a053e42c,a20f2280) at
> netbsd:uhci_dev
> ice_bulk_abort+0x19
> usbd_ar_pipe(a20f2280,a1ccb400,0,a1d73000,a22c6500) at
> netbsd:usbd_ar_pipe+0x29
> usbd_abort_pipe(a20f2280,2,bd99fd5c,a03ad2df,3900) at
> netbsd:usbd_abort_pipe+0x2
> c
> ulptclose(3900,2,2000,bdb9d4cc,40500) at netbsd:ulptclose+0x6e
> spec_close(bd99fdb4,10002,c2503dd8,0,a07a2a80) at netbsd:spec_close+0xbf
> ufsspec_close(bd99fdb4,bd7d3ea0,bd99fddc,a03a5aee,a05d3bc0) at
> netbsd:ufsspec_cl
> ose+0x171
> VOP_CLOSE(c2503dd8,2,bd770ccc,bdb9d4cc,bde07ee0) at netbsd:VOP_CLOSE+0x36
> vn_close(c2503dd8,2,bd770ccc,bdb9d4cc,a25cfaa8) at netbsd:vn_close+0x44
> vn_closefile(bde07ee0,bdb9d4cc,2,bd99fe44,20) at netbsd:vn_closefile+0x29
> closef(bde07ee0,bdb9d4cc,bdd68c38,bdb9d4cc,0) at netbsd:closef+0xd5
> fdfree(bdb9d4cc,0,bd99ff3c,a03a0cd3,bd99fed4) at netbsd:fdfree+0x5f
> exit1(be0af6bc,0,bdb9d4cc,be0af6bc,0) at netbsd:exit1+0x1af
> sys_exit(be0af6bc,bd99ff64,bd99ff5c,bdb9d4cc,1) at netbsd:sys_exit+0x29
> syscall_plain() at netbsd:syscall_plain+0x85
> --- syscall (number 1) ---
> 0x9dbc240f:

  It looks like this is kern/33250.  Rev 1.188 of uhci.c may have fixed
the problem and is not in the netbsd-3 branch.

Matthew Orgass
darkstar@city-net.com