MAC firmware does not load due to (commit 288d5abe usermodehelper)

MAC firmware does not load due to (commit 288d5abe usermodehelper)

[Richard Cochran]

I have some ARM IXP4xx based machines that use the two on chip MAC
ports (aka NPEs). The NPE needs a firmware in order to function. Ever
since the following commit, it is no longer possible to bring up the
interfaces during the init scripts.

    commit 288d5abec8314ae50fe6692f324b0444acae8486
    Author: Linus Torvalds <torvalds@linux-foundation.org>
    Date:   Wed Aug 3 22:03:29 2011 -1000

    Boot up with usermodehelper disabled

A warning appears on the console, shown below.  I don't know what
commit 288d5abe was trying to fix, but it sure made a regression for
me. It probably would also spoil using NFS boot, too.

Any ideas what to do about this?

Thanks,
Richard


------------[ cut here ]------------
WARNING: at /home/richard/git/linux/drivers/base/firmware_class.c:537 _request_firmware+0xc8/0x3b8()
Modules linked in:
Backtrace: 
[<c000bee0>] (dump_backtrace+0x0/0x114) from [<c000c3dc>] (dump_stack+0x18/0x1c)
[<c000c3c4>] (dump_stack+0x0/0x1c) from [<c001a778>] (warn_slowpath_common+0x58/0x70)
[<c001a720>] (warn_slowpath_common+0x0/0x70) from [<c001a7b4>] (warn_slowpath_null+0x24/0x2c)
[<c001a790>] (warn_slowpath_null+0x0/0x2c) from [<c0142710>] (_request_firmware+0xc8/0x3b8)
[<c0142648>] (_request_firmware+0x0/0x3b8) from [<c0142a20>] (request_firmware+0x20/0x28)
[<c0142a00>] (request_firmware+0x0/0x28) from [<c0012c98>] (npe_load_firmware+0x1c/0x488)
[<c0012c7c>] (npe_load_firmware+0x0/0x488) from [<c015dbe8>] (eth_open+0x44/0x6c0)
[<c015dba4>] (eth_open+0x0/0x6c0) from [<c017bf88>] (__dev_open+0xbc/0x114)
[<c017becc>] (__dev_open+0x0/0x114) from [<c017a044>] (__dev_change_flags+0x98/0x124)
[<c0179fac>] (__dev_change_flags+0x0/0x124) from [<c017b8e4>] (dev_change_flags+0x1c/0x50)
[<c017b8c8>] (dev_change_flags+0x0/0x50) from [<c01c2cd0>] (devinet_ioctl+0x2c4/0x674)
[<c01c2a0c>] (devinet_ioctl+0x0/0x674) from [<c01c395c>] (inet_ioctl+0xcc/0x104)
[<c01c3890>] (inet_ioctl+0x0/0x104) from [<c016cb3c>] (sock_ioctl+0x1f8/0x254)
[<c016c944>] (sock_ioctl+0x0/0x254) from [<c0082c28>] (vfs_ioctl+0x28/0x44)
[<c0082c00>] (vfs_ioctl+0x0/0x44) from [<c00832c8>] (do_vfs_ioctl+0x49c/0x4ec)
[<c0082e2c>] (do_vfs_ioctl+0x0/0x4ec) from [<c0083358>] (sys_ioctl+0x40/0x64)
[<c0083318>] (sys_ioctl+0x0/0x64) from [<c0009040>] (ret_fast_syscall+0x0/0x2c)
---[ end trace 6083e7c78b321b84 ]---

MAC firmware does not load due to (commit 288d5abe usermodehelper)

[Linus Torvalds]

On Wed, Sep 28, 2011 at 5:06 AM, Richard Cochran
<richardcochran@gmail.com> wrote:
>
> A warning appears on the console, shown below. ?I don't know what
> commit 288d5abe was trying to fix, but it sure made a regression for
> me. It probably would also spoil using NFS boot, too.
>
> Any ideas what to do about this?

Something seems to be calling ioctl() befor rest_init() has even
completed. Which sounds odd, but I wonder if the schedule() there
before it might have let 'init' get going.

Does something like the attached (UNTESTED) fix it?

                             Linus
-------------- next part --------------
A non-text attachment was scrubbed...
Name: patch.diff
Type: text/x-patch
Size: 678 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20110928/ab9b1404/attachment.bin>

MAC firmware does not load due to (commit 288d5abe usermodehelper)

[Richard Cochran]

On Wed, Sep 28, 2011 at 08:21:05AM -0700, Linus Torvalds wrote:
> 
> Something seems to be calling ioctl() befor rest_init() has even
> completed. Which sounds odd, but I wonder if the schedule() there
> before it might have let 'init' get going.

It does look like it. The init is busybox, and the startup script does
mount, syslogd, and then ifup, so that all can go by quickly.

> Does something like the attached (UNTESTED) fix it?

Yup, problem cured.

Thanks,
Richard