Messages in this thread | | | Date | Wed, 08 May 2013 18:21:11 +0200 | From | Takashi Iwai <> | Subject | Re: [PATCH 1/3] firmware: Avoid superfluous usermodehelper lock |
| |
At Wed, 8 May 2013 23:52:02 +0800, Ming Lei wrote: > > On Wed, May 8, 2013 at 2:56 PM, Takashi Iwai <tiwai@suse.de> wrote: > > When a firmware file can be loaded directly, there is no good reason > > to lock usermodehelper. It's needed only when the direct fw load > > fails and falls back to the user-mode helper. > > I remembered that we discussed the problem before, :-) > > Some crazy drivers might call request_firmware inside resume callback > (for example, USB devices might be rebind in resume), with > usermodehelper_read_lock, we can find the mistake easily and log it. > > I am wondering if it is good to remove the usermodehelper lock. > > Could you let us know any benefit to do it?
Well, the question is whether usermodehelper lock is really an appropriate stuff for *checking* the availability of direct fs access. I find it doesn't fit well any longer, in the situation where no actual user-space call is needed. Though, I'm not quite sure which lock or flag can be used instead...
Takashi
| |