Messages in this thread | | | Date | Tue, 3 Jan 2012 00:00:16 +0000 | From | Alan Cox <> | Subject | Re: loading firmware while usermodehelper disabled. |
| |
On Mon, 2 Jan 2012 13:27:03 -0800 Linus Torvalds <torvalds@linux-foundation.org> wrote:
> On Mon, Jan 2, 2012 at 1:19 PM, Matthew Garrett <mjg@redhat.com> wrote: > > On Mon, Jan 02, 2012 at 12:48:48PM -0800, Linus Torvalds wrote: > > > >> Why are you guys making it any more complicated than that? > > > > Because it's inadequate. You can't guarantee that we ever loaded > > firmware. > > If we didn't load the firmware before the suspend, then the resume > function of a device sure as hell had better not load it at resume > time either.
Get a clue. It is very common that devices keep firmware over OS warm boots, particularly if onboard. "Lesser" beings than yourself consider it a feature of good hardware design for fast booting and it works happily in "lesser" OS products such Windows.
If you'd like a typical worked example the Fujitsu Q550 has a 3G modem on it. So in the following cases we don't load firmware at boot but we do need to after hibernate/resume
Boot windows, reboot to Linux Boot Linux, reboot to Linux (first loads, second doesn't)
Break that and you have a regression and Linus says regressions are wrong and we don't do them.
> And don't make the stupid argument that we don't know. That's just > inane. Either the driver loads the firmware at startup, or it doesn't.
Wrong.
> Fix the 99%. Screw the crazy shit, don't even bother worrying about it > until *after* the 99% is fixed.
This is the 99%. This is PC hardware, the 99% is crazy shit.
Alan
| |