Messages in this thread | | | Date | Fri, 1 Aug 2008 00:50:04 +0200 | From | Ingo Molnar <> | Subject | Re: [patch 0/9] x86, xsave: xsave/xrstor support |
| |
* Linus Torvalds <torvalds@linux-foundation.org> wrote:
> > But thats not what I see with for ex with, w83627hf_wdt.c > > > > Its done at the module_init time. (while I thought it should be > > really done when some user level app opens the device, probably its > > poorly written to take care if the kernel panics before starting > > userland. but kernel can even die before the watchdog driver load > > aswell ;-) > > Yeah, that's a _really_ broken watchdog timer driver. There's no way > that it's correct to start the watchdog at init time, at least when > compiled in. > > It also looks to me like it's not even probing for the hardware - it's > just assuming it's there. That's scary. Am I missing something? > > It really shouldn't be activated until it's opened. And it really > shouldn't just write to ports randomly without checking that they make > sense...
there are a handful of old ISA-ish drivers that can crash randconfig kernels in various ways. [indefinite lockups, crashes, stomped-over hardware, non-working keyboard, etc.]
I mapped most of them out via many months of trial-and-error - but it would still be nice to have some separate config option to disable the known ones. CONFIG_ALLOW_NON_GENERIC or something like that - which i would unset in the randconfig runs.
( They are not CONFIG_BROKEN per se, because often it's hardware that cannot be probed in any reliable way - the driver just assumes it's there. )
Ingo
| |