Re: __check_region in ide code?

Alan Cox (alan@lxorguk.ukuu.org.uk)
06 Jun 2003 16:23:26 +0100


On Gwe, 2003-06-06 at 09:56, Bartlomiej Zolnierkiewicz wrote:
> > There's nothing inherently *wrong* with check_region, it's
> > just deprecated to trap the old (now racy) idiom of "if
> > (check_region(xx)) reserve_region(xx)". There's no reason not to
> > introduce a probe_region if IDE really wants it.
>
> And ide-probe.c does exactly this racy stuff.
>
> I did patch to convert it to request_region() some time ago,
> I just need to double check it and submit.

request_region at that point doesn't actually help you. For PIO devices
its too late if you are handling PCMCIA, for PCI devices its too late
because you want to own the PCI device properly, for MMIO its completely
broken (all the mem region stuff in 2.5)

The only way I can see to fix it properly is to provide ide helpers
for resource allocation that are used by the drivers when needed.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/