Re: [PATCH, RFC] adjust legacy IDE resource setting
On Wed, Feb 14, 2007 at 03:05:24PM +, Jan Beulich wrote: > The change to force legacy mode IDE channels' resources to fixed > non-zero values confuses (at least some versions of) X, because the > values reported by the kernel and those readable from PCI config space > aren't consistent anymore. Therefore, this patch arranges for the > respective BARs to also get updated if possible. > > Signed-off-by: Jan Beulich <[EMAIL PROTECTED]> > > --- linux-2.6.20/drivers/pci/probe.c 2007-02-04 19:44:54.0 +0100 > +++ 2.6.20-pci-ide-legacy/drivers/pci/probe.c 2007-02-14 15:54:58.0 > +0100 ... On Wed, Feb 14, 2007 at 04:09:36PM +, Alan wrote: > > Of course I also opened a bug against X, as I too think it's doing something > > wrong here. > > If you can add a comment about why it is done (X problem) then it looks > fine to me. This would be nice in 2.6.20-stable, too. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH, RFC] adjust legacy IDE resource setting
> The masking is done primarily to (a) calculate the correct length (from a > BAR's > perspective), as I don't want to write the BAR if its length doesn't match the > expectation, and (b) to properly report the new value in the printk. Ok I guess you have to do something like that since you can't properly encode BAR 1 and BAR 3. > Of course I also opened a bug against X, as I too think it's doing something > wrong here. If you can add a comment about why it is done (X problem) then it looks fine to me. Alan - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH, RFC] adjust legacy IDE resource setting
>>> Alan <[EMAIL PROTECTED]> 14.02.07 16:40 >>> >On Wed, 14 Feb 2007 15:05:24 + >"Jan Beulich" <[EMAIL PROTECTED]> wrote: > >> The change to force legacy mode IDE channels' resources to fixed >> non-zero values confuses (at least some versions of) X, because the >> values reported by the kernel and those readable from PCI config space >> aren't consistent anymore. Therefore, this patch arranges for the >> respective BARs to also get updated if possible. > >If X is getting confused fix X. Those BARs are *undefined* in legacy >mode. The value in them is undefined, the results that end up there if >you do write to them are undefined too. If X believes those BAR values >blindly it'll do the wrong thing in some freaky cases. > >Which specific versions of X are problematic ? The one I ran into problems with is reporting X Window System Version 6.9.0 Release Date: 21 December 2005 (used in SLES10, the specific package version is xorg-x11-6.9.0-50.14) >As to the implementation: > start and end as passed are the real I/O values so you don't need >to mask them that I can see. The masking is done primarily to (a) calculate the correct length (from a BAR's perspective), as I don't want to write the BAR if its length doesn't match the expectation, and (b) to properly report the new value in the printk. >I've no fundamental problem with writing the BAR values back to avoid >confusing some apparently broken X, but I'd like to know what X, what >circumstances and that X is also getting fixed. Of course I also opened a bug against X, as I too think it's doing something wrong here. Jan - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH, RFC] adjust legacy IDE resource setting
On Wed, 14 Feb 2007 15:05:24 + "Jan Beulich" <[EMAIL PROTECTED]> wrote: > The change to force legacy mode IDE channels' resources to fixed > non-zero values confuses (at least some versions of) X, because the > values reported by the kernel and those readable from PCI config space > aren't consistent anymore. Therefore, this patch arranges for the > respective BARs to also get updated if possible. If X is getting confused fix X. Those BARs are *undefined* in legacy mode. The value in them is undefined, the results that end up there if you do write to them are undefined too. If X believes those BAR values blindly it'll do the wrong thing in some freaky cases. Which specific versions of X are problematic ? As to the implementation: start and end as passed are the real I/O values so you don't need to mask them that I can see. I've no fundamental problem with writing the BAR values back to avoid confusing some apparently broken X, but I'd like to know what X, what circumstances and that X is also getting fixed. Alan - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/