On 30-09-08 20:13, Linus Torvalds wrote:
On Tue, 30 Sep 2008, Linus Torvalds wrote:
Hmm? Bjorn? Everything that is true about "pnp_system_init" should be
equally true abote pnpacpi_init and pnpbios_init, no?
Side note: unlike the more radical "move all arch driver initcalls
earlier", this one actually works for me. But it does cause
pnp_system_init() to actually run before pnpacpi_init and pnpbios_init,
due to the link order within PnP. That seems ok, since the system/acpi/pnp
drivers should all be just different versions of PnP drivers, but I
thought I'd mention it. Maybe there is some PnP internal reason why
pnp_system_init should run after the pnp_acpi/bios_init functions.
(And if so, then just changing the order in drivers/pnp/Makefile would fix
it up again).
Bjorn, is there any reason why this isn't the right thign to do?
Rene - I have this suspicion that just doing this part should already fix
your issues with _no_ other patch, since now anything PnP does is after
all the PCI setup code anyway. So does your sound card (or whatever it
was) work by just doing the subsys_initcall -> fs_initcall change in PnP?
I'll have to wait for the message to arrive here or in a web-archive to
answer that (...) but please note that I am already fine with the
original code; it is the fix for my own issue that's IN mainline (ie,
not only check for MEM resource overlaps but also IO) that now made
Frans Pop yell due to his machine now spitting out lots of I/O overlap
warnings -- which turned out to not be real overlaps, but due to a
uninialized BAR.
You might want Frans to test this though then...
Rene.
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
