| Message ID | 1302598086-3622-1-git-send-email-jslaby@suse.cz (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 04/12/2011 10:48 AM, Jiri Slaby wrote: > On IBM Bartolo machines, cards in 00:09.0 are defunct (if use > interrupts). DSDT says that this slot (with function 0, i.e. pin A) is > routed to \_SB_.PCI0.PIB_.LNKB. But it's not, it's wired to LNKC, so > interrupts are misrouted. Add a quirk for this to workaround the > issue. > > References: https://bugzilla.novell.com/show_bug.cgi?id=595683 > References: https://bugzilla.kernel.org/show_bug.cgi?id=18092 > > --- > Hi, > > as Robert Hancock suggested at: > http://lkml.org/lkml/2010/9/7/206 > > I reported this in bugzilla (the link above). But I got only replies > requesting another infos which I provided. BTW. the info is (w/ and w/o acpi=noirq): acpidump: https://bugzilla.kernel.org/attachment.cgi?id=29332 dmesg w/: https://bugzilla.kernel.org/attachment.cgi?id=29322 dmesg w/o: https://bugzilla.kernel.org/attachment.cgi?id=29312 lspci w/: https://bugzilla.kernel.org/attachment.cgi?id=29712 lspci w/o: https://bugzilla.kernel.org/attachment.cgi?id=29902 BIOS PRT w/: https://bugzilla.kernel.org/attachment.cgi?id=43192 BIOS PRT w/o: https://bugzilla.kernel.org/attachment.cgi?id=43182 > Could anybody look into > the bug and confirm whether this patch is correct or not. And if yes, > could you apply that? > > --- > > Not-signed-off-by: Jiri Slaby <jslaby@suse.cz> > Cc: Len Brown <lenb@kernel.org> > Cc: Jesse Barnes <jbarnes@virtuousgeek.org> > --- > drivers/acpi/pci_irq.c | 15 +++++++++++++++ > 1 files changed, 15 insertions(+), 0 deletions(-) > > diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c > index f907cfb..e8fb00d 100644 > --- a/drivers/acpi/pci_irq.c > +++ b/drivers/acpi/pci_irq.c > @@ -124,6 +124,18 @@ static const struct dmi_system_id hp_t5710[] = { > { } > }; > > +/* https://bugzilla.novell.com/show_bug.cgi?id=595683 */ > +static const struct dmi_system_id ibm_bartolo[] = { > + { > + .ident = "IBM Bartolo", > + .matches = { > + DMI_MATCH(DMI_SYS_VENDOR, "IBM CORPORATION"), > + DMI_MATCH(DMI_PRODUCT_NAME, "4810320"), > + }, > + }, > + { } > +}; > + > struct prt_quirk { > const struct dmi_system_id *system; > unsigned int segment; > @@ -151,6 +163,9 @@ static const struct prt_quirk prt_quirks[] = { > { hp_t5710, 0, 0, 1, PCI_INTX_PIN('A'), > "\\_SB_.PCI0.LNK1", > "\\_SB_.PCI0.LNK3"}, > + { ibm_bartolo, 0, 0, 9, PCI_INTX_PIN('A'), > + "\\_SB_.PCI0.PIB_.LNKB", > + "\\_SB_.PCI0.PIB_.LNKC"}, > }; > > static void do_prt_fixups(struct acpi_prt_entry *entry, thanks,
On Tue, 12 Apr 2011 10:48:06 +0200 Jiri Slaby <jslaby@suse.cz> wrote: > On IBM Bartolo machines, cards in 00:09.0 are defunct (if use > interrupts). DSDT says that this slot (with function 0, i.e. pin A) is > routed to \_SB_.PCI0.PIB_.LNKB. But it's not, it's wired to LNKC, so > interrupts are misrouted. Add a quirk for this to workaround the > issue. > > References: https://bugzilla.novell.com/show_bug.cgi?id=595683 > References: https://bugzilla.kernel.org/show_bug.cgi?id=18092 > > --- > Hi, > > as Robert Hancock suggested at: > http://lkml.org/lkml/2010/9/7/206 > > I reported this in bugzilla (the link above). But I got only replies > requesting another infos which I provided. Could anybody look into > the bug and confirm whether this patch is correct or not. And if yes, > could you apply that? I'll defer to the ACPI experts, Len or Matthew have you checked this one out?
On Tue, Apr 12, 2011 at 08:56:22AM -0700, Jesse Barnes wrote: > I'll defer to the ACPI experts, Len or Matthew have you checked this > one out? Windows appears to work without this, so it seems likely that we're doing something wrong in some other sense, but I don't think we have a good idea as to what we're doing wrong...
On 04/12/2011 05:59 PM, Matthew Garrett wrote: > On Tue, Apr 12, 2011 at 08:56:22AM -0700, Jesse Barnes wrote: > >> I'll defer to the ACPI experts, Len or Matthew have you checked this >> one out? > > Windows appears to work without this, so it seems likely that we're > doing something wrong in some other sense, but I don't think we have a > good idea as to what we're doing wrong... Is there anything I can do about that? Provide some info, investigate what a particular code does in the kernel (e.g. acpi_pci_irq_enable) etc.? thanks,
diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c index f907cfb..e8fb00d 100644 --- a/drivers/acpi/pci_irq.c +++ b/drivers/acpi/pci_irq.c @@ -124,6 +124,18 @@ static const struct dmi_system_id hp_t5710[] = { { } }; +/* https://bugzilla.novell.com/show_bug.cgi?id=595683 */ +static const struct dmi_system_id ibm_bartolo[] = { + { + .ident = "IBM Bartolo", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "IBM CORPORATION"), + DMI_MATCH(DMI_PRODUCT_NAME, "4810320"), + }, + }, + { } +}; + struct prt_quirk { const struct dmi_system_id *system; unsigned int segment; @@ -151,6 +163,9 @@ static const struct prt_quirk prt_quirks[] = { { hp_t5710, 0, 0, 1, PCI_INTX_PIN('A'), "\\_SB_.PCI0.LNK1", "\\_SB_.PCI0.LNK3"}, + { ibm_bartolo, 0, 0, 9, PCI_INTX_PIN('A'), + "\\_SB_.PCI0.PIB_.LNKB", + "\\_SB_.PCI0.PIB_.LNKC"}, }; static void do_prt_fixups(struct acpi_prt_entry *entry,