2.5.75 out

As you may have seen, Linus just released 2.5.75, which will be the last kernel in the 2.5 series (next one will be pre-2.6). If you've been procrastinating with testing 2.5, now might be a good time to get going... ;-)

There is good and bad news: the bad news is that 2.5.75 doesn't work out-of-the-box on ia64. The good news it _does_ compile. The thing that keeps it from working are a few (late) per-CPU related changes. Hopefully, we can iron those out before pre-2.6 comes out so that it will build _and_ work. Probably wishful thinking though. There always seems to be that last-minute patch that breaks things...

	--david

Compiles, but doesn't link for me ... Linus's tree is missing this patch:

diff ~/BK/david2.5/drivers/char/agp/i460-agp.c drivers/char/agp/i460-agp.c 611c611

---
< .remove = __exit_p(agp_intel_i460_remove),

Tony> Compiles, but doesn't link for me ... Linus's tree is missing Tony> this patch:

Ah, I have the AGP stuff configured as modules so I didn't notice (didn't build the modules).

Tony> diff ~/BK/david2.5/drivers/char/agp/i460-agp.c drivers/char/agp/i460-agp.c Tony> 611c611

Tony> ---
Tony> < .remove = __exit_p(agp_intel_i460_remove),
Tony> -
Tony> To unsubscribe from this list: send the line "unsubscribe linux-ia64" in

Can you submit this directly to Dave Jones? Have you actually tried i460 AGP recently? Does it still work?

	--david

I've built a few 2.5 kernels for a BigSur, but it looks like I haven't actually booted one for many months ... and even when I did, I may not have exercised the AGP as I usually configure for init run level 3, so X doesn't get started.

-Tony

DavidM> There is good and bad news: the bad news is that 2.5.75 doesn't DavidM> work out-of-the-box on ia64. The good news it _does_ compile. The DavidM> thing that keeps it from working are a few (late) per-CPU related DavidM> changes.

There still appears to be an ACPI issue.

On bootup for a single processor I2000 (so the per-cpu problem doesn't arise), the interrupt routing isn't set up properly, so nothing that needs an interrupt work properly.

From the log:
...
ACPI Namespace successfully loaded at root a00000010067b8b0

evxfevnt-0093 [04] acpi_enable           : Transition to ACPI mode successful
evgpeblk-0748 [06] ev_create_gpe_block   : GPE 00 to 15 [_GPE] 2 regs at 0000000000000A0C on int 9
evgpeblk-0748 [06] ev_create_gpe_block   : GPE 16 to 31 [_GPE] 2 regs at 0000000000000A16 on int 9

ACPI: SCI (IRQ-1) allocation failed
evevent-0133: *** Error: Unable to install System Control Interrupt Handler, AE_NOT_ACQUIRED
ACPI: Unable to start the ACPI Interpreter
evxfevnt-0135 [06] acpi_disable : Could not exit ACPI mode to
legacy mode
evmisc-0563 [05] ev_terminate : acpi_disable failed
utalloc-0986 [05] ut_dump_allocations : No outstanding allocations.
Trying to free free IRQ0
SCSI subsystem initialized
drivers/usb/core/usb.c: registered new driver usbfs
drivers/usb/core/usb.c: registered new driver hub
ACPI: ACPI tables contain no PCI IRQ routing entries
PCI: Invalid ACPI-PCI IRQ routing table
...

This should be __devexit_p(), yes?

I'll send a patch in.

Ditto for tiger-4 machine. Doesn't boot because interrupts are all messed up.

- Ken

-----Original Message-----
From: Peter Chubb [mailto:[email protected]]
Sent: Thursday, July 10, 2003 7:54 PM
To: [email protected]
Cc: [email protected]
Subject: RE: 2.5.75 out

DavidM> There is good and bad news: the bad news is that 2.5.75 doesn't DavidM> work out-of-the-box on ia64. The good news it _does_ compile. The
DavidM> thing that keeps it from working are a few (late) per-CPU related
DavidM> changes.

There still appears to be an ACPI issue.

On bootup for a single processor I2000 (so the per-cpu problem doesn't arise), the interrupt routing isn't set up properly, so nothing that needs an interrupt work properly.

From the log:
...
ACPI Namespace successfully loaded at root a00000010067b8b0
evxfevnt-0093 [04] acpi_enable : Transition to ACPI mode
successful
evgpeblk-0748 [06] ev_create_gpe_block : GPE 00 to 15 [_GPE] 2 regs at
0000000000000A0C on int 9
evgpeblk-0748 [06] ev_create_gpe_block : GPE 16 to 31 [_GPE] 2 regs at
0000000000000A16 on int 9
ACPI: SCI (IRQ-1) allocation failed
evevent-0133: *** Error: Unable to install System Control Interrupt
Handler, AE_NOT_ACQUIRED
ACPI: Unable to start the ACPI Interpreter
evxfevnt-0135 [06] acpi_disable : Could not exit ACPI mode to
legacy mode
evmisc-0563 [05] ev_terminate : acpi_disable failed
utalloc-0986 [05] ut_dump_allocations : No outstanding allocations.
Trying to free free IRQ0
SCSI subsystem initialized
drivers/usb/core/usb.c: registered new driver usbfs drivers/usb/core/usb.c: registered new driver hub
ACPI: ACPI tables contain no PCI IRQ routing entries
PCI: Invalid ACPI-PCI IRQ routing table
...

Luming,

Do you know what the status is in getting the ia64 ACPI changes folded into 2.5.xx?

Thanks,

	--david

Ken> Ditto for tiger-4 machine. Doesn't boot because interrupts are Ken> all messed up.

Ken> - Ken

Ken> -----Original Message----- From: Peter Chubb
Ken> [mailto:[email protected]] Sent: Thursday, July 10, 2003 Ken> 7:54 PM To: [email protected] Cc: [email protected] Ken> Subject: RE: 2.5.75 out

DavidM> There is good and bad news: the bad news is that 2.5.75 DavidM> doesn't work out-of-the-box on ia64. The good news it DavidM> _does_ compile.
Ken> The
DavidM> thing that keeps it from working are a few (late) per-CPU Ken> related
DavidM> changes.

Ken> There still appears to be an ACPI issue.

Ken> On bootup for a single processor I2000 (so the per-cpu problem Ken> doesn't arise), the interrupt routing isn't set up properly, so Ken> nothing that needs an interrupt work properly.

Ken> ... ACPI Namespace successfully loaded at root
Ken> a00000010067b8b0 evxfevnt-0093 [04] acpi_enable : Transition to Ken> ACPI mode successful evgpeblk-0748 [06] ev_create_gpe_block :
Ken> GPE 00 to 15 [_GPE] 2 regs at 0000000000000A0C on int 9 Ken> evgpeblk-0748 [06] ev_create_gpe_block : GPE 16 to 31 [_GPE] 2 Ken> regs at 0000000000000A16 on int 9 ACPI: SCI (IRQ-1) allocation Ken> failed evevent-0133: *** Error: Unable to install System Ken> Control Interrupt Handler, AE_NOT_ACQUIRED ACPI: Unable to Ken> start the ACPI Interpreter evxfevnt-0135 [06] acpi_disable :
Ken> Could not exit ACPI mode to legacy mode evmisc-0563 [05] Ken> ev_terminate : acpi_disable failed utalloc-0986 [05] Ken> ut_dump_allocations : No outstanding allocations. Trying to Ken> free free IRQ0 SCSI subsystem initialized
Ken> drivers/usb/core/usb.c: registered new driver usbfs Ken> drivers/usb/core/usb.c: registered new driver hub ACPI: ACPI Ken> tables contain no PCI IRQ routing entries PCI: Invalid ACPI-PCI Ken> IRQ routing table ... - To unsubscribe from this list: send Ken> the line "unsubscribe linux-ia64" in the body of a message to Ken> [email protected] More majordomo info at Ken> http://vger.kernel.org/majordomo-info.html

David,

The straightforward patch is ready.
My part of the patch for removing #ifdef CONFIG_IA64 in ACPI generic code is ready too.

Wilcox,
Is patch for PCI domain-releated stuff ready? (I'm not sure),

I want Andy to apply the straightforward patch first, then apply the patches for generic code.

Thanks,
Luming

-----Original Message-----
From: David Mosberger [mailto:[email protected]]
Sent: 2003?7?12? 2:31
To: Yu, Luming
Cc: Chen, Kenneth W; Peter Chubb; [email protected]; [email protected]
Subject: RE: 2.5.75 out

Luming,

Do you know what the status is in getting the ia64 ACPI changes folded into 2.5.xx?

Thanks,

	--david

Ken> Ditto for tiger-4 machine. Doesn't boot because interrupts are Ken> all messed up.

Ken> - Ken

Ken> -----Original Message----- From: Peter Chubb
Ken> [mailto:[email protected]] Sent: Thursday, July 10, 2003 Ken> 7:54 PM To: [email protected] Cc: [email protected] Ken> Subject: RE: 2.5.75 out

DavidM> There is good and bad news: the bad news is that 2.5.75 DavidM> doesn't work out-of-the-box on ia64. The good news it DavidM> _does_ compile.
Ken> The
DavidM> thing that keeps it from working are a few (late) per-CPU Ken> related
DavidM> changes.

Ken> There still appears to be an ACPI issue.

Ken> On bootup for a single processor I2000 (so the per-cpu problem Ken> doesn't arise), the interrupt routing isn't set up properly, so Ken> nothing that needs an interrupt work properly.

Ken> ... ACPI Namespace successfully loaded at root
Ken> a00000010067b8b0 evxfevnt-0093 [04] acpi_enable : Transition to Ken> ACPI mode successful evgpeblk-0748 [06] ev_create_gpe_block :
Ken> GPE 00 to 15 [_GPE] 2 regs at 0000000000000A0C on int 9 Ken> evgpeblk-0748 [06] ev_create_gpe_block : GPE 16 to 31 [_GPE] 2 Ken> regs at 0000000000000A16 on int 9 ACPI: SCI (IRQ-1) allocation Ken> failed evevent-0133: *** Error: Unable to install System Ken> Control Interrupt Handler, AE_NOT_ACQUIRED ACPI: Unable to Ken> start the ACPI Interpreter evxfevnt-0135 [06] acpi_disable :
Ken> Could not exit ACPI mode to legacy mode evmisc-0563 [05] Ken> ev_terminate : acpi_disable failed utalloc-0986 [05] Ken> ut_dump_allocations : No outstanding allocations. Trying to Ken> free free IRQ0 SCSI subsystem initialized
Ken> drivers/usb/core/usb.c: registered new driver usbfs Ken> drivers/usb/core/usb.c: registered new driver hub ACPI: ACPI Ken> tables contain no PCI IRQ routing entries PCI: Invalid ACPI-PCI Ken> IRQ routing table ... - To unsubscribe from this list: send Ken> the line "unsubscribe linux-ia64" in the body of a message to Ken> [email protected] More majordomo info at Ken> http://vger.kernel.org/majordomo-info.html

Luming> David, The straightforward patch is ready. My part of the Luming> patch for removing #ifdef CONFIG_IA64 in ACPI generic code Luming> is ready too.

Did you see the patch I sent out yesterday? AFAIK, it's the only thing that's left.

	--david

Yes, PCI domain support patches have ben trickling into the kernel since about 2.5.72 and I think we're now done as far as arch/ia64 and drivers/acpi are concerned. I'm now turning my attention to drivers I need to care about. Right now, that means the sym2 driver...

"It's not Hollywood.  War is real, war is primarily not about defeat or
victory, it is about death.  I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk