ACPI for 2.4

Hi Marcelo, please do a

bk pull http://linux-acpi.bkbits.net/linux-acpi-release-2.4.23

Asside from the fixes we discussed, I found some of the ACPI related x86_64 code in need of sync from i386, so I did that and iterated to fix a few broken build configs. Of my standard set, !CONFIG_PCI is still broken, but probably nobody cares. the pci=noacpi code is updated from 2.6, but both trees needs some cleaning...

It would be good to get some feedback from those with x86_64 hardware, since I don't currently have any.

thanks,
-Len

ps. a plain patch is also available here:
ftp://ftp.kernel.org/pub/linux/kernel/people/lenb/acpi/patches/release/2.4.23-rc1/acpi-20031002-2.4.23-rc1.diff.gz

This will update the following files:

 Documentation/kernel-parameters.txt |    9 +
 MAINTAINERS                         |    4 
 arch/i386/kernel/acpi.c             |   58 ++++++-
 arch/i386/kernel/dmi_scan.c         |    1 
 arch/i386/kernel/io_apic.c          |   10 -
 arch/i386/kernel/mpparse.c          |    4 
 arch/i386/kernel/setup.c            |    6 
 arch/x86_64/kernel/Makefile         |    2 
 arch/x86_64/kernel/acpi.c           |   98 ++++++++++---
 arch/x86_64/kernel/e820.c           |   10 +
 arch/x86_64/kernel/io_apic.c        |   20 +-
 arch/x86_64/kernel/mpparse.c        |    8 -
 arch/x86_64/kernel/pci-pc.c         |    8 +
 arch/x86_64/kernel/setup.c          |    4 
 drivers/acpi/bus.c                  |    4 
 drivers/acpi/pci_link.c             |  180 ++++++++++++++++++++----
 drivers/acpi/system.c               |    9 +
 include/acpi/acpi_bus.h             |    1 
 include/asm-i386/acpi.h             |   29 +--
 include/asm-i386/io_apic.h          |   23 ++-
 include/asm-x86_64/acpi.h           |   28 +--
 include/asm-x86_64/apic.h           |   13 -
 include/asm-x86_64/io_apic.h        |   21 ++
 23 files changed, 412 insertions(+), 138 deletions(-)

through these ChangeSets:

<[email protected]> (03/11/18 1.1193.1.10)
"pci=noacpi" use pci_disable_acpi() instead of touching use_acpi_pci
directly

<[email protected]> (03/11/18 1.1063.44.51)
[ACPI] sync some i386 ACPI build fixes into x86_64 to fix
!CONFIG_ACPI build

<[email protected]> (03/11/18 1.1193.1.8)
x86_64 build fix from previous cset

<[email protected]> (03/11/18 1.1063.44.50)
i386 build fix from previous cset

<[email protected]> (03/11/18 1.1193.1.6)
2.4.23 build x86_64 build fixes

<[email protected]> (03/11/18 1.1063.44.49)
[ACPI] fix x86_64 !CONFIG_ACPI build

<[email protected]> (03/11/18 1.1063.44.48)
[ACPI] Maintainer: Andy Grover -> Len Brown

<[email protected]> (03/11/18 1.1063.44.47)
[ACPI] fix x86_64 build errors

<[email protected]> (03/11/18 1.1063.44.46)
[ACPI]
Re-enable IRQ balacning if IOAPIC mode
http://bugzilla.kernel.org/show_bug.cgi?id=1440

Also allow IRQ balancing in PIC mode if "acpi_irq_balance"
http://bugzilla.kernel.org/show_bug.cgi?id=1391

<[email protected]> (03/11/18 1.1063.44.45)
[ACPI ] pci=acpi ineffective fix from i386 2.6 (Thomas Schlichter)
http://bugzilla.kernel.org/show_bug.cgi?id=1219

<[email protected]> (03/11/18 1.1063.44.44)
[ACPI] "acpi_pic_sci=edge" in case platform requires Edge Triggered
SCI
http://bugzilla.kernel.org/show_bug.cgi?id=1390

<[email protected]> (03/11/18 1.1063.44.43)
[ACPI] print_IO_APIC() only after it is actually programmed
http://bugzilla.kernel.org/show_bug.cgi?id=1177

<[email protected]> (03/11/17 1.1063.44.42)
[ACPI] fix ACPI/legacy interrupt sharing issue ala 2.6
http://bugzilla.kernel.org/show_bug.cgi?id=1283

<[email protected]> (03/11/17 1.1063.44.41)
[ACPI] fix poweroff failure ala 2.6 (Ducrot Bruno)
http://bugzilla.kernel.org/show_bug.cgi?id=1456

<[email protected]> (03/11/07 1.1063.44.40)
[ACPI] In ACPI mode, delay print_IO_APIC() to make its output valid.
http://bugzilla.kernel.org/show_bug.cgi?id=1177

<[email protected]> (03/11/07 1.1063.44.39)
[ACPI] If ACPI is disabled by DMI BIOS date, then turn it off completely, including table parsing for HT. This avoids a crash due to ancient garbled tables. acpi=force is available to over-ride this default.
http://bugzilla.kernel.org/show_bug.cgi?id=1434

<[email protected]> (03/10/30 1.1063.44.38)
[ACPI] fix CONFIG_HOTPLUG_PCI_ACPI config (Xose Vazquez Perez)

Hi Marcelo, please do a

bk pull http://linux-acpi.bkbits.net/linux-acpi-release-2.4.23

A panic fix from Matt Wilcox via 2.6, and a build fix from me. (List of kernels built below.)
Same caveat as before, I've not booted on x86_64 hardware.

thanks,
-Len

ps. a plain patch is also available here:
ftp://ftp.kernel.org/pub/linux/kernel/people/lenb/acpi/patches/release/2.4.23-rc2/acpi-20031002-2.4.23-rc2.diff.gz

This will update the following files:

 arch/i386/kernel/acpi.c     |    7 +++----
 arch/i386/kernel/dmi_scan.c |    5 ++---
 arch/i386/kernel/pci-pc.c   |   12 ++----------
 arch/i386/kernel/setup.c    |    9 +++------
 arch/x86_64/kernel/acpi.c   |   19 ++-----------------
 arch/x86_64/kernel/e820.c   |    9 +++++----
 arch/x86_64/kernel/pci-pc.c |   10 +---------
 arch/x86_64/kernel/setup.c  |    2 +-
 drivers/acpi/pci_root.c     |    4 +---
 include/asm-i386/acpi.h     |    6 ++++--
 include/asm-x86_64/acpi.h   |    9 ++++-----
 11 files changed, 28 insertions(+), 64 deletions(-)

through these ChangeSets:

<[email protected]> (03/11/20 1.1208)
[ACPI] "pci=noacpi" -- 2.4.23 specific part of previous 2.4.22 fix

<[email protected]> (03/11/20 1.1063.44.53)
[ACPI] "pci=noacpi" -- replace two sets of flags with one: acpi_noirq

<[email protected]> (03/11/19 1.1063.44.52)
[PATCH] Fix panic-at-boot

This fixes a panic-at-boot when ACPI Hotplug PCI is compiled in, but ACPI is disabled. It just makes sure that the list is properly initialized statically instead of depending on runtime initialization that may or may not happen.

x86_64:
1. errs.ACPI.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
2. errs.ACPI.PCI.X86_LOCAL_APIC.X86_IO_APIC
3. errs.ACPI.SMP.X86_LOCAL_APIC.X86_IO_APIC
4. errs.ACPI.X86_LOCAL_APIC.X86_IO_APIC
5. errs.DEBUG_KERNEL.ACPI.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
6. errs.DEBUG_KERNEL.ACPI.PCI.X86_LOCAL_APIC.X86_IO_APIC
7. errs.DEBUG_KERNEL.ACPI.SMP.X86_LOCAL_APIC.X86_IO_APIC
8. errs.DEBUG_KERNEL.ACPI.X86_LOCAL_APIC.X86_IO_APIC
9. errs.DEBUG_KERNEL.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
10. errs.DEBUG_KERNEL.PCI.X86_LOCAL_APIC.X86_IO_APIC
11. errs.DEBUG_KERNEL.SMP.X86_LOCAL_APIC.X86_IO_APIC
12. errs.DEBUG_KERNEL.X86_LOCAL_APIC.X86_IO_APIC
13. errs.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
14. errs.PCI.X86_LOCAL_APIC.X86_IO_APIC
15. errs.SMP.X86_LOCAL_APIC.X86_IO_APIC
16. errs.X86_LOCAL_APIC.X86_IO_APIC
16 SUCCESS in 16 ATTEMPTS

i386:
1. errs.
2. errs.ACPI
3. errs.ACPI.PCI
4. errs.ACPI.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
5. errs.ACPI.PCI.X86_LOCAL_APIC
6. errs.ACPI.PCI.X86_LOCAL_APIC.X86_IO_APIC
7. errs.ACPI.SMP.X86_LOCAL_APIC.X86_IO_APIC
8. errs.ACPI.X86_LOCAL_APIC
9. errs.ACPI.X86_LOCAL_APIC.X86_IO_APIC
10. errs.DEBUG_KERNEL
11. errs.DEBUG_KERNEL.ACPI
12. errs.DEBUG_KERNEL.ACPI.PCI
13. errs.DEBUG_KERNEL.ACPI.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
14. errs.DEBUG_KERNEL.ACPI.PCI.X86_LOCAL_APIC
15. errs.DEBUG_KERNEL.ACPI.PCI.X86_LOCAL_APIC.X86_IO_APIC
16. errs.DEBUG_KERNEL.ACPI.SMP.X86_LOCAL_APIC.X86_IO_APIC
17. errs.DEBUG_KERNEL.ACPI.X86_LOCAL_APIC
18. errs.DEBUG_KERNEL.ACPI.X86_LOCAL_APIC.X86_IO_APIC
19. errs.DEBUG_KERNEL.PCI
20. errs.DEBUG_KERNEL.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
21. errs.DEBUG_KERNEL.PCI.X86_LOCAL_APIC
22. errs.DEBUG_KERNEL.PCI.X86_LOCAL_APIC.X86_IO_APIC
23. errs.DEBUG_KERNEL.SMP.X86_LOCAL_APIC.X86_IO_APIC
24. errs.DEBUG_KERNEL.X86_LOCAL_APIC
25. errs.DEBUG_KERNEL.X86_LOCAL_APIC.X86_IO_APIC
26. errs.PCI
27. errs.PCI.SMP.X86_LOCAL_APIC.X86_IO_APIC
28. errs.PCI.X86_LOCAL_APIC
29. errs.PCI.X86_LOCAL_APIC.X86_IO_APIC
30. errs.SMP.X86_LOCAL_APIC.X86_IO_APIC
31. errs.X86_LOCAL_APIC
32. errs.X86_LOCAL_APIC.X86_IO_APIC
32 SUCCESS in 32 ATTEMPTS

Hi Len,

This version, as well as the original 2.4.23-rc2 code cannot reboot my VAIO. I don't know yet when the problem got in, since I've been using a 2.4.21 base + acpi-20030424-2.4.21-rc1.diff for a long time without this problem, and don't know if this was the case with intermediate versions since it's not something that I immediately notice.

I will try to identify which pre-release brought this problem.

Basically, when I reboot, the screen goes black, and I hear the speakers shut down exactly as when it will reboot. But nothing appears on the screen anymore. I must say that I also had this behaviour with the above version only when I used LOCAL_APIC. But I tried both with and without, without success. I clearly suspect my BIOS since it's a real crap, but since it once worked, I'll search a bit.

Other than that, power down works pretty well (this has always been a problem on this crap too), and the rest of the kernel is really fine and stable.

Regards,
Willy

Hi again,

in fact this is not an ACPI problem. It's the ATYFB driver which never worked before on this machine which prevents it from rebooting. That doesn't surprize me, since this bios doesn't reinitialize anything properly (previously, this bios hung on the IDE controller if APIC was set).

Once back to VESAFB, everything's OK, so there's no regression at all.

Sorry for the noise,
Willy