Skip to contentSkip to conversation

no boot with 2.6.21-rc3 and later

BUG New Bob Tracy 4 participants 14 messages Mar 21, 2007 – Mar 22, 2007 Clocksource, clockevent drivers

lore
Share & tools
lore

Bob Tracy

author
#0 lore raw
Quoted 10 lines · I originally

Platform is a Dell CPxJ 650GT notebook. Attempts to boot 2.6.21-rc3 and -rc4 produce the following console output:

(apparently normal boot)
...
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
PIIX4: IDE controller at PCI slot 0000:00:07.1
PIIX4: chipset revision 1
PIIX4: not 100% native mode: will probe irqs later

Quoted 7 lines
    ide0: BM-DMA at 0x0860-0x0867, BIOS settings: hda:DMA, hdb:pio
    ide1: BM-DMA at 0x0868-0x086f, BIOS settings: hdc:DMA, hdd:pio

(hangs here)

The next two lines of output for 2.6.21-rc2 and earlier are:
Probing IDE interface ide0...
hda: TOSHIBA MK3017GAP, ATA DISK drive

After nine iterations of git-bisect --> compile --> reboot over the past two days, the commit that caused this problem has been identified.

6bb74df481223731af6c7e0ff3adb31f6442cfcd is first bad commit commit 6bb74df481223731af6c7e0ff3adb31f6442cfcd

Author: john stultz <[email protected]>
Date:   Mon Mar 5 00:30:50 2007 -0800

[PATCH] clocksource init adjustments (fix bug #7426)

This patch resolves the issue found here:
http://bugme.osdl.org/show_bug.cgi?id=7426

The basic summary is:
(omitted for brevity)

Signed-off-by John Stultz <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "David S. Miller" <[email protected]>
Signed-off-by Andrew Morton <[email protected]>
Signed-off-by Linus Torvalds <[email protected]>
:040000 040000 b259cf01f6ad791a0f58c9956e6e0111cb9aa310 0e8fbf31c0bae9ca6039fc7781b53c12f7a67592 M      arch
:040000 040000 19bd5cf3c55c6b7947ec66b68eab643c99f15bb8 ef8fda0445287e34eeb466985901ca64191b6539 M      drivers
:040000 040000 4f84ad7bdf509ed83efd95d174796f1a6a948b0e 51624663be2fdce8aaeb3b49e188520aa3735d90 M      include
:040000 040000 ef82986d4590fed84b9a128da90e65f7e43f43a9 778d0cf7273d7514b436ca4e91a6948f34694972 M      kernel
Signature · 4 lines
-----------------------------------------------------------------------
Bob Tracy                   WTO + WIPO = DMCA? http://www.anti-dmca.org
[email protected]
-----------------------------------------------------------------------
#1 lore raw

Thanks for narrowing it down, I'm adding John to cc:.

Quoted 52 lines · Bob Tracy
Quoted 10 lines · I originally

Platform is a Dell CPxJ 650GT notebook. Attempts to boot 2.6.21-rc3 and -rc4 produce the following console output:

(apparently normal boot)
...
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
PIIX4: IDE controller at PCI slot 0000:00:07.1
PIIX4: chipset revision 1
PIIX4: not 100% native mode: will probe irqs later

Quoted 7 lines
    ide0: BM-DMA at 0x0860-0x0867, BIOS settings: hda:DMA, hdb:pio
    ide1: BM-DMA at 0x0868-0x086f, BIOS settings: hdc:DMA, hdd:pio

(hangs here)

The next two lines of output for 2.6.21-rc2 and earlier are:
Probing IDE interface ide0...
hda: TOSHIBA MK3017GAP, ATA DISK drive

After nine iterations of git-bisect --> compile --> reboot over the past two days, the commit that caused this problem has been identified.

6bb74df481223731af6c7e0ff3adb31f6442cfcd is first bad commit commit 6bb74df481223731af6c7e0ff3adb31f6442cfcd

Author: john stultz <[email protected]>
Date:   Mon Mar 5 00:30:50 2007 -0800

[PATCH] clocksource init adjustments (fix bug #7426)

This patch resolves the issue found here:
http://bugme.osdl.org/show_bug.cgi?id=7426

The basic summary is:
(omitted for brevity)

Signed-off-by John Stultz <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "David S. Miller" <[email protected]>
Signed-off-by Andrew Morton <[email protected]>
Signed-off-by Linus Torvalds <[email protected]>
:040000 040000 b259cf01f6ad791a0f58c9956e6e0111cb9aa310 0e8fbf31c0bae9ca6039fc7781b53c12f7a67592 M      arch
:040000 040000 19bd5cf3c55c6b7947ec66b68eab643c99f15bb8 ef8fda0445287e34eeb466985901ca64191b6539 M      drivers
:040000 040000 4f84ad7bdf509ed83efd95d174796f1a6a948b0e 51624663be2fdce8aaeb3b49e188520aa3735d90 M      include
:040000 040000 ef82986d4590fed84b9a128da90e65f7e43f43a9 778d0cf7273d7514b436ca4e91a6948f34694972 M      kernel
Signature · 4 lines
-----------------------------------------------------------------------
Bob Tracy                   WTO + WIPO = DMCA? http://www.anti-dmca.org
[email protected]
-----------------------------------------------------------------------
#2 lore raw
Quoted 21 lines · Bartlomiej Zolnierkiewicz

Thanks for narrowing it down, I'm adding John to cc:.

Quoted 18 lines · Bob Tracy
Quoted 17 lines · I originally

Platform is a Dell CPxJ 650GT notebook. Attempts to boot 2.6.21-rc3 and -rc4 produce the following console output:

(apparently normal boot)
...
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
PIIX4: IDE controller at PCI slot 0000:00:07.1
PIIX4: chipset revision 1
PIIX4: not 100% native mode: will probe irqs later

    ide0: BM-DMA at 0x0860-0x0867, BIOS settings: hda:DMA, hdb:pio
    ide1: BM-DMA at 0x0868-0x086f, BIOS settings: hdc:DMA, hdd:pio

(hangs here)

The next two lines of output for 2.6.21-rc2 and earlier are:
Probing IDE interface ide0...
hda: TOSHIBA MK3017GAP, ATA DISK drive

Hmmm. Are you able to capture a full dmesg? If so could you send it to me? Also, does booting w/ "clocksource=jiffies" change the behavior?

thanks
-john

author
#3 lore raw
Quoted 1 line · john stultz

Hmmm. Are you able to capture a full dmesg?

For the "it hangs" case, short of either trying to boot with the console I/O redirected to a serial port or hooking up a printer to the console and scanning in the resulting output, I don't think so. There's too much output prior to the hang point to make manual transcription feasible.

Quoted 1 line

Also, does booting w/ "clocksource=jiffies" change the behavior?

I'll build a post-"bad"-commit kernel later this evening and let you know. Unfortunately, real life has to intrude for a few hours :-). If specifying the clocksource allows for a normal bootup, would you like to see the corresponding dmesg output?

Signature · 4 lines
-----------------------------------------------------------------------
Bob Tracy                   WTO + WIPO = DMCA? http://www.anti-dmca.org
[email protected]
-----------------------------------------------------------------------
#4 lore raw
Quoted 8 lines · Bob Tracy
Quoted 1 line · john stultz

Hmmm. Are you able to capture a full dmesg?

For the "it hangs" case, short of either trying to boot with the console I/O redirected to a serial port or hooking up a printer to the console and scanning in the resulting output, I don't think so. There's too much output prior to the hang point to make manual transcription feasible.

Hmm. You could give netconsole a try(see Documentation/networking/netconsole.txt), although if the hang is before that inits you might be out of luck.

Quoted 4 lines
Quoted 1 line

Also, does booting w/ "clocksource=jiffies" change the behavior?

I'll build a post-"bad"-commit kernel later this evening and let you know. Unfortunately, real life has to intrude for a few hours :-).

No problem. When you can get to it. :)

Also trying booting w/ "notsc" would be a useful data point.

Quoted 2 lines

If specifying the clocksource allows for a normal bootup, would you like to see the corresponding dmesg output?

No, although a pre-bad-commit dmesg would help.

thanks
-john

author
#5 lore raw
Quoted 1 line · john stultz
Quoted 1 line
Quoted 1 line

Also, does booting w/ "clocksource=jiffies" change the behavior?

Works fine with 2.6.21-rc4. I'm running on that kernel as I type this.

Quoted 1 line

Also trying booting w/ "notsc" would be a useful data point.

Boot hangs at the point indicated in my original message. I *did* notice the blurb in the console messages about the pit clocksource being selected/used. There was also a complaint about it being unstable, with a negative delta. I think these messages are consistent with the pre-bad-commit case other than where they appear in the boot messages.

Quoted 1 line

(...) a pre-bad-commit dmesg would help.

Sent under separate cover to John.

Signature · 4 lines
-----------------------------------------------------------------------
Bob Tracy                   WTO + WIPO = DMCA? http://www.anti-dmca.org
[email protected]
-----------------------------------------------------------------------
#6 lore raw
Quoted 14 lines · Bob Tracy
Quoted 1 line · john stultz
Quoted 1 line
Quoted 1 line

Also, does booting w/ "clocksource=jiffies" change the behavior?

Works fine with 2.6.21-rc4. I'm running on that kernel as I type this.

Quoted 1 line

Also trying booting w/ "notsc" would be a useful data point.

Boot hangs at the point indicated in my original message. I *did* notice the blurb in the console messages about the pit clocksource being selected/used. There was also a complaint about it being unstable, with a negative delta. I think these messages are consistent with the pre-bad-commit case other than where they appear in the boot messages.

Quoted 3 lines
Quoted 1 line

(...) a pre-bad-commit dmesg would help.

Sent under separate cover to John.

Yea. From looking at your boot log, it does look like the PIT clocksource is being used.

This does point to an issue w/ that clocksource, however I noticed its getting selected over the ACPI PM timer, because you have the PIIX4 bug. That bug drops the ACPI PM timers rating to the same as the PIT, so that's why you're getting the PIT.

Realistically, the PIT is much slower then even the triple read ACPI PM, so the de-ranking code is probably dropping it too far.

Try this patch and let me know if it does the right thing. Although I do still need to dig a bit on the PIT hang issue.

-john

Don't drop ACPI PM quite so low if we see the PIIX4 bug.

Signed-off-by John Stultz <[email protected]>
drivers/clocksource/acpi_pm.c +1 -1
index d42060e..5ac309e 100644
@@ -90,7 +90,7 @@ __setup("acpi_pm_good", acpi_pm_good_set
 static inline void acpi_pm_need_workaround(void)
 {
 	clocksource_acpi_pm.read = acpi_pm_read_slow;
-	clocksource_acpi_pm.rating = 110;
+	clocksource_acpi_pm.rating = 120;
 }
 
 /*
author
#7 lore raw
Quoted 1 line · john stultz

Try this patch and let me know if it does the right thing.

Will do. I'll report back in a few hours.

Quoted 2 lines

Although I do
still need to dig a bit on the PIT hang issue.

Any chance this might be related to the APIC issues currently being discussed in other threads?

Signature · 4 lines
-----------------------------------------------------------------------
Bob Tracy                   WTO + WIPO = DMCA? http://www.anti-dmca.org
[email protected]
-----------------------------------------------------------------------
#8 lore raw
Quoted 10 lines · Bob Tracy
Quoted 1 line · john stultz

Try this patch and let me know if it does the right thing.

Will do. I'll report back in a few hours.

Quoted 2 lines

Although I do
still need to dig a bit on the PIT hang issue.

Any chance this might be related to the APIC issues currently being discussed in other threads?

Hmmm. Good thought, I'll have to look into it. It could be that if the PIT is disabled in favor of the local apic, we'll have to make sure its not being used as a clocksource.

-john

#9 lore raw
Quoted 9 lines · john stultz
Quoted 8 lines · Bob Tracy
Quoted 1 line · john stultz

Try this patch and let me know if it does the right thing.

Will do. I'll report back in a few hours.

Quoted 2 lines

Although I do
still need to dig a bit on the PIT hang issue.

Quoted 6 lines
Quoted 2 lines

Any chance this might be related to the APIC issues currently being discussed in other threads?

Hmmm. Good thought, I'll have to look into it. It could be that if the PIT is disabled in favor of the local apic, we'll have to make sure its not being used as a clocksource.

Ouch, yes. That's fatal and can happen. Not sure, what to do about that.

tglx

author
#10 lore raw
In reply to john stultz
Quoted 9 lines · john stultz
Quoted 8 lines · Bob Tracy
Quoted 1 line · john stultz

Try this patch and let me know if it does the right thing.

Will do. I'll report back in a few hours.

Quoted 2 lines

Although I do
still need to dig a bit on the PIT hang issue.

Quoted 6 lines
Quoted 2 lines

Any chance this might be related to the APIC issues currently being discussed in other threads?

Hmmm. Good thought, I'll have to look into it. It could be that if the PIT is disabled in favor of the local apic, we'll have to make sure its not being used as a clocksource.

According to the dmesg output, the local apic is disabled in the BIOS on this system, and I'm not booting with "lapic" to enable it.

Anyway, the "few hours" turned out to be more like a few minutes :-). The patch worked as far as selecting/installing the acpi_pm clocksource instead of the pit clocksource, and there was no hang during boot. If you would like to see the dmesg output, let me know and I'll forward it to you in a separate message.

Signature · 4 lines
-----------------------------------------------------------------------
Bob Tracy                   WTO + WIPO = DMCA? http://www.anti-dmca.org
[email protected]
-----------------------------------------------------------------------
#11 lore raw
Quoted 17 lines · Bob Tracy
Quoted 9 lines · john stultz
Quoted 8 lines · Bob Tracy
Quoted 1 line · john stultz

Try this patch and let me know if it does the right thing.

Will do. I'll report back in a few hours.

Quoted 2 lines

Although I do
still need to dig a bit on the PIT hang issue.

Quoted 6 lines
Quoted 2 lines

Any chance this might be related to the APIC issues currently being discussed in other threads?

Hmmm. Good thought, I'll have to look into it. It could be that if the PIT is disabled in favor of the local apic, we'll have to make sure its not being used as a clocksource.

Quoted 8 lines

According to the dmesg output, the local apic is disabled in the BIOS on this system, and I'm not booting with "lapic" to enable it.

Anyway, the "few hours" turned out to be more like a few minutes :-). The patch worked as far as selecting/installing the acpi_pm clocksource instead of the pit clocksource, and there was no hang during boot. If you would like to see the dmesg output, let me know and I'll forward it to you in a separate message.

Great to hear! Please do send me the dmesg output so I can look to narrow down what's happening w/ the PIT.

I'll also go ahead and push that acpi_pm change to akpm.

thanks
-john

author
#12 lore raw
Quoted 2 lines · john stultz

Great to hear! Please do send me the dmesg output so I can look to narrow down what's happening w/ the PIT.

Sent under separate cover...

Signature · 4 lines
-----------------------------------------------------------------------
Bob Tracy                   WTO + WIPO = DMCA? http://www.anti-dmca.org
[email protected]
-----------------------------------------------------------------------
#13 lore raw

PATCH 1/1 correct slow acpi_pm rating (fixes no boot with 2.6.21-rc3 and later)

In reply to Bob Tracy

Tested by Bob Tracy.

Quoted 14 lines · Bob Tracy
Quoted 1 line · john stultz
Quoted 1 line
Quoted 1 line

Also, does booting w/ "clocksource=jiffies" change the behavior?

Works fine with 2.6.21-rc4. I'm running on that kernel as I type this.

Quoted 1 line

Also trying booting w/ "notsc" would be a useful data point.

Boot hangs at the point indicated in my original message. I *did* notice the blurb in the console messages about the pit clocksource being selected/used. There was also a complaint about it being unstable, with a negative delta. I think these messages are consistent with the pre-bad-commit case other than where they appear in the boot messages.

Quoted 3 lines
Quoted 1 line

(...) a pre-bad-commit dmesg would help.

Sent under separate cover to John.

Yea. From looking at your boot log, it does look like the PIT clocksource is being used.

This does point to an issue w/ that clocksource, however I noticed its getting selected over the ACPI PM timer, because you have the PIIX4 bug. That bug drops the ACPI PM timers rating to the same as the PIT, so that's why you're getting the PIT.

Realistically, the PIT is much slower then even the triple read ACPI PM, so the de-ranking code is probably dropping it too far.

Try this patch and let me know if it does the right thing. Although I do still need to dig a bit on the PIT hang issue.

-john

Don't drop ACPI PM quite so low if we see the PIIX4 bug.

Signed-off-by John Stultz <[email protected]>
drivers/clocksource/acpi_pm.c +1 -1
index d42060e..5ac309e 100644
@@ -90,7 +90,7 @@ __setup("acpi_pm_good", acpi_pm_good_set
 static inline void acpi_pm_need_workaround(void)
 {
 	clocksource_acpi_pm.read = acpi_pm_read_slow;
-	clocksource_acpi_pm.rating = 110;
+	clocksource_acpi_pm.rating = 120;
 }
 
 /*