Re: v2.6.23-rc2 locks up during boot (without acpi=off)

From: Andrew Morton
Date: Sun Aug 12 2007 - 12:28:48 EST

On Sun, 12 Aug 2007 14:20:46 +0300 (EEST) "Ilpo Järvinen" <ilpo.jarvinen@xxxxxxxxxxx> wrote:

> On Sun, 12 Aug 2007, Andrew Morton wrote:
>
> > On Sat, 11 Aug 2007 15:39:30 +0300 (EEST) "Ilpo Järvinen" <ilpo.jarvinen@xxxxxxxxxxx> wrote:
> >
> > > I noticed that v2.6.23-rc1 locks up during boot, same thing happens now
> > > with the latest linus' tree (+net-2.6.24 and tcp-2.6 tree stuff on top
> > > of it; in -rc1 test they weren't though). The exact location of hang
> > > varies a bit though. No OOPS, does not respond to sysrq or anything else
> > > besides reset. Last known bootable one is something like 2.6.22-rc4
> > > (I usually run 2.6.21.5 on this machine, haven't tried any 2.6.22 on
> > > this after those rcs). Problem seems to start after this line:
> > >
> > > Time: acpi_pm clocksource has been installed.
> > >
> > > ...the power led starts blinking (not periodic cycle but more or less
> > > varying on-off cycle, never seen that led blink before at all, didn't know
> > > that one can make it blink :-)) and the machine gets consideably slower
> > > too. Never have it been able to complete booting all they way up to login
> > > prompt before lock up.
> > >
> > > Tried with acpi=off, boots just fine.
> >
> > It'd be great if you could run a git bisection search please.
>
> ...was already in process... :-)
>
> Here is the result:
>
> git-bisect start
> # bad: [7d57c74238cdf570bca20b711b2c0b31a553c1e5] Linux 2.6.23-rc1
> git-bisect bad 7d57c74238cdf570bca20b711b2c0b31a553c1e5
> # good: [5b78c77092a64e253fe1fde9fbbe818b49330ffc] Linux 2.6.22-rc4
> git-bisect good 5b78c77092a64e253fe1fde9fbbe818b49330ffc
> # good: [0a6d3a2a3813e7b25267366cfbf9a4a4698dd1c2] uml: fix request->sector update
> git-bisect good 0a6d3a2a3813e7b25267366cfbf9a4a4698dd1c2
> # good: [29a68ee73ec6a5510cbf9d803cbf6190b615e276] Chinese translation of Documentation/stable_api_nonsense.txt
> git-bisect good 29a68ee73ec6a5510cbf9d803cbf6190b615e276
> # good: [7f46e6ca0183568a688e6bfe40e3ab9adb305d03] Merge branch 'linus' of master.kernel.org:/pub/scm/linux/kernel/git/perex/alsa
> git-bisect good 7f46e6ca0183568a688e6bfe40e3ab9adb305d03
> # good: [753811dc82a6a39554c34c13c996c3de9f4aa634] x86_64: arch/x86_64/kernel/aperture.c lower printk severity
> git-bisect good 753811dc82a6a39554c34c13c996c3de9f4aa634
> # bad: [dc79747019b43c28d1f50aad69b8039f8d8db301] Merge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
> git-bisect bad dc79747019b43c28d1f50aad69b8039f8d8db301
> # bad: [d6da5ce8cc71a13e2f3671361c5a8bd9b82e014d] Pull sony into release branch
> git-bisect bad d6da5ce8cc71a13e2f3671361c5a8bd9b82e014d
> # good: [b43035a5ec4deecd43019728ab9347df82dd121f] Pull sbs into release branch
> git-bisect good b43035a5ec4deecd43019728ab9347df82dd121f
> # good: [e8b495fe09bc793ae26774e7b2667f7f658d56e2] Pull dock-bay into release branch
> git-bisect good e8b495fe09bc793ae26774e7b2667f7f658d56e2
> # bad: [8b8eb7d8cfc6cd95ed00cd58754e8493322505bd] ACPI: update ACPI proc I/F removal schedule
> git-bisect bad 8b8eb7d8cfc6cd95ed00cd58754e8493322505bd
> # good: [33ce2033433195ccc1fbad00d26ad854b2ab68d0] ACPI: suspend: delete toshiba S1 quirk
> git-bisect good 33ce2033433195ccc1fbad00d26ad854b2ab68d0
> # bad: [4ebf83c8cf89ab13bc23e46b0fcb6178ca23b43c] ACPI: fix empty macros found by -Wextra
> git-bisect bad 4ebf83c8cf89ab13bc23e46b0fcb6178ca23b43c
> # bad: [0dc070bb0242481a6100c95e5deaa07b267399a8] ACPI: drivers/acpi/pci_link.c: lower printk severity
> git-bisect bad 0dc070bb0242481a6100c95e5deaa07b267399a8
>
> ...didn't bother to go any further as the other one just deals with printk
> string... So this is the main suspect:
>
> commit 18eab8550397f1f3d4b8b2c5257c88dae25d58ed
> Author: Venkatesh Pallipadi <venkatesh.pallipadi@xxxxxxxxx>
> Date: Fri Jun 15 19:37:00 2007 -0400
>
> ACPI: Enable C3 even when PM2_control is zero
>
> On systems that do not have pm2_control_block, we cannot really use
> ARB_DISABLE before C3. We used to disable C3 totally on such systems.
>
> To be compatible with Windows, we need to enable C3 on such systems now.
> We just skip ARB_DISABLE step before entering the C3-state and assume
> hardware is handling things correctly. Also, ACPI spec is not clear
> about pm2_control is _needed_ for C3 or not.
>
> We have atleast one system that need this to enable C3.
>
> Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@xxxxxxxxx>
> Signed-off-by: Len Brown <len.brown@xxxxxxxxx>
>

OK, that's great, thanks. So just to double-check, could you please
confirm that the below reversion fixes this post-2.6.22 regression?



From: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>

Revert 18eab8550397f1f3d4b8b2c5257c88dae25d58ed. Due to

"Ilpo Jarvinen" <ilpo.jarvinen@xxxxxxxxxxx> wrote:
>
> I noticed that v2.6.23-rc1 locks up during boot, same thing happens now
> with the latest linus' tree (+net-2.6.24 and tcp-2.6 tree stuff on top
> of it; in -rc1 test they weren't though). The exact location of hang
> varies a bit though. No OOPS, does not respond to sysrq or anything else
> besides reset. Last known bootable one is something like 2.6.22-rc4
> (I usually run 2.6.21.5 on this machine, haven't tried any 2.6.22 on
> this after those rcs). Problem seems to start after this line:
>
> Time: acpi_pm clocksource has been installed.
>
> ...the power led starts blinking (not periodic cycle but more or less
> varying on-off cycle, never seen that led blink before at all, didn't know
> that one can make it blink :-)) and the machine gets consideably slower
> too. Never have it been able to complete booting all they way up to login
> prompt before lock up.

Cc: "Ilpo Jarvinen" <ilpo.jarvinen@xxxxxxxxxxx>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@xxxxxxxxx>
Cc: Len Brown <len.brown@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---

drivers/acpi/processor_idle.c | 20 +++++---------------
1 file changed, 5 insertions(+), 15 deletions(-)

diff -puN drivers/acpi/processor_idle.c~a drivers/acpi/processor_idle.c
--- a/drivers/acpi/processor_idle.c~a
+++ a/drivers/acpi/processor_idle.c
@@ -490,17 +490,7 @@ static void acpi_processor_idle(void)

case ACPI_STATE_C3:

- /*
- * disable bus master
- * bm_check implies we need ARB_DIS
- * !bm_check implies we need cache flush
- * bm_control implies whether we can do ARB_DIS
- *
- * That leaves a case where bm_check is set and bm_control is
- * not set. In that case we cannot do much, we enter C3
- * without doing anything.
- */
- if (pr->flags.bm_check && pr->flags.bm_control) {
+ if (pr->flags.bm_check) {
if (atomic_inc_return(&c3_cpu_count) ==
num_online_cpus()) {
/*
@@ -509,7 +499,7 @@ static void acpi_processor_idle(void)
*/
acpi_set_register(ACPI_BITREG_ARB_DISABLE, 1);
}
- } else if (!pr->flags.bm_check) {
+ } else {
/* SMP with no shared cache... Invalidate cache */
ACPI_FLUSH_CPU_CACHE();
}
@@ -521,7 +511,7 @@ static void acpi_processor_idle(void)
acpi_cstate_enter(cx);
/* Get end time (ticks) */
t2 = inl(acpi_gbl_FADT.xpm_timer_block.address);
- if (pr->flags.bm_check && pr->flags.bm_control) {
+ if (pr->flags.bm_check) {
/* Enable bus master arbitration */
atomic_dec(&c3_cpu_count);
acpi_set_register(ACPI_BITREG_ARB_DISABLE, 0);
@@ -971,9 +961,9 @@ static void acpi_processor_power_verify_
if (pr->flags.bm_check) {
/* bus mastering control is necessary */
if (!pr->flags.bm_control) {
- /* In this case we enter C3 without bus mastering */
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
- "C3 support without bus mastering control\n"));
+ "C3 support requires bus mastering control\n"));
+ return;
}
} else {
/*
_

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/