2.6.25-mm1

In reply to Arjan van de Ven

Quoted 1 line · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

The GCC stackprotector option is a no-go for me, and causes 100% repeatable fatal oopses on boot with my x86_64 box.

This is not new to 2.6.25-mm1 - but was also present in 2.6.24-rc8-mm2 (2.6.24-rc8-mm1 was good, but this option didn't exist then).

It seems that enabling the stackprotector option:

tornado boot # diff -u config-2.6.25-mm1 config-2.6.25-mm1.old

config-2.6.25-mm1.old 2008-04-18 20:09:38.000000000 +1000 renamed from config-2.6.25-mm1 2008-04-18 22:40:15.000000000 +1000 +3 -2

@@ -1,7 +1,7 @@
 #
 # Automatically generated make config: don't edit
 # Linux kernel version: 2.6.25-mm1
-# Fri Apr 18 22:25:04 2008
+# Fri Apr 18 19:57:17 2008
 #
 CONFIG_64BIT=y
 # CONFIG_X86_32 is not set

@@ -256,7 +256,8 @@
 CONFIG_X86_PAT=y
 # CONFIG_EFI is not set
 CONFIG_SECCOMP=y
-# CONFIG_CC_STACKPROTECTOR is not set
+CONFIG_CC_STACKPROTECTOR_ALL=y
+CONFIG_CC_STACKPROTECTOR=y
 # CONFIG_HZ_100 is not set
 # CONFIG_HZ_250 is not set
 CONFIG_HZ_300=y

is enough to prevent my system booting, viz:

input: Belkin Components Belkin OmniView KVM Switch as /devices/pci0000:00/0000:00:1d.1/usb3/3-1/3-1.1/3-1.1:1.0/input/input2
input: USB HID v1.00 Keyboard [Belkin Components Belkin OmniView KVM Switch] on usb-0000:00:1d.1-1.1
input: Belkin Components Belkin OmniView KVM Switch as /devices/pci0000:00/0000:00:1d.1/usb3/3-1/3-1.1/3-1.1:1.1/input/input3
input: USB HID v1.00 Mouse [Belkin Components Belkin OmniView KVM Switch] on usb-0000:00:1d.1-1.1
usbcore: registered new interface driver usbhid
usbhid: v2.6:USB HID core driver
TCP bic registered
NET: Registered protocol family 1
NET: Registered protocol family 10
lo: Disabled Privacy Extensions
NET: Registered protocol family 17
Testing -fstack-protector-all feature

------------[ cut here ]------------
WARNING: at ™š:-2145164734 0x0()
Modules linked in:

Pid: 1, comm: swapper Not tainted 2.6.25-mm1 #1

Call Trace:
 [<ffffffff802362a9>] warn_on_slowpath+0x67/0x98
 [<ffffffff802f31da>] ? proc_register+0x104/0x1b0
 [<ffffffff80237e2a>] ? printk+0x79/0x94
 [<ffffffff804f1d05>] ? register_netdevice_notifier+0xed/0x1c9
 [<ffffffff8023da80>] ? insert_resource+0x3c/0x117
 [<ffffffff8023630d>] ? __stack_chk_test+0x33/0x7b
 [<ffffffff80740ff0>] ? kernel_init+0x16d/0x30d
 [<ffffffff8020c7b8>] ? child_rip+0xa/0x12
 [<ffffffff80740e83>] ? kernel_init+0x0/0x30d
 [<ffffffff8020c7ae>] ? child_rip+0x0/0x12

---[ end trace 8d584356702633c0 ]---
BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
IP: [<0000000000000000>]
PGD 0
Oops: 0010 [1] SMP
last sysfs file:
CPU 0
Modules linked in:
Pid: 1, comm: swapper Tainted: G        W 2.6.25-mm1 #1
RIP: 0010:[<0000000000000000>]  [<0000000000000000>]
RSP: 0000:ffff8100bf05de88  EFLAGS: 00010296
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000000 RSI: 0000000000000046 RDI: 0000000000000200
RBP: ffff8100bf05de90 R08: 0000000000000000 R09: ffff8100000bcce0
R10: 0720072007200720 R11: 0720072007200720 R12: 0000000000000000
R13: ffffffff80787530 R14: 0000000000000000 R15: ffffffff8067fd3c
FS:  0000000000000000(0000) GS:ffffffff80721000(0000) knlGS:0000000000000000
CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 0000000000201000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process swapper (pid: 1, threadinfo ffff8100bf05c000, task ffff8100bf060000)
Stack:  0000000000000000 ffff8100bf05deb0 ffffffff8023630d 0000000000000000
 0000000090e2a955 ffff8100bf05df40 ffffffff80740ff0 aa55aa0000000000
 aa55aa55aa55aa55 0000000000000003 55aa55aa55aa55aa 55aa55aa55aa55aa
Call Trace:
 [<ffffffff8023630d>] __stack_chk_test+0x33/0x7b
 [<ffffffff80740ff0>] kernel_init+0x16d/0x30d
 [<ffffffff8020c7b8>] child_rip+0xa/0x12
 [<ffffffff80740e83>] ? kernel_init+0x0/0x30d
 [<ffffffff8020c7ae>] ? child_rip+0x0/0x12


Code:  Bad RIP value.
RIP  [<0000000000000000>]
 RSP <ffff8100bf05de88>
CR2: 0000000000000000
---[ end trace 8d584356702633c0 ]---
Kernel panic - not syncing: Attempted to kill init!

Pid: 1, comm: swapper Tainted: G D W 2.6.25-mm1 #1

Call Trace:
 [<ffffffff80236716>] panic+0xb2/0x187
 [<ffffffff802547c7>] ? blocking_notifier_call_chain+0x24/0x42
 [<ffffffff8023a5b7>] do_exit+0x772/0x7eb
 [<ffffffff8020cd1f>] oops_end+0x9a/0x9f
 [<ffffffff80224349>] do_page_fault+0x61d/0x7c4
 [<ffffffff802f31da>] ? proc_register+0x104/0x1b0
 [<ffffffff805a51f9>] error_exit+0x0/0x51
 [<ffffffff8023630d>] ? __stack_chk_test+0x33/0x7b
 [<ffffffff80740ff0>] ? kernel_init+0x16d/0x30d
 [<ffffffff8020c7b8>] ? child_rip+0xa/0x12
 [<ffffffff80740e83>] ? kernel_init+0x0/0x30d
 [<ffffffff8020c7ae>] ? child_rip+0x0/0x12

Rebooting in 30 seconds..
----------

gcc version 4.2.3 (Gentoo 4.2.3 p1.0)

I have put the config and full dmesg of 2.6.25-mm1 both working and not working, up at http://www.reub.net/files/kernel/2.6.25-mm1/

It is the exact same oops with 2.6.24-rc8-mm1.

Reuben

Ingo Molnar

2008-04-18 13:37 UTC #3 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

Quoted 1 line · Reuben Farrelly

is enough to prevent my system booting, viz:

hm, does it boot up fine with the attached patch and stackprotector enabled? It appears that your system got to the self-test so stackprotector is working mostly - it's just that the self-test went wrong.

Ingo

----------------->
Subject: x86: disable stackprotector selftest
From: Ingo Molnar <[email protected]>
Date: Fri Apr 18 15:21:49 CEST 2008

Signed-off-by Ingo Molnar <[email protected]>

---
 kernel/panic.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

kernel/panic.c +1 -1

Index: linux-x86.q/kernel/panic.c
===================================================================
@@ -394,5 +394,5 @@ void __stack_chk_fail(void)
 }
 EXPORT_SYMBOL(__stack_chk_fail);
 
-late_initcall(__stack_chk_test);
+/* late_initcall(__stack_chk_test); */
 #endif

Arjan van de Ven

2008-04-18 13:52 UTC #4 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

Quoted 10 lines · On Fri, 18 Apr 2008 15:36:45 +0200 Ingo Molnar <[email protected]> wrote:

Quoted 1 line · Reuben Farrelly
is enough to prevent my system booting, viz:
hm, does it boot up fine with the attached patch and stackprotector enabled? It appears that your system got to the self-test so stackprotector is working mostly - it's just that the self-test went wrong.

yes it'll be interesting to see if this is due to the test triggering or due to something else.

Reuben:
I assume your gcc is pretty vanilla and doesn't have weird patches in this area?

Can you do a (with the test enabled in the config/code) do a
make kernel/panic.s
and send me that file? (this allows me to see what code your gcc generated for the test)

Signature · 3 lines

If you want to reach me at my work email, use [email protected]
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org

Reuben Farrelly

2008-04-18 14:41 UTC #5 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

Quoted 6 lines · Arjan van de Ven

Quoted 3 lines · On Fri, 18 Apr 2008 15:36:45 +0200 Ingo Molnar <[email protected]> wrote:
Quoted 1 line · Reuben Farrelly
is enough to prevent my system booting, viz:

Quoted 8 lines

Quoted 2 lines
hm, does it boot up fine with the attached patch and stackprotector enabled? It appears that your system got to the self-test so stackprotector is working mostly - it's just that the self-test went wrong.
yes it'll be interesting to see if this is due to the test triggering or due to something else.
Reuben:
I assume your gcc is pretty vanilla and doesn't have weird patches in this area?

I think so. Well, put it this way... I haven't made any changes to it, this is the standard/current gcc that has been in Gentoo Portage for the last while.

Quoted 3 lines

Can you do a (with the test enabled in the config/code) do a
make kernel/panic.s
and send me that file? (this allows me to see what code your gcc generated for the test)

Done - posted up in the web directory along with the other files (saves the possible grief of MUA mangling).

I'm about to reboot to try Ingo's test also.

Reuben

Reuben Farrelly

2008-04-18 14:49 UTC #6 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

In reply to Ingo Molnar

Quoted 5 lines · Ingo Molnar

Quoted 1 line · Reuben Farrelly
is enough to prevent my system booting, viz:
hm, does it boot up fine with the attached patch and stackprotector enabled? It appears that your system got to the self-test so stackprotector is working mostly - it's just that the self-test went wrong.

It boots up fine with that patch below and:

tornado boot # grep STACKPROTECT /boot/config-2.6.25-mm1-wip CONFIG_CC_STACKPROTECTOR_ALL=y CONFIG_CC_STACKPROTECTOR=y

In fact I'm running with it applied right now and it all seems good so far, so I guess that's confirmation that it is just the test itself which is problematic?

Reuben

Ingo Molnar

2008-04-21 15:06 UTC #7 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

Quoted 7 lines · Reuben Farrelly

Quoted 4 lines
hm, does it boot up fine with the attached patch and stackprotector enabled? It appears that your system got to the self-test so stackprotector is working mostly - it's just that the self-test went wrong.
It boots up fine with that patch below and:

Quoted 7 lines

tornado boot # grep STACKPROTECT /boot/config-2.6.25-mm1-wip CONFIG_CC_STACKPROTECTOR_ALL=y CONFIG_CC_STACKPROTECTOR=y
In fact I'm running with it applied right now and it all seems good so far, so I guess that's confirmation that it is just the test itself which is problematic?

yeah. Arjan - any new patches to try that might fix the bootup test?

Ingo

Arjan van de Ven

2008-04-22 01:49 UTC #8 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

Quoted 14 lines · On Mon, 21 Apr 2008 17:06:04 +0200 Ingo Molnar <[email protected]> wrote:

Quoted 11 lines · Reuben Farrelly
Quoted 4 lines
hm, does it boot up fine with the attached patch and stackprotector enabled? It appears that your system got to the self-test so stackprotector is working mostly - it's just that the self-test went wrong.
It boots up fine with that patch below and:
tornado boot # grep STACKPROTECT /boot/config-2.6.25-mm1-wip CONFIG_CC_STACKPROTECTOR_ALL=y CONFIG_CC_STACKPROTECTOR=y

Quoted 6 lines

Quoted 3 lines
In fact I'm running with it applied right now and it all seems good so far, so I guess that's confirmation that it is just the test itself which is problematic?
yeah. Arjan - any new patches to try that might fix the bootup test?

I've looked at the disassembly and compared it to mine, and the gcc is doing something... rather unexpected.
The only thing I can think of is the patch below, it should make it a ton more robust...

From: Arjan van de Ven <[email protected]>
Subject: x86: be more conversative about the stack-protector test

This patch makes the stack-protector self-test more robust against weird stack layouts; rather than assuming that a local variable is layed out in a certain way, we first check this against the known canary value (before we poison it).

Signed-off-by Arjan van de Ven <[email protected]>

kernel/panic.c +4 -1

index c92c1e2..b4a6a05 100644
@@ -351,7 +351,10 @@ static noinline void __stack_chk_test_func(void)
 	}
 #endif
 	barrier();
-	memset(&foo, 0, 2*sizeof(foo)); /* deliberate buffer overflow */
+	if (current->stack_canary == *(((unsigned long *)&foo)+1))
+		*(((unsigned long *)&foo)+1) = 0;
+	else
+		printk(KERN_ERR "No -ftack-protector canary found\n");
 	barrier();
 }

[email protected]

2008-04-22 02:05 UTC #9 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

On Mon, 21 Apr 2008 18:48:59 PDT, Arjan van de Ven said:

Typo alert?

kernel/panic.c static noinline void __stack_chk_test_func(void) · line 357

Quoted 13 lines

kernel/panic.c +4 -1

index c92c1e2..b4a6a05 100644
@@ -351,7 +351,10 @@ static noinline void __stack_chk_test_func(void)
 	}
 #endif
 	barrier();
-	memset(&foo, 0, 2*sizeof(foo)); /* deliberate buffer overflow */
+	if (current->stack_canary == *(((unsigned long *)&foo)+1))
+		*(((unsigned long *)&foo)+1) = 0;
+	else
+		printk(KERN_ERR "No -ftack-protector canary found\n");

-fstack?

Quoted 2 lines

kernel/panic.c +0 -0
 	barrier();
 }

Ingo Molnar

2008-04-22 08:34 UTC #10 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

kernel/panic.c static noinline void __stack_chk_test_func(void) · line 357

Quoted 11 lines · Arjan van de Ven

Quoted 1 line
yeah. Arjan - any new patches to try that might fix the bootup test?
I've looked at the disassembly and compared it to mine, and the gcc is doing something... rather unexpected. The only thing I can think of is the patch below, it should make it a ton more robust...
kernel/panic.c +4 -1
-	memset(&foo, 0, 2*sizeof(foo)); /* deliberate buffer overflow */
+	if (current->stack_canary == *(((unsigned long *)&foo)+1))
+		*(((unsigned long *)&foo)+1) = 0;
+	else
+		printk(KERN_ERR "No -ftack-protector canary found\n");

ok, i queued this up. (with the typo that Valdis noticed fixed)

but ... this bug needs to be figured out, not worked around.

Ingo

Arjan van de Ven

2008-04-22 14:30 UTC #11 lore raw

Re: StackProtector Oopses - Re: 2.6.25-mm1

Quoted 20 lines · On Tue, 22 Apr 2008 10:34:08 +0200 Ingo Molnar <[email protected]> wrote:

Quoted 13 lines · Arjan van de Ven
Quoted 2 lines
yeah. Arjan - any new patches to try that might fix the bootup test?
I've looked at the disassembly and compared it to mine, and the gcc is doing something... rather unexpected. The only thing I can think of is the patch below, it should make it a ton more robust...
(quoted diff) +0 -1
-	memset(&foo, 0, 2*sizeof(foo)); /* deliberate buffer
overflow */
kernel/panic.c +4 -0
+	if (current->stack_canary == *(((unsigned long *)&foo)+1))
+		*(((unsigned long *)&foo)+1) = 0;
+	else
+		printk(KERN_ERR "No -ftack-protector canary
Quoted 1 line
found\n");
ok, i queued this up. (with the typo that Valdis noticed fixed)

Quoted 1 line

but ... this bug needs to be figured out, not worked around.

well what I figured out was that the stack layout was "different". Why/how I don't know, but being more robust against that is a good idea in general.

Signature · 3 lines

If you want to reach me at my work email, use [email protected]
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org

Randy Dunlap

2008-04-18 16:41 UTC #12 lore raw

Re: 2.6.25-mm1 (build error: driver core)

Quoted 2 lines · On Fri, 18 Apr 2008 01:47:57 -0700 Andrew Morton wrote:

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

with CONFIG_BLOCK=n:

linux-2.6.25-mm1/drivers/base/core.c: In function 'device_to_dev_kobj':
linux-2.6.25-mm1/drivers/base/core.c:768: error: 'block_class' undeclared (first use in this function) make[3]: *** [drivers/base/core.o] Error 1

---
~Randy

Greg KH

2008-04-18 16:59 UTC #13 lore raw

Re: 2.6.25-mm1 (build error: driver core)

Quoted 10 lines · Randy Dunlap

Quoted 2 lines · On Fri, 18 Apr 2008 01:47:57 -0700 Andrew Morton wrote:
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
with CONFIG_BLOCK=n:
linux-2.6.25-mm1/drivers/base/core.c: In function 'device_to_dev_kobj':
linux-2.6.25-mm1/drivers/base/core.c:768: error: 'block_class' undeclared (first use in this function) make[3]: *** [drivers/base/core.o] Error 1

Ah, more fun caused by Dan's /sys/dev/... patch.

Dan, this is causing a lot of problems, I'm going to drop it for now until the build and run-time errors get resolved.

thanks,

greg k-h

Dan Williams

2008-04-18 18:39 UTC #14 lore raw

Re: 2.6.25-mm1 (build error: driver core)

Quoted 8 lines · Greg KH

Quoted 7 lines · Randy Dunlap
Quoted 2 lines · On Fri, 18 Apr 2008 01:47:57 -0700 Andrew Morton wrote:
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
with CONFIG_BLOCK=n:

Quoted 9 lines

Quoted 3 lines
linux-2.6.25-mm1/drivers/base/core.c: In function 'device_to_dev_kobj':
linux-2.6.25-mm1/drivers/base/core.c:768: error: 'block_class' undeclared (first use in this function) make[3]: *** [drivers/base/core.o] Error 1
Ah, more fun caused by Dan's /sys/dev/... patch.
Dan, this is causing a lot of problems, I'm going to drop it for now until the build and run-time errors get resolved.

Ok, I'll have an updated version with Kay's duplicate-entry fix and a resolution for CONFIG_BLOCK=n shortly...

Quoted 4 lines

thanks,
greg k-h

Regards,
Dan

Randy Dunlap

2008-04-18 16:48 UTC #15 lore raw

Re: 2.6.25-mm1 (build error: trace selftest)

Quoted 2 lines · On Fri, 18 Apr 2008 01:47:57 -0700 Andrew Morton wrote:

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

with CONFIG_RT_MUTEXES=N:

In file included from linux-2.6.25-mm1/kernel/trace/trace.c:2432:
linux-2.6.25-mm1/kernel/trace/trace_selftest.c: In function 'trace_wakeup_test_thread':
linux-2.6.25-mm1/kernel/trace/trace_selftest.c:413: error: implicit declaration of function 'rt_mutex_setprio' make[3]: *** [kernel/trace/trace.o] Error 1

---
~Randy

[email protected]

2008-04-18 20:14 UTC #16 lore raw

Quoted 1 line · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

I may have reported this same one previously against an earlier -mm, or somebody did, or something... ;)

x86_64 kernel, Core2 Duo T7200, Dell Latitude D820 laptop...

[    0.060388] ACPI: Core revision 20080321
[    0.070079] ------------[ cut here ]------------
[    0.070082] WARNING: at arch/x86/kernel/genapic_64.c:86 read_apic_id+0x41/0x7c()
[    0.070085] Modules linked in:
[    0.070089] Pid: 1, comm: swapper Not tainted 2.6.25-mm1 #1
[    0.070091]
[    0.070092] Call Trace:
[    0.070099]  [<ffffffff8023c702>] warn_on_slowpath+0x67/0xb7
[    0.070105]  [<ffffffff805d096e>] ? _spin_lock+0x25/0x54
[    0.070111]  [<ffffffff8021e399>] ? __cpus_weight+0x4b/0x68
[    0.070114]  [<ffffffff802245c1>] read_apic_id+0x41/0x7c
[    0.070119]  [<ffffffff807a0ded>] verify_local_APIC+0xb4/0x177
[    0.070123]  [<ffffffff805d4153>] ? sub_preempt_count+0x44/0x6e
[    0.070126]  [<ffffffff8079faf4>] native_smp_prepare_cpus+0x238/0x36a
[    0.070129]  [<ffffffff805d4153>] ? sub_preempt_count+0x44/0x6e
[    0.070134]  [<ffffffff80794712>] kernel_init+0x69/0x2a1
[    0.070137]  [<ffffffff805d0d42>] ? _spin_unlock_irq+0x43/0x62
[    0.070143]  [<ffffffff802366b7>] ? finish_task_switch+0x3e/0xb4
[    0.070147]  [<ffffffff8020d6e8>] child_rip+0xa/0x12
[    0.070150]  [<ffffffff8020cdd0>] ? restore_args+0x0/0x30
[    0.070154]  [<ffffffff807946a9>] ? kernel_init+0x0/0x2a1
[    0.070157]  [<ffffffff8020d6de>] ? child_rip+0x0/0x12
[    0.070159]
[    0.070166] ---[ end trace a7919e7f17c0a725 ]---
[    0.070169] ------------[ cut here ]------------
[    0.070171] WARNING: at arch/x86/kernel/genapic_64.c:86 read_apic_id+0x41/0x7c()
[    0.070174] Modules linked in:
[    0.070177] Pid: 1, comm: swapper Tainted: G        W 2.6.25-mm1 #1
[    0.070179]
[    0.070179] Call Trace:
[    0.070182]  [<ffffffff8023c702>] warn_on_slowpath+0x67/0xb7
[    0.070186]  [<ffffffff805d096e>] ? _spin_lock+0x25/0x54
[    0.070189]  [<ffffffff8021e399>] ? __cpus_weight+0x4b/0x68
[    0.070192]  [<ffffffff802245c1>] read_apic_id+0x41/0x7c
[    0.070195]  [<ffffffff807a0e1f>] verify_local_APIC+0xe6/0x177
[    0.070199]  [<ffffffff805d4153>] ? sub_preempt_count+0x44/0x6e
[    0.070202]  [<ffffffff8079faf4>] native_smp_prepare_cpus+0x238/0x36a
[    0.070205]  [<ffffffff805d4153>] ? sub_preempt_count+0x44/0x6e
[    0.070209]  [<ffffffff80794712>] kernel_init+0x69/0x2a1
[    0.070212]  [<ffffffff805d0d42>] ? _spin_unlock_irq+0x43/0x62
[    0.070215]  [<ffffffff802366b7>] ? finish_task_switch+0x3e/0xb4
[    0.070218]  [<ffffffff8020d6e8>] child_rip+0xa/0x12
[    0.070221]  [<ffffffff8020cdd0>] ? restore_args+0x0/0x30
[    0.070225]  [<ffffffff807946a9>] ? kernel_init+0x0/0x2a1
[    0.070228]  [<ffffffff8020d6de>] ? child_rip+0x0/0x12
[    0.070230]
[    0.070231] ---[ end trace a7919e7f17c0a725 ]---
[    0.081710] CPU0: Intel(R) Core(TM)2 CPU         T7200  @ 2.00GHz stepping 06

Other than that, it's lived for an hour without barfing yet, but I won't have a chance to kick the tires and try stupid sysadmin tricks like I usually do until Sunday-ish... Have a good weekend...

Alexey Dobriyan

2008-04-18 22:14 UTC #17 lore raw

2.6.25-mm1: orphaned files after build

At least the following files aren't removed by "make mrproper":

Module.markers
arch/x86/kernel/acpi/realmode/wakeup.lds
crypto/.tmp_aes_generic.ver
fs/.tmp_buffer.ver
kernel/time/.tmp_timekeeping.ver

Noticed by "git-ls-files -o".

*.ver , I think, remain after abruptdly terminated build.

Joseph Fannin

2008-04-19 02:14 UTC #18 lore raw

http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

Quoted 2 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

I've been seeing the following backtraces since 2.6.25-rc8-mm1 -- at least, since that's the earliest -mm I've built in a while. I don't get the same in mainline.

No idea who to CC: I've sat on this report long enough.

I'm going to send a few different reports in separate mails, so I'll put my dmesg and .config up on a server:

[  451.915553] sysfs: duplicate filename 'pcspkr' can not be created
[  451.915731] ------------[ cut here ]------------
[  451.915851] WARNING: at fs/sysfs/dir.c:427 sysfs_add_one+0x85/0xe0()
[  451.915981] Modules linked in: snd_pcsp(+) ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx  scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  451.918960] Pid: 2740, comm: modprobe Tainted: G        W 2.6.25-mm1 #7
[  451.929271]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  451.929500]  [<c0132400>] ? vprintk+0x2f0/0x4a0
[  451.929723]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.929918]  [<c01c6a7a>] ? ifind+0x4a/0xa0
[  451.930126]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.930334]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  451.930534]  [<c01325d0>] ? printk+0x20/0x30
[  451.930727]  [<c01fcc45>] sysfs_add_one+0x85/0xe0
[  451.930900]  [<c01fd89e>] create_dir+0x4e/0xb0
[  451.931064]  [<c01fd930>] sysfs_create_dir+0x30/0x50
[  451.931291]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.931485]  [<c023dac6>] kobject_add_internal+0xb6/0x190
[  451.931656]  [<c023dc22>] ? kobject_set_name_vargs+0x32/0x40
[  451.931857]  [<c023dc8a>] kobject_add_varg+0x5a/0x60
[  451.932022]  [<c023e12f>] kobject_init_and_add+0x2f/0x40
[  451.932188]  [<c02a3e44>] bus_add_driver+0x94/0x250
[  451.932364]  [<c02a5062>] driver_register+0x42/0xf0
[  451.932533]  [<c02a6c56>] platform_driver_register+0x66/0x70
[  451.932702]  [<cc04b02a>] pcsp_init+0x2a/0x2c [snd_pcsp]
[  451.932877]  [<c015e137>] sys_init_module+0x87/0x1a0
[  451.933043]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.933246]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  451.933453]  [<c0104077>] sysenter_past_esp+0x78/0xc5
[  451.933636]  =======================
[  451.933744] ---[ end trace 4eaa2a86a8e2da22 ]---
[  451.933905] kobject_add_internal failed for pcspkr with -EEXIST, don't try to register things with the same name in the same directory.
[  451.934130] Pid: 2740, comm: modprobe Tainted: G        W 2.6.25-mm1 #7
[  451.934269]  [<c023db53>] kobject_add_internal+0x143/0x190
[  451.934439]  [<c023dc22>] ? kobject_set_name_vargs+0x32/0x40
[  451.934641]  [<c023dc8a>] kobject_add_varg+0x5a/0x60
[  451.934805]  [<c023e12f>] kobject_init_and_add+0x2f/0x40
[  451.934969]  [<c02a3e44>] bus_add_driver+0x94/0x250
[  451.979314]  [<c02a5062>] driver_register+0x42/0xf0
[  451.979535]  [<c02a6c56>] platform_driver_register+0x66/0x70
[  451.979709]  [<cc04b02a>] pcsp_init+0x2a/0x2c [snd_pcsp]
[  451.979888]  [<c015e137>] sys_init_module+0x87/0x1a0
[  451.980060]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.980264]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  451.980479]  [<c0104077>] sysenter_past_esp+0x78/0xc5
[  451.980660]  =======================

Anyone?

--
Joseph Fannin
[email protected]

Andrew Morton

author maintainer

2008-04-19 03:02 UTC #19 lore raw

Quoted 9 lines · On Fri, 18 Apr 2008 22:13:43 -0400 Joseph Fannin <[email protected]> wrote:

Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
I've been seeing the following backtraces since 2.6.25-rc8-mm1 -- at least, since that's the earliest -mm I've built in a while. I don't get the same in mainline.
No idea who to CC:

I have a few ideas.

Quoted 1 line

I've sat on this report long enough.

Thanks for the report.

Quoted 26 lines

I'm going to send a few different reports in separate mails, so I'll put my dmesg and .config up on a server:

http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

[  451.915553] sysfs: duplicate filename 'pcspkr' can not be created
[  451.915731] ------------[ cut here ]------------
[  451.915851] WARNING: at fs/sysfs/dir.c:427 sysfs_add_one+0x85/0xe0()
[  451.915981] Modules linked in: snd_pcsp(+) ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx  scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  451.918960] Pid: 2740, comm: modprobe Tainted: G        W 2.6.25-mm1 #7
[  451.929271]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  451.929500]  [<c0132400>] ? vprintk+0x2f0/0x4a0
[  451.929723]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.929918]  [<c01c6a7a>] ? ifind+0x4a/0xa0
[  451.930126]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.930334]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  451.930534]  [<c01325d0>] ? printk+0x20/0x30
[  451.930727]  [<c01fcc45>] sysfs_add_one+0x85/0xe0
[  451.930900]  [<c01fd89e>] create_dir+0x4e/0xb0
[  451.931064]  [<c01fd930>] sysfs_create_dir+0x30/0x50
[  451.931291]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.931485]  [<c023dac6>] kobject_add_internal+0xb6/0x190
[  451.931656]  [<c023dc22>] ? kobject_set_name_vargs+0x32/0x40
[  451.931857]  [<c023dc8a>] kobject_add_varg+0x5a/0x60
[  451.932022]  [<c023e12f>] kobject_init_and_add+0x2f/0x40

Quoted 8 lines

[  451.932188]  [<c02a3e44>] bus_add_driver+0x94/0x250
[  451.932364]  [<c02a5062>] driver_register+0x42/0xf0
[  451.932533]  [<c02a6c56>] platform_driver_register+0x66/0x70
[  451.932702]  [<cc04b02a>] pcsp_init+0x2a/0x2c [snd_pcsp]
[  451.932877]  [<c015e137>] sys_init_module+0x87/0x1a0
[  451.933043]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.933246]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  451.933453]  [<c0104077>] sysenter_past_esp+0x78/0xc5

Yes, there have been lots of these lately. I expect some of them _will_ go into mainline and they'll then slowly get weeded out.

Dmitry Torokhov

maintainer

2008-04-19 04:14 UTC #20 lore raw

Quoted 53 lines · Andrew Morton

Quoted 9 lines · On Fri, 18 Apr 2008 22:13:43 -0400 Joseph Fannin <[email protected]> wrote:

Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
I've been seeing the following backtraces since 2.6.25-rc8-mm1 -- at least, since that's the earliest -mm I've built in a while. I don't get the same in mainline.
No idea who to CC:

I have a few ideas.

Quoted 1 line

I've sat on this report long enough.

Thanks for the report.

Quoted 26 lines

I'm going to send a few different reports in separate mails, so I'll put my dmesg and .config up on a server:

http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

[  451.915553] sysfs: duplicate filename 'pcspkr' can not be created
[  451.915731] ------------[ cut here ]------------
[  451.915851] WARNING: at fs/sysfs/dir.c:427 sysfs_add_one+0x85/0xe0()
[  451.915981] Modules linked in: snd_pcsp(+) ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx  scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  451.918960] Pid: 2740, comm: modprobe Tainted: G        W 2.6.25-mm1 #7
[  451.929271]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  451.929500]  [<c0132400>] ? vprintk+0x2f0/0x4a0
[  451.929723]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.929918]  [<c01c6a7a>] ? ifind+0x4a/0xa0
[  451.930126]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.930334]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  451.930534]  [<c01325d0>] ? printk+0x20/0x30
[  451.930727]  [<c01fcc45>] sysfs_add_one+0x85/0xe0
[  451.930900]  [<c01fd89e>] create_dir+0x4e/0xb0
[  451.931064]  [<c01fd930>] sysfs_create_dir+0x30/0x50
[  451.931291]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.931485]  [<c023dac6>] kobject_add_internal+0xb6/0x190
[  451.931656]  [<c023dc22>] ? kobject_set_name_vargs+0x32/0x40
[  451.931857]  [<c023dc8a>] kobject_add_varg+0x5a/0x60
[  451.932022]  [<c023e12f>] kobject_init_and_add+0x2f/0x40

Quoted 8 lines

[  451.932188]  [<c02a3e44>] bus_add_driver+0x94/0x250
[  451.932364]  [<c02a5062>] driver_register+0x42/0xf0
[  451.932533]  [<c02a6c56>] platform_driver_register+0x66/0x70
[  451.932702]  [<cc04b02a>] pcsp_init+0x2a/0x2c [snd_pcsp]
[  451.932877]  [<c015e137>] sys_init_module+0x87/0x1a0
[  451.933043]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.933246]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  451.933453]  [<c0104077>] sysenter_past_esp+0x78/0xc5

Quoted 3 lines

Yes, there have been lots of these lately. I expect some of them _will_ go into mainline and they'll then slowly get weeded out.

It looks like it is coming from snd_pcsp module from alsa tree.

Cool things there:

+#ifdef CONFIG_DEBUG_PAGEALLOC
+ /* Well, CONFIG_DEBUG_PAGEALLOC makes the sound horrible. Lets
alert */
+ printk(KERN_WARNING
+ "PCSP: Warning, CONFIG_DEBUG_PAGEALLOC is enabled!\n"
+ "You have to disable it if you want to use the PC-Speaker
"

+	       "driver.\n"
+	       "Unless it is disabled, enjoy the horrible, distorted "
+	       "and crackling noise.\n");

+#endif

Signature · 1 line

Dmitry

Andrew Morton

author maintainer

2008-04-19 04:30 UTC #21 lore raw

Quoted 54 lines · On Sat, 19 Apr 2008 00:14:29 -0400 Dmitry Torokhov <[email protected]>…

Quoted 53 lines · Andrew Morton

Quoted 9 lines · On Fri, 18 Apr 2008 22:13:43 -0400 Joseph Fannin <[email protected]> wrote:

Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
I've been seeing the following backtraces since 2.6.25-rc8-mm1 -- at least, since that's the earliest -mm I've built in a while. I don't get the same in mainline.
No idea who to CC:

I have a few ideas.

Quoted 1 line

I've sat on this report long enough.

Thanks for the report.

Quoted 34 lines

I'm going to send a few different reports in separate mails, so I'll put my dmesg and .config up on a server:

http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

[  451.915553] sysfs: duplicate filename 'pcspkr' can not be created
[  451.915731] ------------[ cut here ]------------
[  451.915851] WARNING: at fs/sysfs/dir.c:427 sysfs_add_one+0x85/0xe0()
[  451.915981] Modules linked in: snd_pcsp(+) ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx  scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  451.918960] Pid: 2740, comm: modprobe Tainted: G        W 2.6.25-mm1 #7
[  451.929271]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  451.929500]  [<c0132400>] ? vprintk+0x2f0/0x4a0
[  451.929723]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.929918]  [<c01c6a7a>] ? ifind+0x4a/0xa0
[  451.930126]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.930334]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  451.930534]  [<c01325d0>] ? printk+0x20/0x30
[  451.930727]  [<c01fcc45>] sysfs_add_one+0x85/0xe0
[  451.930900]  [<c01fd89e>] create_dir+0x4e/0xb0
[  451.931064]  [<c01fd930>] sysfs_create_dir+0x30/0x50
[  451.931291]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.931485]  [<c023dac6>] kobject_add_internal+0xb6/0x190
[  451.931656]  [<c023dc22>] ? kobject_set_name_vargs+0x32/0x40
[  451.931857]  [<c023dc8a>] kobject_add_varg+0x5a/0x60
[  451.932022]  [<c023e12f>] kobject_init_and_add+0x2f/0x40
[  451.932188]  [<c02a3e44>] bus_add_driver+0x94/0x250
[  451.932364]  [<c02a5062>] driver_register+0x42/0xf0
[  451.932533]  [<c02a6c56>] platform_driver_register+0x66/0x70
[  451.932702]  [<cc04b02a>] pcsp_init+0x2a/0x2c [snd_pcsp]
[  451.932877]  [<c015e137>] sys_init_module+0x87/0x1a0
[  451.933043]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.933246]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  451.933453]  [<c0104077>] sysenter_past_esp+0x78/0xc5

Quoted 5 lines

Quoted 3 lines
Yes, there have been lots of these lately. I expect some of them _will_ go into mainline and they'll then slowly get weeded out.
It looks like it is coming from snd_pcsp module from alsa tree.

Ah, OK. More cc's.

Quoted 10 lines

Cool things there:

(quoted diff) +2 -0

+#ifdef CONFIG_DEBUG_PAGEALLOC
+	/* Well, CONFIG_DEBUG_PAGEALLOC makes the sound horrible. Lets

alert */

(quoted diff) +3 -0

+	printk(KERN_WARNING
+	       "PCSP: Warning, CONFIG_DEBUG_PAGEALLOC is enabled!\n"
+	       "You have to disable it if you want to use the PC-Speaker

Review

Quoted 4 lines

(quoted diff) +4 -0

+	       "driver.\n"
+	       "Unless it is disabled, enjoy the horrible, distorted "
+	       "and crackling noise.\n");
+#endif

heh.

CONFIG_DEBUG_PAGEALLOC is a very heavy consumer of CPU cycles. I'm not surprised that it would whack what I presume to be a very latency-sensitive driver.

Joseph Fannin

2008-04-19 06:34 UTC #22 lore raw

Quoted 81 lines · Andrew Morton

Quoted 54 lines · On Sat, 19 Apr 2008 00:14:29 -0400 Dmitry Torokhov <[email protected]>…
Quoted 53 lines · Andrew Morton
Quoted 9 lines · On Fri, 18 Apr 2008 22:13:43 -0400 Joseph Fannin <[email protected]> wrote:
Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
I've been seeing the following backtraces since 2.6.25-rc8-mm1 -- at least, since that's the earliest -mm I've built in a while. I don't get the same in mainline.
No idea who to CC:
I have a few ideas.
Quoted 1 line
I've sat on this report long enough.
Thanks for the report.
Quoted 34 lines
I'm going to send a few different reports in separate mails, so I'll put my dmesg and .config up on a server:
http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt
[  451.915553] sysfs: duplicate filename 'pcspkr' can not be created
[  451.915731] ------------[ cut here ]------------
[  451.915851] WARNING: at fs/sysfs/dir.c:427 sysfs_add_one+0x85/0xe0()
[  451.915981] Modules linked in: snd_pcsp(+) ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx  scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  451.918960] Pid: 2740, comm: modprobe Tainted: G        W 2.6.25-mm1 #7
[  451.929271]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  451.929500]  [<c0132400>] ? vprintk+0x2f0/0x4a0
[  451.929723]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.929918]  [<c01c6a7a>] ? ifind+0x4a/0xa0
[  451.930126]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.930334]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  451.930534]  [<c01325d0>] ? printk+0x20/0x30
[  451.930727]  [<c01fcc45>] sysfs_add_one+0x85/0xe0
[  451.930900]  [<c01fd89e>] create_dir+0x4e/0xb0
[  451.931064]  [<c01fd930>] sysfs_create_dir+0x30/0x50
[  451.931291]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.931485]  [<c023dac6>] kobject_add_internal+0xb6/0x190
[  451.931656]  [<c023dc22>] ? kobject_set_name_vargs+0x32/0x40
[  451.931857]  [<c023dc8a>] kobject_add_varg+0x5a/0x60
[  451.932022]  [<c023e12f>] kobject_init_and_add+0x2f/0x40
[  451.932188]  [<c02a3e44>] bus_add_driver+0x94/0x250
[  451.932364]  [<c02a5062>] driver_register+0x42/0xf0
[  451.932533]  [<c02a6c56>] platform_driver_register+0x66/0x70
[  451.932702]  [<cc04b02a>] pcsp_init+0x2a/0x2c [snd_pcsp]
[  451.932877]  [<c015e137>] sys_init_module+0x87/0x1a0
[  451.933043]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.933246]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  451.933453]  [<c0104077>] sysenter_past_esp+0x78/0xc5
Quoted 5 lines
Quoted 3 lines
Yes, there have been lots of these lately. I expect some of them _will_ go into mainline and they'll then slowly get weeded out.
It looks like it is coming from snd_pcsp module from alsa tree.
Ah, OK. More cc's.
Quoted 10 lines
Cool things there:
(quoted diff) +2 -0
+#ifdef CONFIG_DEBUG_PAGEALLOC
+	/* Well, CONFIG_DEBUG_PAGEALLOC makes the sound horrible. Lets
alert */
(quoted diff) +3 -0
+	printk(KERN_WARNING
+	       "PCSP: Warning, CONFIG_DEBUG_PAGEALLOC is enabled!\n"
+	       "You have to disable it if you want to use the PC-Speaker
"
Review
Quoted 4 lines
(quoted diff) +4 -0
+	       "driver.\n"
+	       "Unless it is disabled, enjoy the horrible, distorted "
+	       "and crackling noise.\n");
+#endif
heh.

Quoted 3 lines

CONFIG_DEBUG_PAGEALLOC is a very heavy consumer of CPU cycles. I'm not surprised that it would whack what I presume to be a very latency-sensitive driver.

Um...

[jhf@Susa ~]$ uname -a
Linux Susa 2.6.25-mm1 #7 SMP PREEMPT Fri Apr 18 17:05:14 EDT 2008 i686 GNU/Linux
[jhf@Susa ~]$ zgrep PAGEALLOC /proc/config.gz
# CONFIG_DEBUG_PAGEALLOC is not set
[jhf@Susa ~]$

I thought that might have snuck in as =y, but it didn't.

--
Joseph Fannin
[email protected]

Takashi Iwai

2008-04-21 11:08 UTC #23 lore raw

[Added snd-pcsp author to Cc]

At Fri, 18 Apr 2008 21:29:34 -0700,

Quoted 48 lines · Andrew Morton

Quoted 45 lines · On Sat, 19 Apr 2008 00:14:29 -0400 Dmitry Torokhov <[email protected]>…
Quoted 44 lines · Andrew Morton
Quoted 9 lines · On Fri, 18 Apr 2008 22:13:43 -0400 Joseph Fannin <[email protected]> wrote:
Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
I've been seeing the following backtraces since 2.6.25-rc8-mm1 -- at least, since that's the earliest -mm I've built in a while. I don't get the same in mainline.
No idea who to CC:
I have a few ideas.
Quoted 1 line
I've sat on this report long enough.
Thanks for the report.
Quoted 26 lines
I'm going to send a few different reports in separate mails, so I'll put my dmesg and .config up on a server:
http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt
[  451.915553] sysfs: duplicate filename 'pcspkr' can not be created
[  451.915731] ------------[ cut here ]------------
[  451.915851] WARNING: at fs/sysfs/dir.c:427 sysfs_add_one+0x85/0xe0()
[  451.915981] Modules linked in: snd_pcsp(+) ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx  scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  451.918960] Pid: 2740, comm: modprobe Tainted: G        W 2.6.25-mm1 #7
[  451.929271]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  451.929500]  [<c0132400>] ? vprintk+0x2f0/0x4a0
[  451.929723]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.929918]  [<c01c6a7a>] ? ifind+0x4a/0xa0
[  451.930126]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.930334]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  451.930534]  [<c01325d0>] ? printk+0x20/0x30
[  451.930727]  [<c01fcc45>] sysfs_add_one+0x85/0xe0
[  451.930900]  [<c01fd89e>] create_dir+0x4e/0xb0
[  451.931064]  [<c01fd930>] sysfs_create_dir+0x30/0x50
[  451.931291]  [<c0356adc>] ? _spin_unlock+0x2c/0x50
[  451.931485]  [<c023dac6>] kobject_add_internal+0xb6/0x190
[  451.931656]  [<c023dc22>] ? kobject_set_name_vargs+0x32/0x40
[  451.931857]  [<c023dc8a>] kobject_add_varg+0x5a/0x60
[  451.932022]  [<c023e12f>] kobject_init_and_add+0x2f/0x40

Quoted 8 lines

[  451.932188]  [<c02a3e44>] bus_add_driver+0x94/0x250
[  451.932364]  [<c02a5062>] driver_register+0x42/0xf0
[  451.932533]  [<c02a6c56>] platform_driver_register+0x66/0x70
[  451.932702]  [<cc04b02a>] pcsp_init+0x2a/0x2c [snd_pcsp]
[  451.932877]  [<c015e137>] sys_init_module+0x87/0x1a0
[  451.933043]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  451.933246]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  451.933453]  [<c0104077>] sysenter_past_esp+0x78/0xc5

Seems that snd-pcsp registers as "pcspkr", which is identical with input pc-speaker driver. Does the patch below fix the problem?

Takashi

---

sound/drivers/pcsp/pcsp.c +2 -2

diff -r e8f61dd0b153 sound/drivers/pcsp/pcsp.c
@@ -21,7 +21,7 @@
 MODULE_DESCRIPTION("PC-Speaker driver");
 MODULE_LICENSE("GPL");
 MODULE_SUPPORTED_DEVICE("{{PC-Speaker, pcsp}}");
-MODULE_ALIAS("platform:pcspkr");
+MODULE_ALIAS("platform:snd_pcsp");
 
 static int index = SNDRV_DEFAULT_IDX1;	/* Index 0-MAX */
 static char *id = SNDRV_DEFAULT_STR1;	/* ID for this card */

@@ -214,7 +214,7 @@
 
 static struct platform_driver pcsp_platform_driver = {
 	.driver		= {
-		.name	= "pcspkr",
+		.name	= "snd_pcsp",
 		.owner	= THIS_MODULE,
 	},
 	.probe		= pcsp_probe,

Stas Sergeev

Acked-by Takashi Iwai

2008-04-21 18:25 UTC #24 lore raw

Re: 2.6.25-mm1 (snd-pcsp causes driver conflict)

Hello.

Quoted 1 line · Takashi Iwai

[Added snd-pcsp author to Cc]

Thanks.

Quoted 2 lines

Seems that snd-pcsp registers as "pcspkr", which is identical with input pc-speaker driver. Does the patch below fix the problem?

Actually it does not. The reason is that then it fails to match the platform device, which is created in arch/x86/kernel/pcspeaker.c, with the name of "pcspkr".

But we already had the patch for that in an alsa tree, it probably got forgotten. Here it is:
http://hg-mirror.alsa-project.org/alsa-driver/raw-file/90eeee75052f/utils/patches/pcsp-kernel-2.6.22-01.diff

Also attaching it here.
It simply disables the pcspkr driver in Kconfig. snd-pcsp has the copy of that driver, so that only one driver would drive the device.

Does that fix look good? (presumably acked by Takashi, otherwise the patch wouldn't be in an alsa tree)

---------------

- Prevent pcspkr driver from being built together with snd-pcsp. snd-pcsp fully superceeds pcspkr.
- Update CREDITS file. :)

Signed-off-by Stas Sergeev <[email protected]>

Acked-by Takashi Iwai <[email protected]>

CREDITS +8 -0

diff -urN linux-2.6.24/CREDITS linux-2.6.24-pcsp-kern/CREDITS
@@ -403,6 +403,8 @@
 N: Erik Inge Bolsø
 E: [email protected]
 D: Misc kernel hacks
+D: Updated PC speaker driver for 2.3
+S: Norway
 
 N: Andreas E. Bombe
 E: [email protected]

@@ -3109,6 +3111,12 @@
 S: Sunnyvale, California 94088-4132
 S: USA
 
+N: Stas Sergeev
+E: [email protected]
+D: PCM PC-Speaker driver
+D: misc fixes
+S: Russia
+
 N: Simon Shapiro
 E: [email protected]
 W: http://www.-i-Connect.Net/~shimon

drivers/input/misc/Kconfig +2 -1

diff -urN linux-2.6.24/drivers/input/misc/Kconfig linux-2.6.24-pcsp-kern/drivers/input/misc/Kconfig
@@ -14,7 +14,8 @@
 
 config INPUT_PCSPKR
 	tristate "PC Speaker support"
-	depends on ALPHA || X86 || MIPS || PPC_PREP || PPC_CHRP || PPC_PSERIES
+	depends on (ALPHA || X86 || MIPS || PPC_PREP || PPC_CHRP || \
+		PPC_PSERIES) && SND_PCSP=n
 	help
 	  Say Y here if you want the standard PC Speaker to be used for
 	  bells and whistles.

Takashi Iwai

2008-04-22 10:09 UTC #25 lore raw

Re: 2.6.25-mm1 (snd-pcsp causes driver conflict)

At Mon, 21 Apr 2008 21:44:33 +0400,

Quoted 5 lines · Stas Sergeev

Hello.
Quoted 1 line · Takashi Iwai
[Added snd-pcsp author to Cc]

Quoted 8 lines

Thanks.
Quoted 2 lines
Seems that snd-pcsp registers as "pcspkr", which is identical with input pc-speaker driver. Does the patch below fix the problem?
Actually it does not. The reason is that then it fails to match the platform device, which is created in arch/x86/kernel/pcspeaker.c, with the name of "pcspkr".

Hm, the hardcoded string is no good thing.
It should be defined in the common header if it's used in multiple places.

I'm not 100% certain whether restrictng this in Kconfig is the correct fix. Basically this doesn't stop building both drivers. In theory, we can switch them dynamically.

But, it's the easiest way to avoid unnecessary bugs right now, so let's merge it.

Quoted 4 lines

But we already had the patch for that in an alsa tree, it probably got forgotten. Here it is:
http://hg-mirror.alsa-project.org/alsa-driver/raw-file/90eeee75052f/utils/patches/pcsp-kernel-2.6.22-01.diff

Err, no, this wasn't merged to sound git tree because apparently the patch is to 2.6.22 and you didn't resubmit it...

Quoted 9 lines

Also attaching it here.
It simply disables the pcspkr driver in Kconfig. snd-pcsp has the copy of that driver, so that only one driver would drive the device.
Does that fix look good? (presumably acked by Takashi, otherwise the patch wouldn't be in an alsa tree)

No, I gave no ACK yet. The alsa-driver tree is our playground, and the patch merged to that tree doesn't mean that I approved it for linux-kernel merge.

drivers/input/misc/Kconfig line 18

Quoted 10 lines

diff -urN linux-2.6.24/drivers/input/misc/Kconfig linux-2.6.24-pcsp-kern/drivers/input/misc/Kconfig
drivers/input/misc/Kconfig +2 -1
@@ -14,7 +14,8 @@
 
 config INPUT_PCSPKR
 	tristate "PC Speaker support"
-	depends on ALPHA || X86 || MIPS || PPC_PREP || PPC_CHRP || PPC_PSERIES
+	depends on (ALPHA || X86 || MIPS || PPC_PREP || PPC_CHRP || \
+		PPC_PSERIES) && SND_PCSP=n

I'd rather add a new line with a single "depends on SND_PCSP=n". You see a clear difference in the art of dependencies, one for architectures and one for driver-specific.

thanks,

Takashi

Stas Sergeev

2008-04-22 17:55 UTC #26 lore raw

Re: 2.6.25-mm1 (snd-pcsp causes driver conflict)

Hello.

Quoted 2 lines · Takashi Iwai

Err, no, this wasn't merged to sound git tree because apparently the patch is to 2.6.22 and you didn't resubmit it...

That's simply because the old one
applies fine. :)

Quoted 1 line

No, I gave no ACK yet.

OK, sorry.

Quoted 3 lines

I'd rather add a new line with a single "depends on SND_PCSP=n". You see a clear difference in the art of dependencies, one for architectures and one for driver-specific.

Done.

---
- Update CREDITS with the pc-speaker
driver authors.
- Prevent pcspkr from being built together
with snd-pcsp.

Signed-off-by Stas Sergeev <[email protected]>

CREDITS +8 -0

diff -ur linux-2.6.25/CREDITS linux-2.6.25-pcsp-kern/CREDITS
@@ -403,6 +403,8 @@
 N: Erik Inge Bolsø
 E: [email protected]
 D: Misc kernel hacks
+D: Updated PC speaker driver for 2.3
+S: Norway
 
 N: Andreas E. Bombe
 E: [email protected]

@@ -3116,6 +3118,12 @@
 S: Sunnyvale, California 94088-4132
 S: USA
 
+N: Stas Sergeev
+E: [email protected]
+D: PCM PC-Speaker driver
+D: misc fixes
+S: Russia
+
 N: Simon Shapiro
 E: [email protected]
 W: http://www.-i-Connect.Net/~shimon

drivers/input/misc/Kconfig +1 -0

diff -ur linux-2.6.25/drivers/input/misc/Kconfig linux-2.6.25-pcsp-kern/drivers/input/misc/Kconfig
@@ -15,6 +15,7 @@
 config INPUT_PCSPKR
 	tristate "PC Speaker support"
 	depends on ALPHA || X86 || MIPS || PPC_PREP || PPC_CHRP || PPC_PSERIES
+	depends on SND_PCSP=n
 	help
 	  Say Y here if you want the standard PC Speaker to be used for
 	  bells and whistles.

Takashi Iwai

Applied

2008-04-23 08:55 UTC #27 lore raw

Re: 2.6.25-mm1 (snd-pcsp causes driver conflict)

At Tue, 22 Apr 2008 21:54:46 +0400,

Quoted 17 lines · Stas Sergeev

Hello.
Quoted 2 lines · Takashi Iwai
Err, no, this wasn't merged to sound git tree because apparently the patch is to 2.6.22 and you didn't resubmit it...
That's simply because the old one
applies fine. :)
Quoted 1 line
No, I gave no ACK yet.
OK, sorry.
Quoted 3 lines
I'd rather add a new line with a single "depends on SND_PCSP=n". You see a clear difference in the art of dependencies, one for architectures and one for driver-specific.
Done.

Quoted 7 lines

---
- Update CREDITS with the pc-speaker
driver authors.
- Prevent pcspkr from being built together
with snd-pcsp.
Signed-off-by Stas Sergeev <[email protected]>

Thanks, applied to my git tree.

	git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git

Meanwhile, we need to add a similar depenency to snd-pcsp as well, no?

Takashi

Takashi Iwai

2008-04-23 14:14 UTC #28 lore raw

Re: 2.6.25-mm1 (snd-pcsp causes driver conflict)

At Wed, 23 Apr 2008 10:55:08 +0200,

Quoted 26 lines · I wrote:

At Tue, 22 Apr 2008 21:54:46 +0400,
Quoted 23 lines · Stas Sergeev
Hello.
Quoted 2 lines · Takashi Iwai
Err, no, this wasn't merged to sound git tree because apparently the patch is to 2.6.22 and you didn't resubmit it...
That's simply because the old one
applies fine. :)
Quoted 1 line
No, I gave no ACK yet.
OK, sorry.
Quoted 3 lines
I'd rather add a new line with a single "depends on SND_PCSP=n". You see a clear difference in the art of dependencies, one for architectures and one for driver-specific.
Done.
---
- Update CREDITS with the pc-speaker
driver authors.
- Prevent pcspkr from being built together
with snd-pcsp.

Quoted 6 lines

Quoted 1 line
Signed-off-by Stas Sergeev <[email protected]>
Thanks, applied to my git tree.
	git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
Meanwhile, we need to add a similar depenency to snd-pcsp as well, no?

Fixed on ALSA tree now.

Takashi

Stas Sergeev

2008-04-21 19:46 UTC #29 lore raw

Hi.

Quoted 6 lines · On Fri, 18 Apr 2008 22:13:43 -0400 Joseph Fannin <[email protected]> wrote:

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines · Andrew Morton
Quoted 1 line
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
I've been seeing the following backtraces since 2.6.25-rc8-mm1 -- at least, since that's the earliest -mm I've built in a while. I don't get the same in mainline.

As a quick workaround, say N to INPUT_PCSPKR:

  │ Prompt: PC Speaker support
  │   Location:
  │     -> Device Drivers
  │       -> Input device support
  │         -> Generic input layer (needed for keyboard, mouse, ...)
(INPUT │
  │           -> Miscellaneous devices (INPUT_MISC [=y])

Takashi Iwai

2008-04-21 14:06 UTC #30 lore raw

At Fri, 18 Apr 2008 21:29:34 -0700,

Quoted 18 lines · Andrew Morton

Quoted 11 lines

Cool things there:

(quoted diff) +2 -0

+#ifdef CONFIG_DEBUG_PAGEALLOC
+	/* Well, CONFIG_DEBUG_PAGEALLOC makes the sound horrible. Lets

alert */

(quoted diff) +3 -0

+	printk(KERN_WARNING
+	       "PCSP: Warning, CONFIG_DEBUG_PAGEALLOC is enabled!\n"
+	       "You have to disable it if you want to use the PC-Speaker

Review

Quoted 4 lines

(quoted diff) +4 -0

+	       "driver.\n"
+	       "Unless it is disabled, enjoy the horrible, distorted "
+	       "and crackling noise.\n");
+#endif

heh.

Quoted 3 lines

CONFIG_DEBUG_PAGEALLOC is a very heavy consumer of CPU cycles. I'm not surprised that it would whack what I presume to be a very latency-sensitive driver.

We can add simply a dependncy to Kconfig if this really matters.

Takashi

---

sound/drivers/Kconfig +1 -0

diff -r e8f61dd0b153 sound/drivers/Kconfig
@@ -7,6 +7,7 @@
 config SND_PCSP
 	tristate "Internal PC speaker support"
 	depends on X86_PC && HIGH_RES_TIMERS
+	depends on !DEBUG_PAGEALLOC
 	help
 	  If you don't have a sound card in your computer, you can include a
 	  driver for the PC speaker which allows it to act like a primitive

Stas Sergeev

2008-04-21 17:55 UTC #31 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 4 lines · Takashi Iwai

Quoted 3 lines
CONFIG_DEBUG_PAGEALLOC is a very heavy consumer of CPU cycles. I'm not surprised that it would whack what I presume to be a very latency-sensitive driver.
We can add simply a dependncy to Kconfig if this really matters.

I think this is a bit too heavy-handed. That thing adds a lot of noise to the sound, but it doesn't really prevent the driver from working properly. And perhaps there are other options with the same effect? Also, that would motivate people to optimize DEBUG_PAGEALLOC, while otherwise they wouldn't know. :)

What is the problem with the warning exactly? If it makes problems, we can just remove it and update the help text instead, for example.

Takashi Iwai

2008-04-22 10:14 UTC #32 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

At Mon, 21 Apr 2008 21:55:27 +0400,

Quoted 17 lines · Stas Sergeev

Hello.
Quoted 4 lines · Takashi Iwai
Quoted 3 lines
CONFIG_DEBUG_PAGEALLOC is a very heavy consumer of CPU cycles. I'm not surprised that it would whack what I presume to be a very latency-sensitive driver.
We can add simply a dependncy to Kconfig if this really matters.
I think this is a bit too heavy-handed. That thing adds a lot of noise to the sound, but it doesn't really prevent the driver from working properly. And perhaps there are other options with the same effect? Also, that would motivate people to optimize DEBUG_PAGEALLOC, while otherwise they wouldn't know. :)

Quoted 4 lines

What is the problem with the warning exactly? If it makes problems, we can just remove it and update the help text instead, for example.

It's no big problem but I just find the phrase "you have to disable..." too strict. It's just a sound-quality problem as you mentioned in the above.

BTW, I noticed that the message includes a few line breaks. You need a proper KERN_ prefix for each line in such a case. A fix patch would be appreciated.

thanks,

Takashi

Dmitry Torokhov

maintainer

2008-04-22 14:01 UTC #33 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hi, Takashi, Stas,

Quoted 5 lines · Takashi Iwai

It's no big problem but I just find the phrase "you have to disable..." too strict. It's just a sound-quality problem as you mentioned in the above.

Just out of curiosity, what is the sound quality with the PC speaker? Not to belittle Stas's effort, but how relevant is this driver for mainline given that it os pretty much impossible nowadays to find a motherboard without on-board sound?

Signature · 1 line

Dmitry

Stas Sergeev

2008-04-22 16:42 UTC #34 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 1 line · Dmitry Torokhov

Just out of curiosity, what is the sound quality with the PC speaker?

That depends on the speaker itself. If it is large enough (not a piezo tablet), then the sound probably matches the telephone. :)

Quoted 3 lines

Not to belittle Stas's effort, but how relevant is this driver for mainline given that it os pretty much impossible nowadays to find a motherboard without on-board sound?

I know it is used because people are mailing me about it. Mainly on servers I think, as on a desktop its value it really questionable. :)
The real problem is that the board manufacturers put the piezo beepers these days, and also nvidia already excluded the necessary capabilities from their NForce chipset. So in the future it may indeed became out of the use.
By the way, you may be surprised, but I am still being asked by various people to write an LPT DAC (Covox) sound driver (the ancient OSS-based pc-speaker driver supported also that).
I guess people just like to use all the hardware they have, even if it is not really very usefull.

Stas Sergeev

2008-04-22 18:31 UTC #35 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 6 lines · Takashi Iwai

It's no big problem but I just find the phrase "you have to disable..." too strict. It's just a sound-quality problem as you mentioned in the above.
BTW, I noticed that the message includes a few line breaks. You need a proper KERN_ prefix for each line in such a case. A fix patch would be appreciated.

Like the attached one?

# HG changeset patch
# User Stas Sergeev <[email protected]> # Date 1208888978 -14400
# Node ID 00c36aa3903d861198e8f57c46e75194f3e172d6 # Parent f8fbce7ba45922a73a9679d68c4a4b07ebef42de
pcsp: fix wording in DEBUG_PAGEALLOC warning.

Signed-off-by Stas Sergeev <[email protected]>

drivers/pcsp/pcsp.c +2 -6

diff -r f8fbce7ba459 -r 00c36aa3903d drivers/pcsp/pcsp.c
@@ -147,12 +147,8 @@ static int __devinit alsa_card_pcsp_init
 
 #ifdef CONFIG_DEBUG_PAGEALLOC
 	/* Well, CONFIG_DEBUG_PAGEALLOC makes the sound horrible. Lets alert */
-	printk(KERN_WARNING
-	       "PCSP: Warning, CONFIG_DEBUG_PAGEALLOC is enabled!\n"
-	       "You have to disable it if you want to use the PC-Speaker "
-	       "driver.\n"
-	       "Unless it is disabled, enjoy the horrible, distorted "
-	       "and crackling noise.\n");
+	printk(KERN_WARNING "PCSP: CONFIG_DEBUG_PAGEALLOC is enabled, "
+		KERN_WARNING "which may make the sound noisy.\n");
 #endif
 
 	return 0;

Takashi Iwai

2008-04-23 08:49 UTC #36 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

At Tue, 22 Apr 2008 22:31:01 +0400,

Quoted 15 lines · Stas Sergeev

# HG changeset patch
# User Stas Sergeev <[email protected]> # Date 1208888978 -14400
# Node ID 00c36aa3903d861198e8f57c46e75194f3e172d6 # Parent f8fbce7ba45922a73a9679d68c4a4b07ebef42de
pcsp: fix wording in DEBUG_PAGEALLOC warning.
Signed-off-by Stas Sergeev <[email protected]>
diff -r f8fbce7ba459 -r 00c36aa3903d drivers/pcsp/pcsp.c
drivers/pcsp/pcsp.c +0 -0
@@ -147,12 +147,8 @@ static int __devinit alsa_card_pcsp_init
 

drivers/pcsp/pcsp.c static int __devinit alsa_card_pcsp_init · line 151

Quoted 10 lines

drivers/pcsp/pcsp.c +2 -6

 #ifdef CONFIG_DEBUG_PAGEALLOC
 	/* Well, CONFIG_DEBUG_PAGEALLOC makes the sound horrible. Lets alert */
-	printk(KERN_WARNING
-	       "PCSP: Warning, CONFIG_DEBUG_PAGEALLOC is enabled!\n"
-	       "You have to disable it if you want to use the PC-Speaker "
-	       "driver.\n"
-	       "Unless it is disabled, enjoy the horrible, distorted "
-	       "and crackling noise.\n");
+	printk(KERN_WARNING "PCSP: CONFIG_DEBUG_PAGEALLOC is enabled, "
+		KERN_WARNING "which may make the sound noisy.\n");

Missing \n in the first line?

thanks,

Takashi

Takashi Iwai

2008-04-23 14:18 UTC #37 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

At Wed, 23 Apr 2008 10:49:00 +0200,

Quoted 18 lines · I wrote:

At Tue, 22 Apr 2008 22:31:01 +0400,
Quoted 15 lines · Stas Sergeev
# HG changeset patch
# User Stas Sergeev <[email protected]> # Date 1208888978 -14400
# Node ID 00c36aa3903d861198e8f57c46e75194f3e172d6 # Parent f8fbce7ba45922a73a9679d68c4a4b07ebef42de
pcsp: fix wording in DEBUG_PAGEALLOC warning.
Signed-off-by Stas Sergeev <[email protected]>
diff -r f8fbce7ba459 -r 00c36aa3903d drivers/pcsp/pcsp.c
drivers/pcsp/pcsp.c +0 -0
@@ -147,12 +147,8 @@ static int __devinit alsa_card_pcsp_init
 

Quoted 12 lines

drivers/pcsp/pcsp.c static int __devinit alsa_card_pcsp_init · line 151

Quoted 10 lines

drivers/pcsp/pcsp.c +2 -6

 #ifdef CONFIG_DEBUG_PAGEALLOC
 	/* Well, CONFIG_DEBUG_PAGEALLOC makes the sound horrible. Lets alert */
-	printk(KERN_WARNING
-	       "PCSP: Warning, CONFIG_DEBUG_PAGEALLOC is enabled!\n"
-	       "You have to disable it if you want to use the PC-Speaker "
-	       "driver.\n"
-	       "Unless it is disabled, enjoy the horrible, distorted "
-	       "and crackling noise.\n");
+	printk(KERN_WARNING "PCSP: CONFIG_DEBUG_PAGEALLOC is enabled, "
+		KERN_WARNING "which may make the sound noisy.\n");

Missing \n in the first line?

Fixed on ALSA tree now, too.

Takashi

Stas Sergeev

2008-04-23 20:02 UTC #38 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 3 lines · Takashi Iwai

drivers/pcsp/pcsp.c static int __devinit alsa_card_pcsp_init · line 151
Quoted 2 lines
drivers/pcsp/pcsp.c +2 -0
+	printk(KERN_WARNING "PCSP: CONFIG_DEBUG_PAGEALLOC is enabled, "
+		KERN_WARNING "which may make the sound noisy.\n");
Missing \n in the first line?

This was intentional - wanted it to print in a single line. You see a space there for that reason.
Should the second KERN_WARNING be removed then, or what is the problem exactly?

Quoted 7 lines

Quoted 4 lines
- Prevent pcspkr from being built together
Quoted 3 lines
with snd-pcsp.
Signed-off-by Stas Sergeev <[email protected]>
Thanks, applied to my git tree.
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
Meanwhile, we need to add a similar depenency to snd-pcsp as well, no?

I personally don't think so.
snd-pcsp has the excact copy of the pcspkr code built-in, so I thought pcspkr can be obsoleted in the future. From that point of view, having snd-pcsp enabled and not even seeing pcspkr in a menuconfig is fine. While otherwise (you ocasionally enable pcspkr and don't even see snd-pcsp then) is not fine.
You mentioned earlier that you would like to be able to swap those drivers dynamically, but... what's the use? With such a dependancy added, many people will not even know about snd-pcsp as they already have pcspkr enabled.

Quoted 1 line

Fixed on ALSA tree now.

Oh...

Takashi Iwai

2008-04-24 09:41 UTC #39 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

At Thu, 24 Apr 2008 00:02:06 +0400,

Quoted 4 lines · Stas Sergeev

Hello.
Takashi Iwai wrote:

Quoted 8 lines

Quoted 3 lines
drivers/pcsp/pcsp.c static int __devinit alsa_card_pcsp_init · line 151
Quoted 2 lines
drivers/pcsp/pcsp.c +2 -0
+	printk(KERN_WARNING "PCSP: CONFIG_DEBUG_PAGEALLOC is enabled, "
+		KERN_WARNING "which may make the sound noisy.\n");
Missing \n in the first line?
This was intentional - wanted it to print in a single line. You see a space there for that reason.
Should the second KERN_WARNING be removed then, or what is the problem exactly?

Then you don't need KERN_WARNING there.
The problem is that you need KERN_* prefix again after the line break (\n). Your original patch didn't do that properly.

Quoted 7 lines

Quoted 7 lines
Quoted 4 lines
- Prevent pcspkr from being built together
Quoted 3 lines
with snd-pcsp.
Signed-off-by Stas Sergeev <[email protected]>
Thanks, applied to my git tree.
git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git
Meanwhile, we need to add a similar depenency to snd-pcsp as well, no?

Quoted 9 lines

I personally don't think so.
snd-pcsp has the excact copy of the pcspkr code built-in, so I thought pcspkr can be obsoleted in the future. From that point of view, having snd-pcsp enabled and not even seeing pcspkr in a menuconfig is fine. While otherwise (you ocasionally enable pcspkr and don't even see snd-pcsp then) is not fine.

No, I don't think input-pcspkr would be ever easily obsoleted by snd-pcsp. People definitely want a system without the sound subsystem but with a beep.

Quoted 5 lines

You mentioned earlier that you would like to be able to swap those drivers dynamically, but... what's the use? With such a dependancy added, many people will not even know about snd-pcsp as they already have pcspkr enabled.

Think about distro. They could distribute both modules if both modules can be replacible. Otherwise, snd-pcsp won't be enabled on most distros, I guess.

Takashi

Stas Sergeev

2008-04-25 03:52 UTC #40 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 3 lines · Takashi Iwai

No, I don't think input-pcspkr would be ever easily obsoleted by snd-pcsp. People definitely want a system without the sound subsystem but with a beep.

If you don't enable sound, then you can't enable snd-pcsp, and so the pcspkr can be enabled.
But if you have enabled the sound, what's the use in not seeing the snd-pcsp in the config? Because most people already had pcspkr enabled I guess, this just reduces the availability of snd-pcsp by an order of magnitude. And the reason for that, is... ?

Takashi Iwai

2008-04-25 06:28 UTC #41 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

At Fri, 25 Apr 2008 07:51:58 +0400,

Quoted 10 lines · Stas Sergeev

Hello.
Quoted 3 lines · Takashi Iwai
No, I don't think input-pcspkr would be ever easily obsoleted by snd-pcsp. People definitely want a system without the sound subsystem but with a beep.
If you don't enable sound, then you can't enable snd-pcsp, and so the pcspkr can be enabled.

Yeah, of course. Don't mix up with the argument about Kconfig dependency issues. I just reacted against your statement that input pcspkr driver will be obsoleted.

Quoted 8 lines

But if you have enabled the sound, what's the use in not seeing the snd-pcsp in the config? Because most people already had pcspkr enabled I guess, this just reduces the availability of snd-pcsp by an order of magnitude. And the reason for that, is... ?

I actually removed the dependency from snd-pcsp Kconfig again since this causes a dependency loop (at least, kbuild can't handle properly).

But anyway, as mentioned in my previous post, snd-pcsp wouldn't be enabled on most of distros' kernels because it cannot be built together with input-pcspkr driver, unfortunately...

Takashi

Stas Sergeev

2008-04-25 16:46 UTC #42 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 3 lines · Takashi Iwai

Yeah, of course. Don't mix up with the argument about Kconfig dependency issues. I just reacted against your statement that input pcspkr driver will be obsoleted.

Those are the different ones of course. For me the dependancy issue is a primary problem of course, so the other one I simply ducked.

Quoted 3 lines

I actually removed the dependency from snd-pcsp Kconfig again since this causes a dependency loop (at least, kbuild can't handle properly).

OK, good to know this. :)

Quoted 3 lines

But anyway, as mentioned in my previous post, snd-pcsp wouldn't be enabled on most of distros' kernels because it cannot be built together with input-pcspkr driver, unfortunately...

Yes, making them to play well together is something to think about too. But I don't see how it can affect the distributors. If the distro doesn't have the sound enabled, then it won't enable snd-pcsp no matter what. If it does have the sound enabled, then why would it ever want to use pcspkr instead of snd-pcsp?
I see this issue mostly as a theoretical one. If you see the real problems with the current situation, please let me know.

Takashi Iwai

2008-04-25 16:51 UTC #43 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

At Fri, 25 Apr 2008 20:45:51 +0400,

Quoted 4 lines · Stas Sergeev

Quoted 4 lines
But anyway, as mentioned in my previous post, snd-pcsp wouldn't be enabled on most of distros' kernels because it cannot be built together with input-pcspkr driver, unfortunately...

Quoted 8 lines

Yes, making them to play well together is something to think about too. But I don't see how it can affect the distributors. If the distro doesn't have the sound enabled, then it won't enable snd-pcsp no matter what. If it does have the sound enabled, then why would it ever want to use pcspkr instead of snd-pcsp?

The sound subsystem is enabled (loaded) only when necessary. As long as there are many systems without the sound subsystem (i.e. servers), snd-pcsp wouldn't be built (thus not provided even as a module) for their kernels because it prevents input-pcspkr.

Takashi

Stas Sergeev

2008-04-25 17:25 UTC #44 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 3 lines · Takashi Iwai

The sound subsystem is enabled (loaded) only when necessary. As long as there are many systems without the sound subsystem (i.e. servers),

Is it really a problem for the server admin to just tolerate a few extra modules being loaded? I really don't know, I just thought it is not.
Of course the one may not like the fact that after the kernel update, he gets a few modules more in lsmod output; someone may even call it a bloat...

$ lsmod |wc -l
79
Doesn't look too small already. :)

Anyway, I guess we'll soon find out the numbers, unless someone will come up with the good way to make those drivers to play better together and to not depend on each other, which looks a bit difficult from the first glance.

Takashi Iwai

2008-05-02 16:44 UTC #45 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

t Fri, 25 Apr 2008 18:51:04 +0200,

Quoted 16 lines · I wrote:

At Fri, 25 Apr 2008 20:45:51 +0400,
Quoted 4 lines · Stas Sergeev
Quoted 4 lines
But anyway, as mentioned in my previous post, snd-pcsp wouldn't be enabled on most of distros' kernels because it cannot be built together with input-pcspkr driver, unfortunately...
Quoted 8 lines
Yes, making them to play well together is something to think about too. But I don't see how it can affect the distributors. If the distro doesn't have the sound enabled, then it won't enable snd-pcsp no matter what. If it does have the sound enabled, then why would it ever want to use pcspkr instead of snd-pcsp?

Quoted 4 lines

The sound subsystem is enabled (loaded) only when necessary. As long as there are many systems without the sound subsystem (i.e. servers), snd-pcsp wouldn't be built (thus not provided even as a module) for their kernels because it prevents input-pcspkr.

... and I completely missed the viewpoint of device allocation. Yeah, that'll be a bit hackish.

Takashi

Stas Sergeev

2008-05-02 16:57 UTC #46 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Quoted 6 lines · Takashi Iwai

Quoted 4 lines
The sound subsystem is enabled (loaded) only when necessary. As long as there are many systems without the sound subsystem (i.e. servers), snd-pcsp wouldn't be built (thus not provided even as a module) for their kernels because it prevents input-pcspkr.
... and I completely missed the viewpoint of device allocation. Yeah, that'll be a bit hackish.

I'll appreciate if your replies
became a bit less cryptic... ;)
What will be a bit hackish?

Takashi Iwai

2008-05-06 10:20 UTC #47 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

At Fri, 02 May 2008 20:57:00 +0400,

Quoted 11 lines · Stas Sergeev

Quoted 6 lines · Takashi Iwai
Quoted 4 lines
The sound subsystem is enabled (loaded) only when necessary. As long as there are many systems without the sound subsystem (i.e. servers), snd-pcsp wouldn't be built (thus not provided even as a module) for their kernels because it prevents input-pcspkr.
... and I completely missed the viewpoint of device allocation. Yeah, that'll be a bit hackish.
I'll appreciate if your replies
became a bit less cryptic... ;)
What will be a bit hackish?

[Oops, overseen this follow up]

One problem is that we cannot load two drivers to a single device right now. Even if you have input pcspkr and snd-pcsp modules, you have to blacklist one of two modules so that udev loads the one properly. Because of this, snd-pcsp will be unlikely activated for most systems as default.

For avoiding this, you'll have a few choices:
a) implement pcspkr-core driver, and make input-pcspkr and snd-pcsp on that core module
b) make snd-pcsp copmletely rely on input pcspkr, implement as an add-on by adding hook to each driver callback and event handler of pcspkr
c) implement snd-pcsp as another individual platform driver and adds a hook to pcskr event handler of pcspkr

The case (a) would make things more complicated and give less solution.

In the case (b), the modification of pcspkr.c would be big, and would be ugly.

The case (c) was my proposal. But in this case, the driver will become likely self consistent; it allocates its own device at init.

In anyway, there is no sexy way to auto-load snd-pcsp (partly because it's the purpose -- avoid loading the sound subsystem unless really necessary). That's why I called it hackish.

Takashi

Stas Sergeev

2008-05-06 16:51 UTC #48 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 5 lines · Takashi Iwai

The case (c) was my proposal. But in this case, the driver will become likely self consistent; it allocates its own device at init. In anyway, there is no sexy way to auto-load snd-pcsp (partly because it's the purpose -- avoid loading the sound subsystem unless really necessary). That's why I called it hackish.

I thought all (or most) alsa drivers are allocating device on init, even though this is explicitly discouraged in the docs. So I was considering this as a possible solution with the minimal drawback.
But... as long as the autoloading by default is not needed, and both drivers can be at least built together, and not too much distros have pcspkr built-in, I thought the current solution - having pcspkr as the default but to let the user to choose snd-pcsp, is not all that bad too. I guess it costs only adding a single alias into modprobe.conf to choose snd-pcsp. And I also think _most_ distros do not mind having the sound subsystem loaded by default, but some certainly do. For those that do, the user will have to add an alias. For others - he may get snd-pcsp right away. IMHO this is rather acceptable.

Dmitry Torokhov

maintainer

2008-04-25 18:12 UTC #49 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

In reply to Stas Sergeev

Hi Stas,

Quoted 7 lines · Stas Sergeev

Hello.
Quoted 3 lines · Takashi Iwai
But anyway, as mentioned in my previous post, snd-pcsp wouldn't be enabled on most of distros' kernels because it cannot be built together with input-pcspkr driver, unfortunately...

Quoted 8 lines

Yes, making them to play well together is something to think about too. But I don't see how it can affect the distributors. If the distro doesn't have the sound enabled, then it won't enable snd-pcsp no matter what. If it does have the sound enabled, then why would it ever want to use pcspkr instead of snd-pcsp?

Given agerage quality of the speakers in the current boxes and abudance of on-board sound I doubt any distriution would want to have snd-pcsp enabled since there is a chance it may come up as a primary (default) sound device. At least my experience with F[C]7 that juggling 3 sound cards is not very easy.

Signature · 1 line

Dmitry

Stas Sergeev

2008-04-25 18:31 UTC #50 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Hello.

Quoted 3 lines · Dmitry Torokhov

abudance of on-board sound I doubt any distriution would want to have snd-pcsp enabled since there is a chance it may come up as a primary (default) sound device.

I avoid that by adding
options snd-pcsp index=1
into /etc/modprobe.conf.
Maybe index=5 or alike would be
safer for those who have many cards.

Dmitry Torokhov

maintainer

2008-04-25 18:37 UTC #51 lore raw

Re: 2.6.25-mm1 (snd-pcsp doesn't like DEBUG_PAGEALLOC)

Quoted 11 lines · Stas Sergeev

Hello.
Quoted 3 lines · Dmitry Torokhov
abudance of on-board sound I doubt any distriution would want to have snd-pcsp enabled since there is a chance it may come up as a primary (default) sound device.
I avoid that by adding
options snd-pcsp index=1
into /etc/modprobe.conf.
Maybe index=5 or alike would be
safer for those who have many cards.

Yes, it could be. But face it, the number of users wanting to play musuc through PC speaker is quite small and unlikely to increase in the future.

Signature · 1 line

Dmitry

Joseph Fannin

2008-04-19 02:26 UTC #52 lore raw

http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

Quoted 2 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

I've been seeing the following backtrace since (I think) 2.6.25-rc8-mm2.

I'm sending multiple reports vs. 2.6.25-mm1, so I'm putting the dmesg and .config on a server:

[  842.795144] hm, dftrace overflow: 265 changes (0 total) in 428 usecs
[  842.795182] ------------[ cut here ]------------
[  842.795192] WARNING: at kernel/trace/ftrace.c:658 ftraced+0x1a4/0x1b0()
[  842.795200] Modules linked in: af_packet rfcomm l2cap bluetooth ppdev ipv6 cpufreq_conservative cpufreq_stats cpufreq_userspace cpufreq_powersave video output wmi pci_slot container dock sbs sbshcbattery iptable_filter ip_tables x_tables ext2 ac lp loop snd_via82xx gameport snd_ac97_codec ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  842.795470] Pid: 13, comm: ftraced Tainted: G        W 2.6.25-mm1 #7
[  842.795497]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  842.795541]  [<c013244f>] ? vprintk+0x33f/0x4a0
[  842.795589]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  842.795622]  [<c0354eb0>] ? __mutex_lock_common+0x2b0/0x3c0
[  842.795667]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  842.795688]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  842.795709]  [<c017e4d0>] ? __ftrace_update_code+0x0/0x110
[  842.795730]  [<c017e9f0>] ? ftraced+0x0/0x1b0
[  842.795746]  [<c01325d0>] ? printk+0x20/0x30
[  842.795764]  [<c017e9f0>] ? ftraced+0x0/0x1b0
[  842.795780]  [<c017eb94>] ftraced+0x1a4/0x1b0
[  842.795807]  [<c0146527>] kthread+0x47/0x80
[  842.795836]  [<c01464e0>] ? kthread+0x0/0x80
[  842.795855]  [<c0104d9b>] kernel_thread_helper+0x7/0x10
[  842.795890]  =======================
[  842.795899] ---[ end trace 4eaa2a86a8e2da22 ]---

--
Joseph Fannin
[email protected]

Andrew Morton

author maintainer

2008-04-19 03:14 UTC #53 lore raw

Quoted 4 lines · On Fri, 18 Apr 2008 22:25:59 -0400 Joseph Fannin <[email protected]> wrote:

Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

Quoted 8 lines

I've been seeing the following backtrace since (I think) 2.6.25-rc8-mm2.
I'm sending multiple reports vs. 2.6.25-mm1, so I'm putting the dmesg and .config on a server:
http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

Thanks.

Quoted 1 line

[  400.781246] WARNING: at arch/x86/mm/ioremap.c:158 __ioremap_caller+0x27d/0x2e0()

I don't think I've seen that one before.

Quoted 1 line

[  406.935080] sysfs: duplicate filename '189:0' can not be created

Seen plenty of them - I think Greg today dropped the offending patch(es).

[  451.915553] sysfs: duplicate filename 'pcspkr' can not be created

Yup. Maybe the above droppage will make this go away, I'm not sure.

Quoted 13 lines

[  842.795144] hm, dftrace overflow: 265 changes (0 total) in 428 usecs
[  842.795182] ------------[ cut here ]------------
[  842.795192] WARNING: at kernel/trace/ftrace.c:658 ftraced+0x1a4/0x1b0()
[  842.795200] Modules linked in: af_packet rfcomm l2cap bluetooth ppdev ipv6 cpufreq_conservative cpufreq_stats cpufreq_userspace cpufreq_powersave video output wmi pci_slot container dock sbs sbshcbattery iptable_filter ip_tables x_tables ext2 ac lp loop snd_via82xx gameport snd_ac97_codec ac97_bus snd_pcm_oss snd_mixer_oss snd_pcm snd_mpu401_uart snd_seq_dummy snd_seq_oss snd_seq_midi psmouse snd_rawmidi serio_raw snd_seq_midi_event snd_seq button i2c_viapro snd_timer snd_seq_device pcspkr i2c_core snd snd_page_alloc via686a shpchp pci_hotplug parport_pc parport via_agp agpgart soundcore evdev sg sr_mod cdrom sd_mod 8139cp aic7xxx scsi_transport_spi scsi_mod 8139too mii uhci_hcd usbcore raid10 raid456 async_xor async_memcpy async_tx xor raid1 raid0 multipath linear md_mod thermal processor fan fuse ext4dev mbcache jbd2 crc16
[  842.795470] Pid: 13, comm: ftraced Tainted: G        W 2.6.25-mm1 #7
[  842.795497]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  842.795541]  [<c013244f>] ? vprintk+0x33f/0x4a0
[  842.795589]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  842.795622]  [<c0354eb0>] ? __mutex_lock_common+0x2b0/0x3c0
[  842.795667]  [<c0155216>] ? trace_hardirqs_on_caller+0x16/0x150
[  842.795688]  [<c015535b>] ? trace_hardirqs_on+0xb/0x10
[  842.795709]  [<c017e4d0>] ? __ftrace_update_code+0x0/0x110
[  842.795730]  [<c017e9f0>] ? ftraced+0x0/0x1b0

Quoted 8 lines

[  842.795746]  [<c01325d0>] ? printk+0x20/0x30
[  842.795764]  [<c017e9f0>] ? ftraced+0x0/0x1b0
[  842.795780]  [<c017eb94>] ftraced+0x1a4/0x1b0
[  842.795807]  [<c0146527>] kthread+0x47/0x80
[  842.795836]  [<c01464e0>] ? kthread+0x0/0x80
[  842.795855]  [<c0104d9b>] kernel_thread_helper+0x7/0x10
[  842.795890]  =======================
[  842.795899] ---[ end trace 4eaa2a86a8e2da22 ]---

I haven't seen that one before.

Joseph Fannin

2008-04-19 03:10 UTC #54 lore raw

http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

Quoted 2 lines · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

New, in 2.6.25-mm1 is a hang I'm seeing, just after the kernel prints:

"[ 0.160375] NET: Registered protocol family 16"

The hang lasts about five minutes, and then boot continues. Just after that, a backtrace is printed; I don't know if it's related. The backtrace will follow.

This does not occur in mainline. It seems it might be related to OLPC support -- I enabled all those options -- but that's not good behavior, and I see no warning of thus in the help.

I'm sending a number or reports against 2.6.25-mm1, so I've put my dmesg and .config on a server:

[    0.160375] NET: Registered protocol family 16
[  400.782683] ------------[ cut here ]------------
[  400.782832] WARNING: at arch/x86/mm/ioremap.c:158 __ioremap_caller+0x27d/0x2e0()
[  400.783022] Modules linked in:
[  400.783169] Pid: 1, comm: swapper Not tainted 2.6.25-mm1 #7
[  400.783300]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  400.783480]  [<c0106c2e>] ? profile_pc+0x3e/0x50
[  400.783682]  [<c01374ee>] ? irq_exit+0x4e/0xa0
[  400.783879]  [<c0115aec>] ? smp_apic_timer_interrupt+0x5c/0x90
[  400.784087]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784298]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.784506]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784706]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.784906]  [<c011d0e6>] ? page_is_ram+0xa6/0xd0
[  400.785059]  [<c011d4ed>] __ioremap_caller+0x27d/0x2e0
[  400.785221]  [<c03569d8>] ? _spin_unlock_irqrestore+0x48/0x80
[  400.785421]  [<c017f4cd>] ? ftrace_record_ip+0x7d/0x250
[  400.785621]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.785817]  [<c011d59f>] ioremap_nocache+0x1f/0x30
[  400.785976]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.786165]  [<c0474801>] olpc_init+0x31/0x140
[  400.786318]  [<c0464992>] kernel_init+0x142/0x2d0
[  400.786479]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.786680]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.786879]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787069]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787260]  [<c0104d9b>] kernel_thread_helper+0x7/0x10
[  400.787422]  =======================
[  400.787727] ---[ end trace 4eaa2a86a8e2da22 ]---

--
Joseph Fannin
[email protected]

Andrew Morton

author maintainer

2008-04-19 03:29 UTC #55 lore raw

Quoted 9 lines · On Fri, 18 Apr 2008 23:10:24 -0400 Joseph Fannin <[email protected]> wrote:

Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
New, in 2.6.25-mm1 is a hang I'm seeing, just after the kernel prints:
"[ 0.160375] NET: Registered protocol family 16"
The hang lasts about five minutes, and then boot continues.

Please add initcall_debug to the kernel boot command line - that should narrow it down.

Quoted 35 lines

Just
after that, a backtrace is printed; I don't know if it's related. The backtrace will follow.

This does not occur in mainline. It seems it might be related to OLPC support -- I enabled all those options -- but that's not good behavior, and I see no warning of thus in the help.

I'm sending a number or reports against 2.6.25-mm1, so I've put my dmesg and .config on a server:

http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt

[    0.160375] NET: Registered protocol family 16
[  400.782683] ------------[ cut here ]------------
[  400.782832] WARNING: at arch/x86/mm/ioremap.c:158 __ioremap_caller+0x27d/0x2e0()
[  400.783022] Modules linked in:
[  400.783169] Pid: 1, comm: swapper Not tainted 2.6.25-mm1 #7
[  400.783300]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  400.783480]  [<c0106c2e>] ? profile_pc+0x3e/0x50
[  400.783682]  [<c01374ee>] ? irq_exit+0x4e/0xa0
[  400.783879]  [<c0115aec>] ? smp_apic_timer_interrupt+0x5c/0x90
[  400.784087]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784298]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.784506]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784706]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.784906]  [<c011d0e6>] ? page_is_ram+0xa6/0xd0
[  400.785059]  [<c011d4ed>] __ioremap_caller+0x27d/0x2e0
[  400.785221]  [<c03569d8>] ? _spin_unlock_irqrestore+0x48/0x80
[  400.785421]  [<c017f4cd>] ? ftrace_record_ip+0x7d/0x250
[  400.785621]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.785817]  [<c011d59f>] ioremap_nocache+0x1f/0x30
[  400.785976]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.786165]  [<c0474801>] olpc_init+0x31/0x140

Quoted 8 lines

[  400.786318]  [<c0464992>] kernel_init+0x142/0x2d0
[  400.786479]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.786680]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.786879]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787069]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787260]  [<c0104d9b>] kernel_thread_helper+0x7/0x10
[  400.787422]  =======================
[  400.787727] ---[ end trace 4eaa2a86a8e2da22 ]---

That's

                WARN_ON_ONCE(is_ram);

the changelog for the patch which added that warning is information-free and there's no code comment explaining what went wrong, which makes things rather harder than they ought to be.

Yes it's due to the new OLPC code. olpc_init() has

	romsig = ioremap(0xffffffc0, 16);

which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed.

Andres Salomon

2008-04-19 13:23 UTC #56 lore raw

Quoted 69 lines · On Fri, 18 Apr 2008 20:29:25 -0700 Andrew Morton <[email protected]>…

Quoted 9 lines · On Fri, 18 Apr 2008 23:10:24 -0400 Joseph Fannin <[email protected]> wrote:
Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
New, in 2.6.25-mm1 is a hang I'm seeing, just after the kernel prints:
"[ 0.160375] NET: Registered protocol family 16"
The hang lasts about five minutes, and then boot continues.
Please add initcall_debug to the kernel boot command line - that should narrow it down.
Quoted 35 lines
Just
after that, a backtrace is printed; I don't know if it's related. The backtrace will follow.
This does not occur in mainline. It seems it might be related to OLPC support -- I enabled all those options -- but that's not good behavior, and I see no warning of thus in the help.
I'm sending a number or reports against 2.6.25-mm1, so I've put my dmesg and .config on a server:
http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt
[    0.160375] NET: Registered protocol family 16
[  400.782683] ------------[ cut here ]------------
[  400.782832] WARNING: at arch/x86/mm/ioremap.c:158 __ioremap_caller+0x27d/0x2e0()
[  400.783022] Modules linked in:
[  400.783169] Pid: 1, comm: swapper Not tainted 2.6.25-mm1 #7
[  400.783300]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  400.783480]  [<c0106c2e>] ? profile_pc+0x3e/0x50
[  400.783682]  [<c01374ee>] ? irq_exit+0x4e/0xa0
[  400.783879]  [<c0115aec>] ? smp_apic_timer_interrupt+0x5c/0x90
[  400.784087]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784298]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.784506]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784706]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.784906]  [<c011d0e6>] ? page_is_ram+0xa6/0xd0
[  400.785059]  [<c011d4ed>] __ioremap_caller+0x27d/0x2e0
[  400.785221]  [<c03569d8>] ? _spin_unlock_irqrestore+0x48/0x80
[  400.785421]  [<c017f4cd>] ? ftrace_record_ip+0x7d/0x250
[  400.785621]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.785817]  [<c011d59f>] ioremap_nocache+0x1f/0x30
[  400.785976]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.786165]  [<c0474801>] olpc_init+0x31/0x140
Quoted 8 lines
[  400.786318]  [<c0464992>] kernel_init+0x142/0x2d0
[  400.786479]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.786680]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.786879]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787069]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787260]  [<c0104d9b>] kernel_thread_helper+0x7/0x10
[  400.787422]  =======================
[  400.787727] ---[ end trace 4eaa2a86a8e2da22 ]---
<looks at this again>
That's
                WARN_ON_ONCE(is_ram);
the changelog for the patch which added that warning is information-free and there's no code comment explaining what went wrong, which makes things rather harder than they ought to be.

Quoted 7 lines

Yes it's due to the new OLPC code. olpc_init() has
	romsig = ioremap(0xffffffc0, 16);
which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed.

Hm. We could either protect that code with an:

if (!is_geode())
return;

Or I could add the OpenFirmware patches which would allow us to get rid of this code, and instead check for the existence of OFW using that.

The former is quick and easy; the latter is (imo) nicer, so long as people don't have problems w/ the OFW code. :)

Signature · 1 line

Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Andres Salomon

2008-04-19 17:36 UTC #57 lore raw

PATCH 1/2 OLPC: Add support for calling into Open Firmware

This adds 32-bit support for calling into OFW from the kernel. It's useful for querying the firmware for misc hardware information, fetching the device tree, etc.

There's potentially no reason why other platforms couldn't use this, but currently OLPC is the main user of it.

This work was originally done by Mitch Bradley.

---
 arch/x86/Kconfig          |    8 +++++
 arch/x86/kernel/Makefile  |    1 +
 arch/x86/kernel/head_32.S |   27 ++++++++++++++++
 arch/x86/kernel/ofw.c     |   75 +++++++++++++++++++++++++++++++++++++++++++++
 include/asm-x86/ofw.h     |   50 ++++++++++++++++++++++++++++++
 include/asm-x86/setup.h   |    1 +
 6 files changed, 162 insertions(+), 0 deletions(-)
 create mode 100644 arch/x86/kernel/ofw.c
 create mode 100644 include/asm-x86/ofw.h

arch/x86/Kconfig +8 -0

index 3b9089b..ce56105 100644
@@ -661,6 +661,14 @@ config I8K
 	  Say Y if you intend to run this kernel on a Dell Inspiron 8000.
 	  Say N otherwise.
 
+config OPEN_FIRMWARE
+	bool "Support for Open Firmware"
+	default y if OLPC
+	---help---
+	  This option adds support for the implementation of Open Firmware
+	  that is used on the OLPC XO laptop.
+	  If unsure, say N here.
+
 config X86_REBOOTFIXUPS
 	def_bool n
 	prompt "Enable X86 board specific fixups for reboot"

arch/x86/kernel/Makefile +1 -0

index 9575754..d33600e 100644
@@ -54,6 +54,7 @@ obj-$(CONFIG_X86_TRAMPOLINE)	+= trampoline_$(BITS).o
 obj-$(CONFIG_X86_MPPARSE)	+= mpparse_$(BITS).o
 obj-$(CONFIG_X86_LOCAL_APIC)	+= apic_$(BITS).o nmi_$(BITS).o
 obj-$(CONFIG_X86_IO_APIC)	+= io_apic_$(BITS).o
+obj-$(CONFIG_OPEN_FIRMWARE)	+= ofw.o
 obj-$(CONFIG_X86_REBOOTFIXUPS)	+= reboot_fixups_32.o
 obj-$(CONFIG_KEXEC)		+= machine_kexec_$(BITS).o
 obj-$(CONFIG_KEXEC)		+= relocate_kernel_$(BITS).o crash.o

arch/x86/kernel/head_32.S +27 -0

index 74d87ea..c9d2d00 100644
@@ -132,6 +132,33 @@ ENTRY(startup_32)
 	movsl
 1:
 
+#ifdef CONFIG_OPEN_FIRMWARE
+/*
+ * If Open Firmware booted us, save the OFW client interface callback address
+ * and preserve the OFW page mappings by priming the kernel's new page
+ * directory area with a copy of the OFW page directory.  That lets OFW stay
+ * resident in high memory (high in both the virtual and physical spaces)
+ * for at least long enough to copy out the device tree.
+ */
+	movl $pa(boot_params + OFW_INFO_OFFSET), %ebp
+	cmpl $0x2057464F, (%ebp)		/* Magic number "OFW " */
+	jne 4f
+
+	mov 0x8(%ebp), %eax			/* Save callback address */
+	mov %eax, pa(call_firmware)
+
+	/* Copy the OFW pdir into swapper_pg_dir */
+	movl %esi, %edx		/* save %esi */
+	movl $pa(swapper_pg_dir), %edi
+	movl %cr3, %esi		/* Source is current pg_dir base address */
+	movl $1024, %ecx	/* Number of page directory entries */
+	rep
+	movsl
+	movl %edx, %esi		/* restore %esi */
+4:
+
+#endif
+
 #ifdef CONFIG_PARAVIRT
 	/* This is can only trip for a broken bootloader... */
 	cmpw $0x207, pa(boot_params + BP_version)

arch/x86/kernel/ofw.c new file +75 -0

75 lines not on this page · raw message

include/asm-x86/ofw.h new file +50 -0

50 lines not on this page · raw message

include/asm-x86/setup.h +1 -0

index 071e054..8e2b674 100644
@@ -28,6 +28,7 @@ char *machine_specific_memory_setup(void);
 #define OLD_CL_MAGIC		0xA33F
 #define OLD_CL_ADDRESS		0x020	/* Relative to real mode data */
 #define NEW_CL_POINTER		0x228	/* Relative to real mode data */
+#define OFW_INFO_OFFSET		0xb0	/* Relative to real mode data */
 
 #ifndef __ASSEMBLY__
 #include <asm/bootparam.h>

Signature · 5 lines

1.5.4.4


-- 
Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Yinghai Lu

2008-04-20 10:34 UTC #58 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 110 lines · Andres Salomon

This adds 32-bit support for calling into OFW from the kernel. It's useful for querying the firmware for misc hardware information, fetching the device tree, etc.
There's potentially no reason why other platforms couldn't use this, but currently OLPC is the main user of it.
This work was originally done by Mitch Bradley.
Signed-off-by Andres Salomon <[email protected]>
---
  arch/x86/Kconfig          |    8 +++++
  arch/x86/kernel/Makefile  |    1 +
  arch/x86/kernel/head_32.S |   27 ++++++++++++++++
  arch/x86/kernel/ofw.c     |   75 +++++++++++++++++++++++++++++++++++++++++++++
  include/asm-x86/ofw.h     |   50 ++++++++++++++++++++++++++++++
  include/asm-x86/setup.h   |    1 +
  6 files changed, 162 insertions(+), 0 deletions(-)
create mode 100644 arch/x86/kernel/ofw.c
create mode 100644 include/asm-x86/ofw.h
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 3b9089b..ce56105 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -661,6 +661,14 @@ config I8K
Say Y if you intend to run this kernel on a Dell Inspiron 8000. Say N otherwise.
+config OPEN_FIRMWARE
 +       bool "Support for Open Firmware"
 +       default y if OLPC
 +       ---help---
 +         This option adds support for the implementation of Open Firmware
 +         that is used on the OLPC XO laptop.
 +         If unsure, say N here.
 +
  config X86_REBOOTFIXUPS
        def_bool n
        prompt "Enable X86 board specific fixups for reboot"
 diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
 index 9575754..d33600e 100644
 --- a/arch/x86/kernel/Makefile
 +++ b/arch/x86/kernel/Makefile
 @@ -54,6 +54,7 @@ obj-$(CONFIG_X86_TRAMPOLINE)  += trampoline_$(BITS).o
  obj-$(CONFIG_X86_MPPARSE)      += mpparse_$(BITS).o
  obj-$(CONFIG_X86_LOCAL_APIC)   += apic_$(BITS).o nmi_$(BITS).o
  obj-$(CONFIG_X86_IO_APIC)      += io_apic_$(BITS).o
 +obj-$(CONFIG_OPEN_FIRMWARE)    += ofw.o
obj-$(CONFIG_X86_REBOOTFIXUPS) += reboot_fixups_32.o
obj-$(CONFIG_KEXEC) += machine_kexec_$(BITS).o
obj-$(CONFIG_KEXEC) += relocate_kernel_$(BITS).o crash.o
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S index 74d87ea..c9d2d00 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -132,6 +132,33 @@ ENTRY(startup_32)
movsl
1:
+#ifdef CONFIG_OPEN_FIRMWARE
+/*
+ * If Open Firmware booted us, save the OFW client interface callback address
+ * and preserve the OFW page mappings by priming the kernel's new page
+ * directory area with a copy of the OFW page directory. That lets OFW stay
+ * resident in high memory (high in both the virtual and physical spaces)
+ * for at least long enough to copy out the device tree.
+ */
 +       movl $pa(boot_params + OFW_INFO_OFFSET), %ebp
 +       cmpl $0x2057464F, (%ebp)                /* Magic number "OFW " */
 +       jne 4f
 +
 +       mov 0x8(%ebp), %eax                     /* Save callback address */
 +       mov %eax, pa(call_firmware)
 +
 +       /* Copy the OFW pdir into swapper_pg_dir */
 +       movl %esi, %edx         /* save %esi */
 +       movl $pa(swapper_pg_dir), %edi
 +       movl %cr3, %esi         /* Source is current pg_dir base address */
 +       movl $1024, %ecx        /* Number of page directory entries */
 +       rep
 +       movsl
 +       movl %edx, %esi         /* restore %esi */
 +4:
 +
 +#endif
 +
  #ifdef CONFIG_PARAVIRT
        /* This is can only trip for a broken bootloader... */
        cmpw $0x207, pa(boot_params + BP_version)
diff --git a/arch/x86/kernel/ofw.c b/arch/x86/kernel/ofw.c new file mode 100644
index 0000000..14036aa
--- /dev/null
+++ b/arch/x86/kernel/ofw.c
@@ -0,0 +1,75 @@
+/*
+ * Open Firmware client interface for 32-bit systems.
+ *
+ * Copyright (c) 2007 Mitch Bradley <[email protected]>
+ * Copyright (c) 2007-2008 Andres Salomon <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/kernel.h>

Quoted 8 lines

+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <asm/ofw.h>
+
+/*
+ * This code is intended to be portable to any 32-bit Open Firmware
+ * implementation with a standard client interface that can be
+ * called when Linux is running.

how about changing to ofw_32.c?

H. Peter Anvin

reviewer

2008-04-20 12:18 UTC #59 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 10 lines · Yinghai Lu

Quoted 9 lines · Andres Salomon
This adds 32-bit support for calling into OFW from the kernel. It's useful for querying the firmware for misc hardware information, fetching the device tree, etc.
There's potentially no reason why other platforms couldn't use this, but currently OLPC is the main user of it.
This work was originally done by Mitch Bradley.

Hm. This interface seems more than a bit ad hoc. In particular, I *really* don't like the swapper_pg_dir hack.

"There must be a better way."

-hpa

Andres Salomon

2008-04-20 17:57 UTC #60 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 12 lines · On Sun, 20 Apr 2008 08:07:55 -0400 "H. Peter Anvin" <[email protected]> wrote:

Quoted 10 lines · Yinghai Lu
Quoted 9 lines · Andres Salomon
This adds 32-bit support for calling into OFW from the kernel. It's useful for querying the firmware for misc hardware information, fetching the device tree, etc.
There's potentially no reason why other platforms couldn't use this, but currently OLPC is the main user of it.
This work was originally done by Mitch Bradley.

Quoted 6 lines

Hm. This interface seems more than a bit ad hoc. In particular, I *really* don't like the swapper_pg_dir hack.
"There must be a better way."
-hpa

I'm certainly open to suggestions.. Otherwise, I'll poke around and see if I can come up w/ something.

Signature · 1 line

Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Mitch Bradley

2008-04-20 18:52 UTC #61 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 20 lines · Andres Salomon

Quoted 16 lines · On Sun, 20 Apr 2008 08:07:55 -0400 "H. Peter Anvin" <[email protected]> wrote:
Quoted 6 lines · Yinghai Lu
Quoted 4 lines · Andres Salomon
This adds 32-bit support for calling into OFW from the kernel. It's useful for querying the firmware for misc hardware information, fetching the device tree, etc.
Quoted 6 lines
Quoted 6 lines
There's potentially no reason why other platforms couldn't use this, but currently OLPC is the main user of it.
This work was originally done by Mitch Bradley.
Hm. This interface seems more than a bit ad hoc. In particular, I *really* don't like the swapper_pg_dir hack.

Quoted 8 lines

Quoted 4 lines
"There must be a better way."
-hpa
I'm certainly open to suggestions.. Otherwise, I'll poke around and see if I can come up w/ something.

The x86 architecture doesn't make this problem easy.

The conventional solution is to have the BIOS operate in real mode. When the kernel calls into the BIOS, it has to do a grotesque dance that involves jumping through a chain of several segments of different flavors, thus gradually shutting down the multi-tiered address translation mechanism. Then, if the BIOS is actually operating in protected mode (which is necessary if it is larger than 64K, as all modern BIOSes are), it has to perform the inverse process, do the requested work, then go back into real mode to return to the kernel. The net result is that a "call" into the BIOS involves:

a) Copying the arguments to a real-mode register shadow array
b) Saving all the registers - general ones and a few special ones too
c) Far call to a linear-mapped code segment with an execution address in the first 1M of memory
d) Switching to a different stack
e) Turning off page translation
f) Switching from protected mode to real mode (or in some cases, V86 mode instead, which requires an additional Task State Segment dance to set the IO permission mask)
g) Switching to a real-mode interrupt descriptor table

h) Executing an INT instruction

I) Performing the inverse of a - g inside the BIOS

j) Doing the requested work

K) Performing a - g again to get back into real mode

l) Executing an "iret" instruction

M) Performing the inverse of a-g to return to normal operation

The machinery that you need to do all that is predictably complex - extra segment descriptors that are set up just-so, several little code fragments that must be at special addresses in the first meg, additional stacks, a real-mode interrupt table at a fixed address, and several data save arrays. That machinery has to be in assembly language, spanning several different instruction set modes.

Compared to that, I think that sharing one or two page directory entries at the very top of the virtual address space is pretty clean and simple. With that sharing, the BIOS call is just an ordinary subroutine call. (The setup code copies the entire page directory, but only a couple of entries are actually needed. The reason for copying the whole thing is because it is rather more work to determine the exact number of entries necessary, compared to copying everything and then letting Linux replace the ones it uses.)

Every other solution that I know of requires some sort of heroic dance, either from the OS or from the BIOS or (usually) both.

H. Peter Anvin

reviewer

2008-04-20 19:17 UTC #62 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 3 lines · Mitch Bradley

The x86 architecture doesn't make this problem easy.

[long rant about the x86 architecture]

It would be more useful if you described the actual defined entry conditions from OpenFirmware look like, including if they are well-defined for all OF implementations or only for OLPC.

-hpa

Mitch Bradley

2008-04-21 03:39 UTC #63 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 10 lines · H. Peter Anvin

Quoted 3 lines · Mitch Bradley
The x86 architecture doesn't make this problem easy.
[long rant about the x86 architecture]
It would be more useful if you described the actual defined entry conditions from OpenFirmware look like, including if they are well-defined for all OF implementations or only for OLPC.
-hpa

Fair enough...

To get the second subquestion out of the way: At the present time, on the x86 architecture, "all OF implementations" and "OLPC" are effectively the same. I am unaware of any other x86 OFW deployments in current use. There have been some in the past, on bespoke systems such as Network Appliance servers and at least one settop box, but those have fallen by the wayside as those companies have shifted over to commodity PC hardware. The current market status quo is that x86 boards are primarily designed for Windows, and thus must run legacy BIOS, with some recent migration to EFI, neither of which are open source in the strong sense. While I would like to see more OFW penetration into the larger x86 market, I don't really expect it. x86 motherboard manufacturing is becoming more and more difficult as signal speeds increase, leading to a decline in the number of manufacturers. The existing manufacturers depend on Windows for sales volume and their internal procedures and working knowledge are based on legacy BIOS.

Once upon a time, we had an OFW "binding" document that stipulated the interface conditions, with the intention of making that "standard" across all OFW-on-x86 systems. However, by the time OLPC came around, there were no other systems to consider, so I felt free to make some changes in the interface. I ended up choosing an ABI that resulted in a simple (in the sense of not much code, and no complex state transitions) interface with 2.6 Linux kernels.

The interface defined below is not inherently OLPC-specific - it would be suitable for any ia32 system that used OFW. (At a higher level, the set of OFW callback functions is architecture-neutral; in this message I am focusing on the very low-level details of the ia32 ABI)

The system conditions for the OFW to Linux kernel transition are as follows:

a) OFW can load the Linux kernel from either bzimage format or ELF format (either uncompressed or zlib-compressed.) If the kernel is in ELF format with symbols, OFW can do symbolic kernel debugging. Further discussion will focus on bzimage format, as that is what OLPC uses and is also the "greased path" for kernel builds.

b) OFW loads the bzimage kernel at 0x100000 and the ramdisk image, if any, at 0x800000.

c) OFW runs in 32-bit protected mode with paging enabled. There are two reasons for this. First, it lets OFW locate itself at the top of both physical memory space and virtual address space, the only places that are truly "out of the way". It also lets Linux call back into OFW with a minimum of bother (similarly helping with the OFW-debugs-Linux scenario).

d) OFW itself lives at the top of the virtual address space, just below the ROM. (The ROM is mapped virtual=physical for convenience) OFW uses RAM allocated from the top of physical memory, mapped at the aforementioned high virtual addresses. One page directory entry - the next to last one - is used for that RAM mapping and also for mapping additional miscellaneous I/O devices. The 8MB frame buffer requires 2 additional PDEs, just below. When Linux takes over the display, OFW no longer needs the frame buffer mapping, but it is convenient to preserve that mapping temporarily while using OFW as a debugger.

e) Low memory - everything except the ~1Meg that OFW lives in - is mapped virtual=physical.

f) The code, data, and stack segment descriptors are all for 32-bit 4GB linear segments. (For debugging convenience, OFW initially uses the same segment numbers as the Linux kernel, but that is not an ABI requirement - callbacks into OFW will work from any 32-bit segments that encompass the kernel and OFW virtual address space.)

g) OFW sets up a boot parameter block at 0x90000, with the format as defined in include/asm-x86/bootparam.h . The block includes the cmdline, memory layout information, screen info, and address/length of the ramdisk.

h) OFW also includes in the boot parameter block an additional information block at OFW_INFO_OFFSET. That info block contains:
"OFW " - 4-byte signature string ('O' at byte offset 0, etc)
__u32 version - 1
__u32 callback - (actually a 32-bit function pointer) address of
callback function
__u32 idt - address of OFW's interrupt descriptor table, in case
Linux wants to preserve the breakpoint vector

i) OFW transfers control to the bzimage with the processor in the following state:
* interrupts masked off at the interrupt controller and in the flags
register
* 32-bit protected mode, CS=0x60, DS=ES=FS=GS=SS=0x68
* ebx=ecx=edx=ebp=edi=0
* eax = &callback_function
* esi = 0x90000 (boot parameter block)
* esp = 0x90000 (short-lived initial stack starts below boot
parameter block)
* eip = 0x100000 (entry address of bzimage)
* paging enabled, with V=P mapping of physical memory and one high
pde for OFW, as in (d) and (e) above

j) Linux must save the following information during early startup:
1) The callback function address - either from the initial value of
eax or from the OFW info block.
2) The the next-to-last page directory entry - just the pointer to the
page table. The page table itself lives in OFW's reserved memory.

k) When calling back into OFW, Linux must:
1) Establish a page directory that contains OFW's PDE (saved in j2
above) and that maps the client interface argument array, including any buffer pointers.
2) Call callback_function with the address of the argument array in
eax. (Ordinary 32-bit near call).

For all of the OLPC kernel's current callbacks into OFW, the requirements (j2) and (k1) are easily satisfied by "priming" swapper_pg_dir with the contents of the current page directory (as pointed to by the CR3 register).

Yinghai Lu

2008-04-21 04:54 UTC #64 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 52 lines · Mitch Bradley

Quoted 10 lines · H. Peter Anvin
Quoted 4 lines · Mitch Bradley
The x86 architecture doesn't make this problem easy.
[long rant about the x86 architecture]
It would be more useful if you described the actual defined entry
conditions from OpenFirmware look like, including if they are well-defined for all OF implementations or only for OLPC.
Quoted 3 lines
-hpa
Fair enough...
To get the second subquestion out of the way: At the present time, on the
x86 architecture, "all OF implementations" and "OLPC" are effectively the same. I am unaware of any other x86 OFW deployments in current use. There have been some in the past, on bespoke systems such as Network Appliance servers and at least one settop box, but those have fallen by the wayside as those companies have shifted over to commodity PC hardware. The current market status quo is that x86 boards are primarily designed for Windows, and thus must run legacy BIOS, with some recent migration to EFI, neither of which are open source in the strong sense. While I would like to see more OFW penetration into the larger x86 market, I don't really expect it. x86 motherboard manufacturing is becoming more and more difficult as signal speeds increase, leading to a decline in the number of manufacturers. The existing manufacturers depend on Windows for sales volume and their internal procedures and working knowledge are based on legacy BIOS.
Once upon a time, we had an OFW "binding" document that stipulated the
interface conditions, with the intention of making that "standard" across all OFW-on-x86 systems. However, by the time OLPC came around, there were no other systems to consider, so I felt free to make some changes in the interface. I ended up choosing an ABI that resulted in a simple (in the sense of not much code, and no complex state transitions) interface with 2.6 Linux kernels.
The interface defined below is not inherently OLPC-specific - it would be
suitable for any ia32 system that used OFW. (At a higher level, the set of OFW callback functions is architecture-neutral; in this message I am focusing on the very low-level details of the ia32 ABI)
The system conditions for the OFW to Linux kernel transition are as
follows:

Quoted 8 lines

a) OFW can load the Linux kernel from either bzimage format or ELF format
(either uncompressed or zlib-compressed.) If the kernel is in ELF format with symbols, OFW can do symbolic kernel debugging. Further discussion will focus on bzimage format, as that is what OLPC uses and is also the "greased path" for kernel builds.
b) OFW loads the bzimage kernel at 0x100000 and the ramdisk image, if any,
at 0x800000.

so you are assuming that your uncompressed vmlinux only use less 8M space?

you are supposed to check the bzImage to get uncompressed vmlinux size.

Mitch Bradley

2008-04-21 08:22 UTC #65 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 11 lines · Yinghai Lu

Quoted 6 lines
path" for kernel builds.
b) OFW loads the bzimage kernel at 0x100000 and the ramdisk image, if any,
at 0x800000.
so you are assuming that your uncompressed vmlinux only use less 8M space?
you are supposed to check the bzImage to get uncompressed vmlinux size.

The 0x800000 ramdisk load address is an OLPC-specific firmware implementation detail that could easily be changed without affecting anything else. I probably shouldn't have mentioned it because it isn't really an integral part of the interface "contract".

I certainly hope that the OLPC kernel never gets anywhere near that size. The OLPC hardware has limited configurability, so it's not plausible that the kernel would grow that large to include a huge kit of drivers. If the kernel file becomes large as a result of including the initramfs in the same file, the 0x800000 ramdisk load address won't apply (because there won't be a separate load of the initramfs file), so the kernel could be extend way past that boundary with no problems.

If we get to the point where we do need huge kernels on OLPC, we can release a firmware upgrade along with the new OS. We have mechanisms for coordinating firmware and OS upgrades.

If a new customer for OFW on x86 appears, I'll remember to float the boundary above the bzImage uncompressed size (assuming that the bzimage format is still applicable when that happens).

Quoted 3 lines

YH

H. Peter Anvin

reviewer

2008-04-21 11:41 UTC #66 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

In reply to H. Peter Anvin

Quoted 3 lines · Mitch Bradley

d) OFW itself lives at the top of the virtual address space, just below the ROM. (The ROM is mapped virtual=physical for convenience) OFW uses RAM allocated from the top of physical memory, mapped at the aforementioned high virtual addresses. One page directory entry - the next to last one - is used for that RAM mapping and also for mapping additional miscellaneous I/O devices. The 8MB frame buffer requires 2 additional PDEs, just below. When Linux takes over the display, OFW no longer needs the frame buffer mapping, but it is convenient to preserve that mapping temporarily while using OFW as a debugger.

So let me see here... you want the virtual address range [0xffc00000, 0xfff00000) to be reserved for OFW, and you are prohibiting the kernel from using PAE?

Quoted 1 line

e) Low memory - everything except the ~1Meg that OFW lives in - is mapped virtual=physical.

Are you making this assumption when called from the kernel, too?

Quoted 6 lines

j) Linux must save the following information during early startup:
1) The callback function address - either from the initial value of eax
or from the OFW info block.
2) The the next-to-last page directory entry - just the pointer to the
page table. The page table itself lives in OFW's reserved memory.

Quoted 7 lines

k) When calling back into OFW, Linux must:
1) Establish a page directory that contains OFW's PDE (saved in j2
above) and that maps the client interface argument array, including any buffer pointers.
2) Call callback_function with the address of the argument array in
eax. (Ordinary 32-bit near call).
For all of the OLPC kernel's current callbacks into OFW, the requirements (j2) and (k1) are easily satisfied by "priming" swapper_pg_dir with the contents of the current page directory (as pointed to by the CR3 register).

I do not like it, simply because it amounts to "initialize this otherwise zero-initialized piece of data without making any kind of reservations and blindly hope nothing else overwrites it."

I'm also troubled with the assumption that the kernel doesn't use PAE. I realize that this is not an issue for OLPC, but it certainly makes this a less-than-generic solution.

Having mapped page table entries which are not under kernel control is a very serious problem for PAT - PAT requires, by hardware specification, the kernel to eliminate all potential aliases with different mappings.

One way to deal with this, of course, is to save the firmware-provided PGD and only use it for OFW calls. On the other hand, perhaps a better questions is to what extent it is needed at all.

Furthermore, since you're using a nonstandard OFW interface (not compliant with the x86 OFW binding document), all of this should be called something like OLPC_OFW to make it clear that it's the OLPC variant.

If I had designed this, I would probably have used an SMI; since you have control over the firmware you can do that. SMI saves the entire machine state including all the modes, cleans them all up for you, and puts it all back together at RSM time. It is slow, of course, but it completely decouples the firmware and the OS, which is why it's used.

-hpa

H. Peter Anvin

reviewer

2008-04-21 13:33 UTC #67 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Okay, stepping back a few steps, it's pretty clear that most of my objections aren't really an issue for Geode/OLPC; however, I *really* don't want others to pick it up as being "the" Open Firmware interface.

Within those constraints it makes sense to set up the PDEs in swapper_pg_dir and let them propagate using the normal mechanisms.

** This is assuming that your OF interface does not rely on a 1:1 mapping of low memory being present at the time it makes a call. If it *does*, then a separate page directory needs to be maintained for the OF class. **

Thus, I'm willing to accept this with these changes:

- Please name things specific to the interface (as opposed to Open Firmware in general, like the device tree) olpc_ofw or olpcfw, to denote that this is an OLPC-specific interface. Thus, CONFIG_OLPC_OPEN_FIRMWARE or something along those lines.

- Make it explicit in Kconfig that OLPC_OPEN_FIRMWARE conflicts with X86_PAE, 64BIT, or X86_PAT.

- Change VMALLOC_END in include/asm-x86/pgtable_32.h so the kernel will know to avoid this virtual memory range.

- Add a memory region to arch/x86/mm/dump_tabletables.c.

-hpa

H. Peter Anvin

reviewer

2008-04-21 13:34 UTC #68 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Within those constraints it makes sense to set up the PDEs in swapper_pg_dir and let them propagate using the normal mechanisms.

Thus, I'm willing to accept this with these changes:

- Make it explicit in Kconfig that OLPC_OPEN_FIRMWARE conflicts with X86_PAE, 64BIT, or X86_PAT.

- Change VMALLOC_END in include/asm-x86/pgtable_32.h so the kernel will know to avoid this virtual memory range.

- Add a memory region to arch/x86/mm/dump_tabletables.c.

-hpa

H. Peter Anvin

reviewer

2008-04-21 13:35 UTC #69 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

In reply to H. Peter Anvin

Within those constraints it makes sense to set up the PDEs in swapper_pg_dir and let them propagate using the normal mechanisms.

Thus, I'm willing to accept this with these changes:

- Make it explicit in Kconfig that OLPC_OPEN_FIRMWARE conflicts with X86_PAE, 64BIT, or X86_PAT.

- Change VMALLOC_END in include/asm-x86/pgtable_32.h so the kernel will know to avoid this virtual memory range.

- Add a memory region to arch/x86/mm/dump_tabletables.c.

-hpa

Jordan Crouse

2008-04-21 14:58 UTC #70 lore raw

Re: OLPC: Add support for calling into Open Firmware

Quoted 16 lines · Mitch Bradley

Quoted 5 lines · H. Peter Anvin
Quoted 3 lines · Mitch Bradley
The x86 architecture doesn't make this problem easy.
Quoted 7 lines
[long rant about the x86 architecture]
It would be more useful if you described the actual defined entry conditions from OpenFirmware look like, including if they are well-defined for all OF implementations or only for OLPC.
-hpa
Fair enough...

Quoted 14 lines

To get the second subquestion out of the way: At the present time, on the x86 architecture, "all OF implementations" and "OLPC" are effectively the same. I am unaware of any other x86 OFW deployments in current use. There have been some in the past, on bespoke systems such as Network Appliance servers and at least one settop box, but those have fallen by the wayside as those companies have shifted over to commodity PC hardware. The current market status quo is that x86 boards are primarily designed for Windows, and thus must run legacy BIOS, with some recent migration to EFI, neither of which are open source in the strong sense. While I would like to see more OFW penetration into the larger x86 market, I don't really expect it. x86 motherboard manufacturing is becoming more and more difficult as signal speeds increase, leading to a decline in the number of manufacturers. The existing manufacturers depend on Windows for sales volume and their internal procedures and working knowledge are based on legacy BIOS.

/me puts on his coreboot hat

This is off topic slightly, but let it be known that the coreboot project considers OFW a very valid option for x86 platforms. A kernel that worked happily with OFW would greatly encourage people to adopt it in lieu of other BIOS / firmware solutions.

I return you to your previously scheduled debate.

Jordan

H. Peter Anvin

reviewer

2008-04-21 15:02 UTC #71 lore raw

Re: OLPC: Add support for calling into Open Firmware

Quoted 8 lines · Jordan Crouse

This is off topic slightly, but let it be known that the coreboot project considers OFW a very valid option for x86 platforms. A kernel that worked happily with OFW would greatly encourage people to adopt it in lieu of other BIOS / firmware solutions.
I return you to your previously scheduled debate.

The interface they are proposing is definitely not suitable for upward extension, for the reasons already mentioned. However, they have units in the field, and the amount of changes required to support another interface should be relatively minor.

Hence my insistence that we don't promote it as *the* OFW interface, but *a* OFW interface.

-hpa

H. Peter Anvin

reviewer

2008-04-20 19:18 UTC #72 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 4 lines · Andres Salomon

I'm certainly open to suggestions.. Otherwise, I'll poke around and see if I can come up w/ something.

It pretty much depends on what the invariants look like. The normal/clean way of doing this kind of thing is via a fixmap entry and/or ioremap.

-hpa

Mitch Bradley

2008-04-21 03:10 UTC #73 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

In reply to Yinghai Lu

Quoted 121 lines · Yinghai Lu

Quoted 109 lines · Andres Salomon
This adds 32-bit support for calling into OFW from the kernel. It's useful for querying the firmware for misc hardware information, fetching the device tree, etc.
There's potentially no reason why other platforms couldn't use this, but currently OLPC is the main user of it.
This work was originally done by Mitch Bradley.
Signed-off-by Andres Salomon <[email protected]>
---
  arch/x86/Kconfig          |    8 +++++
  arch/x86/kernel/Makefile  |    1 +
  arch/x86/kernel/head_32.S |   27 ++++++++++++++++
  arch/x86/kernel/ofw.c     |   75 +++++++++++++++++++++++++++++++++++++++++++++
  include/asm-x86/ofw.h     |   50 ++++++++++++++++++++++++++++++
  include/asm-x86/setup.h   |    1 +
  6 files changed, 162 insertions(+), 0 deletions(-)
create mode 100644 arch/x86/kernel/ofw.c
create mode 100644 include/asm-x86/ofw.h
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 3b9089b..ce56105 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -661,6 +661,14 @@ config I8K
Say Y if you intend to run this kernel on a Dell Inspiron 8000. Say N otherwise.
+config OPEN_FIRMWARE
 +       bool "Support for Open Firmware"
 +       default y if OLPC
 +       ---help---
 +         This option adds support for the implementation of Open Firmware
 +         that is used on the OLPC XO laptop.
 +         If unsure, say N here.
 +
  config X86_REBOOTFIXUPS
        def_bool n
        prompt "Enable X86 board specific fixups for reboot"
 diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
 index 9575754..d33600e 100644
 --- a/arch/x86/kernel/Makefile
 +++ b/arch/x86/kernel/Makefile
 @@ -54,6 +54,7 @@ obj-$(CONFIG_X86_TRAMPOLINE)  += trampoline_$(BITS).o
  obj-$(CONFIG_X86_MPPARSE)      += mpparse_$(BITS).o
  obj-$(CONFIG_X86_LOCAL_APIC)   += apic_$(BITS).o nmi_$(BITS).o
  obj-$(CONFIG_X86_IO_APIC)      += io_apic_$(BITS).o
 +obj-$(CONFIG_OPEN_FIRMWARE)    += ofw.o
obj-$(CONFIG_X86_REBOOTFIXUPS) += reboot_fixups_32.o
obj-$(CONFIG_KEXEC) += machine_kexec_$(BITS).o
obj-$(CONFIG_KEXEC) += relocate_kernel_$(BITS).o crash.o
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S index 74d87ea..c9d2d00 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -132,6 +132,33 @@ ENTRY(startup_32)
movsl
1:
+#ifdef CONFIG_OPEN_FIRMWARE
+/*
+ * If Open Firmware booted us, save the OFW client interface callback address
+ * and preserve the OFW page mappings by priming the kernel's new page
+ * directory area with a copy of the OFW page directory. That lets OFW stay
+ * resident in high memory (high in both the virtual and physical spaces)
+ * for at least long enough to copy out the device tree.
+ */
 +       movl $pa(boot_params + OFW_INFO_OFFSET), %ebp
 +       cmpl $0x2057464F, (%ebp)                /* Magic number "OFW " */
 +       jne 4f
 +
 +       mov 0x8(%ebp), %eax                     /* Save callback address */
 +       mov %eax, pa(call_firmware)
 +
 +       /* Copy the OFW pdir into swapper_pg_dir */
 +       movl %esi, %edx         /* save %esi */
 +       movl $pa(swapper_pg_dir), %edi
 +       movl %cr3, %esi         /* Source is current pg_dir base address */
 +       movl $1024, %ecx        /* Number of page directory entries */
 +       rep
 +       movsl
 +       movl %edx, %esi         /* restore %esi */
 +4:
 +
 +#endif
 +
  #ifdef CONFIG_PARAVIRT
        /* This is can only trip for a broken bootloader... */
        cmpw $0x207, pa(boot_params + BP_version)
diff --git a/arch/x86/kernel/ofw.c b/arch/x86/kernel/ofw.c new file mode 100644
index 0000000..14036aa
--- /dev/null
+++ b/arch/x86/kernel/ofw.c
@@ -0,0 +1,75 @@
+/*
+ * Open Firmware client interface for 32-bit systems.
+ *
+ * Copyright (c) 2007 Mitch Bradley <[email protected]>
+ * Copyright (c) 2007-2008 Andres Salomon <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/kernel.h>
Quoted 9 lines
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <asm/ofw.h>
+
+/*
+ * This code is intended to be portable to any 32-bit Open Firmware
+ * implementation with a standard client interface that can be
+ * called when Linux is running.

Quoted 4 lines

how about changing to ofw_32.c?
YH

Is your suggestion to change the filename from "ofw.c" to "ofw_32.c"? That seems like a good idea to me.

Yinghai Lu

2008-04-21 03:15 UTC #74 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 6 lines · Mitch Bradley

Quoted 1 line · Yinghai Lu

Quoted 8 lines

Quoted 5 lines
how about changing to ofw_32.c?
YH
Is your suggestion to change the filename from "ofw.c" to "ofw_32.c"? That
seems like a good idea to me.

Yes.

BTW, why olpc need OFW runtime service?
why not just put the info in in ram with some signiture, so kernel/util just need to loot at the table if needed?

Mitch Bradley

2008-04-21 04:06 UTC #75 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 16 lines · Yinghai Lu

Quoted 4 lines · Mitch Bradley
Yinghai Lu wrote:
Quoted 9 lines
Quoted 6 lines
how about changing to ofw_32.c?
YH
Is your suggestion to change the filename from "ofw.c" to "ofw_32.c"? That
seems like a good idea to me.

Quoted 6 lines

Yes.
BTW, why olpc need OFW runtime service?
why not just put the info in in ram with some signiture, so kernel/util just need to loot at the table if needed?

In SPARC land, at least on SunOS and Solaris, it was very convenient for debugging to interrupt the OS with Stop-A and use OFW to inspect the system state. That was especially handy for live crash analysis. Dumps are useful as far as they go, but they often fail to capture detailed I/O device state.

I was hoping to do that on x86 too. So far we (OLPC) haven't implemented a sysrq hook to enter OFW, but I haven't given up hope yet. It doesn't cost much to leave OFW around, but once you decide to eject it, you can't easily get it back.

Apple made the early decision to eject OFW and just keep a device tree table. That decision was probably due to several factors, including the rather lame state of Apple's first OFW implementation and the complexity of their OS startup process at the time (which included "trampolining" to a 68000 emulator to run their legacy code). Once they went down that path, the die was cast, and the PowerPC community got used to the "OFW ends up as just a table" idea.

Quoted 1 line

David Miller

2008-04-21 04:27 UTC #76 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

From: Mitch Bradley <[email protected]>
Date: Sun, 20 Apr 2008 18:05:26 -1000

Quoted 5 lines

In SPARC land, at least on SunOS and Solaris, it was very convenient for debugging to interrupt the OS with Stop-A and use OFW to inspect the system state. That was especially handy for live crash analysis. Dumps are useful as far as they go, but they often fail to capture detailed I/O device state.

In most current SPARC systems, OFW is not usable and is completely forgotten right after bootup in order to accomodate LDOMs and CPU hotplug.

It's a better idea, anyways, to develop more pervasive and usable in-kernel debugger facilities. Then it doesn't matter if you have "cool" firmware or not. :-)

Yinghai Lu

2008-04-21 04:50 UTC #77 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 39 lines · Mitch Bradley

Quoted 8 lines · Yinghai Lu
Quoted 5 lines · Mitch Bradley
Yinghai Lu wrote:
Quoted 8 lines
Quoted 8 lines
Quoted 7 lines
how about changing to ofw_32.c?
YH
Is your suggestion to change the filename from "ofw.c" to "ofw_32.c"?
That
Quoted 11 lines
Quoted 3 lines
seems like a good idea to me.
Yes.
BTW, why olpc need OFW runtime service?
why not just put the info in in ram with some signiture, so kernel/util just need to loot at the table if needed?
In SPARC land, at least on SunOS and Solaris, it was very convenient for
debugging to interrupt the OS with Stop-A and use OFW to inspect the system state. That was especially handy for live crash analysis. Dumps are useful as far as they go, but they often fail to capture detailed I/O device state.

Quoted 4 lines

I was hoping to do that on x86 too. So far we (OLPC) haven't implemented a
sysrq hook to enter OFW, but I haven't given up hope yet. It doesn't cost much to leave OFW around, but once you decide to eject it, you can't easily get it back.

geode is using SMI to simulate the pci conf space, wonder that could be problem.

later you have 64 runtime service for 64 platform like UEFI?

Mitch Bradley

2008-04-21 08:04 UTC #78 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 4 lines · Yinghai Lu

Quoted 1 line
...
geode is using SMI to simulate the pci conf space, wonder that could be problem.

On the current OLPC system, we don't use the SMI-based PCI config space simulator. The code for that "VSA" module is only partially open sourced (some of it is open, and some of it is just not available). The parts of it for which we do have source can only be compiled with an old proprietary toolchain that is no longer available.

Instead of using the SMI-based simulation, we have added a PCI configuration access method in the kernel that supplies the necessary information from a table. The code for that hardware-specific access method is roughly 40 lines of code plus a few data tables.

In the past few weeks, I have developed a rather complete Open Firmware-based reimplementation of the SMI PCI config hardware emulator. All-told, it requires over 1000 lines. It remains to be seen whether the complicated version will ultimately be deployed. Personally, I find it distasteful to use a lot of code to make the hardware pretend that it is something other than what it really is, when a much smaller driver works just as well. The SMI-based emulator is quite difficult to understand and maintain, because the Geode SMI handling mechanism is complex, incompletely documented, and suffers from many of the multiple-mode-switches problems as real-mode to protected-mode gateway code.

Quoted 2 lines

later you have 64 runtime service for 64 platform like UEFI?

Possibly. 64-bit systems are not a problem per se - there have been 64-bit OFW implementations for 64-bit architectures like SPARC and Alpha dating back to a long time ago. The main issue from my point of view is whether or not there is a customer to motivate the work.

Quoted 3 lines

YH

Andres Salomon

2008-04-21 14:21 UTC #79 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 32 lines · On Sun, 20 Apr 2008 18:05:26 -1000 Mitch Bradley <[email protected]> wrote:

Quoted 16 lines · Yinghai Lu
Quoted 4 lines · Mitch Bradley
Yinghai Lu wrote:
Quoted 9 lines
Quoted 6 lines
how about changing to ofw_32.c?
YH
Is your suggestion to change the filename from "ofw.c" to "ofw_32.c"? That
seems like a good idea to me.
Quoted 6 lines
Yes.
BTW, why olpc need OFW runtime service?
why not just put the info in in ram with some signiture, so kernel/util just need to loot at the table if needed?
In SPARC land, at least on SunOS and Solaris, it was very convenient for debugging to interrupt the OS with Stop-A and use OFW to inspect the system state. That was especially handy for live crash analysis. Dumps are useful as far as they go, but they often fail to capture detailed I/O device state.

Quoted 5 lines

I was hoping to do that on x86 too. So far we (OLPC) haven't implemented a sysrq hook to enter OFW, but I haven't given up hope yet. It doesn't cost much to leave OFW around, but once you decide to eject it, you can't easily get it back.

I'm not actually convinced that we *do* want to keep OFW resident in memory, especially given the memory tricks we need to play. I also don't actually like the OFW interface that we. The debugging aspect of it was a compelling argument up until a week ago (when kernel debuggers started finally finding their way into the kernel).

However, until we clean up the promfs stuff, there's no chance of getting an OFW device tree upstream.

Quoted 9 lines

Apple made the early decision to eject OFW and just keep a device tree table. That decision was probably due to several factors, including the rather lame state of Apple's first OFW implementation and the complexity of their OS startup process at the time (which included "trampolining" to a 68000 emulator to run their legacy code). Once they went down that path, the die was cast, and the PowerPC community got used to the "OFW ends up as just a table" idea.
Quoted 1 line

Signature · 1 line

Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

David Woodhouse

2008-04-21 15:55 UTC #80 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 5 lines · Andres Salomon

I'm not actually convinced that we *do* want to keep OFW resident in memory, especially given the memory tricks we need to play. I also don't actually like the OFW interface that we. The debugging aspect of it was a compelling argument up until a week ago (when kernel debuggers started finally finding their way into the kernel).

I don't actually think that the debugging aspect was _ever_ a compelling argument. It might have made it theoretically possible for _Mitch_ to debug kernel problems, should he be inclined to do so -- but for the rest of us mere mortals it's just a PITA trying to keep OpenFirmware live. A gdb stub is much more useful, in my experience.

Quoted 2 lines

However, until we clean up the promfs stuff, there's no chance of getting an OFW device tree upstream.

I see no reason why we shouldn't be able to create a 'flattened' device-tree during early boot, like the PowerPC kernel does. And use it thereafter, having quiesced OpenFirmware. Haven't we already been working on unifying this between SPARC and PowerPC kernels?

I definitely don't think we need to play these tricks to keep OpenFirmware resident while the kernel is running. Take a look at your second patch -- it's _all_ just lookups in the device-tree, and you're inventing a new way to do it instead of using the existing one.

Signature · 1 line

dwmw2

Andres Salomon

2008-04-21 17:00 UTC #81 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 13 lines · On Mon, 21 Apr 2008 16:54:13 +0100 David Woodhouse <[email protected]> wrote:

Quoted 5 lines · Andres Salomon
I'm not actually convinced that we *do* want to keep OFW resident in memory, especially given the memory tricks we need to play. I also don't actually like the OFW interface that we. The debugging aspect of it was a compelling argument up until a week ago (when kernel debuggers started finally finding their way into the kernel).
I don't actually think that the debugging aspect was _ever_ a compelling argument. It might have made it theoretically possible for _Mitch_ to debug kernel problems, should he be inclined to do so -- but for the rest of us mere mortals it's just a PITA trying to keep OpenFirmware live. A gdb stub is much more useful, in my experience.

Quoted 7 lines

Quoted 2 lines
However, until we clean up the promfs stuff, there's no chance of getting an OFW device tree upstream.
I see no reason why we shouldn't be able to create a 'flattened'b device-tree during early boot, like the PowerPC kernel does. And use it thereafter, having quiesced OpenFirmware. Haven't we already been working on unifying this between SPARC and PowerPC kernels?

Quite simply, it's a lot more work (*and* we have to play nice w/ sparc and ppc). I had intended to eventually do it, but first I wanted to get this stuff in for 2.6.26 so that we could at least boot upstream kernels on XOs.

I was also hoping to not get into this conversation, but alas.. too late. :)

Quoted 6 lines

I definitely don't think we need to play these tricks to keep OpenFirmware resident while the kernel is running. Take a look at your second patch -- it's _all_ just lookups in the device-tree, and you're inventing a new way to do it instead of using the existing one.

Signature · 1 line

Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

David Woodhouse

2008-04-21 19:18 UTC #82 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 2 lines · Andres Salomon

Quite simply, it's a lot more work (*and* we have to play nice w/ sparc and ppc).

It's only more work because we did it the wrong way in the first place. If only someone had pointed it out at the time... :)

For interaction with device-tree properties in generic code, you should be using the functions defined in <linux/of.h>.

Creating the static device-tree before we quiesce OpenFirmware surely shouldn't be so hard? Can't we cut and paste most of that code anyway?

Quoted 3 lines

I had intended to eventually do it, but first I wanted
to get this stuff in for 2.6.26 so that we could at least boot upstream kernels on XOs.

Is it only the things in your second patch which need to be made to work? One of them was already working, by grubbing around in the BIOS directly -- so all we need is the board revision, isn't it? Can we get that from the EC for now?

Signature · 1 line

dwmw2

Andres Salomon

2008-04-21 19:43 UTC #83 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 7 lines · On Mon, 21 Apr 2008 20:18:11 +0100 David Woodhouse <[email protected]> wrote:

Quoted 2 lines · Andres Salomon
Quite simply, it's a lot more work (*and* we have to play nice w/ sparc and ppc).
It's only more work because we did it the wrong way in the first place. If only someone had pointed it out at the time... :)

Yes, and if only we had an infinite number of kernel hackers who had time to work on such things, then we could've done things differently.

Quoted 3 lines

For interaction with device-tree properties in generic code, you should be using the functions defined in <linux/of.h>.

At the time [the OFW interface] was written, linux/of.h didn't *exist*.

Quoted 3 lines

Creating the static device-tree before we quiesce OpenFirmware surely shouldn't be so hard? Can't we cut and paste most of that code anyway?

We're not adding a device tree right now, we're adding a method for querying OFW for information. Eventually that information should be obtained from a device tree. However, that's going to take additional time, and I'd like to get rid of some of these patches that we've been carrying around.

Quoted 9 lines

Quoted 3 lines
I had intended to eventually do it, but first I wanted
to get this stuff in for 2.6.26 so that we could at least boot upstream kernels on XOs.
Is it only the things in your second patch which need to be made to work? One of them was already working, by grubbing around in the BIOS directly -- so all we need is the board revision, isn't it? Can we get that from the EC for now?

Well, no, it wasn't already working; that's the reason this whole thread started. It was crashing someone's machine. That's why the OFW interface, as imperfect as it is, is an _improvement_.

Signature · 1 line

Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

David Woodhouse

2008-04-21 20:25 UTC #84 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 3 lines · Andres Salomon

Well, no, it wasn't already working; that's the reason this whole thread started. It was crashing someone's machine. That's why the OFW interface, as imperfect as it is, is an _improvement_.

You're proposing a new interface between bootloader and kernel as a temporary hack just to work around that until we fix it properly?

That seems like overkill to me. I'd just go for is_geode() as you suggested, and maybe PCI configuration tricks to detect the lack of VSA so we can be _fairly_ sure it's OLPC before we poke at it?

Or why not try '!page_is_ram(0xffffffc0 >> PAGE_SHIFT)' if it's just to avoid that particular warning? :)

Signature · 1 line

dwmw2

Andres Salomon

2008-04-21 21:00 UTC #85 lore raw

PATCH 1/1 OLPC: only check for OFW signature on VSA-less Geodes

Quoted 8 lines · On Mon, 21 Apr 2008 21:25:17 +0100 David Woodhouse <[email protected]> wrote:

Quoted 3 lines · Andres Salomon
Well, no, it wasn't already working; that's the reason this whole thread started. It was crashing someone's machine. That's why the OFW interface, as imperfect as it is, is an _improvement_.
You're proposing a new interface between bootloader and kernel as a temporary hack just to work around that until we fix it properly?

Quoted 7 lines

That seems like overkill to me. I'd just go for is_geode() as you suggested, and maybe PCI configuration tricks to detect the lack of VSA so we can be _fairly_ sure it's OLPC before we poke at it?
Or why not try '!page_is_ram(0xffffffc0 >> PAGE_SHIFT)' if it's just to avoid that particular warning? :)

Okay, does anyone have a problem with this?

The OFW sig check requires an ioremap that is dangerous on non-OLPC systems. Long term, we should be getting the signature from the device tree (/openprom/model), but for right now just limit the check to only run on a subset of Geode (GX2/LX) systems.

---
 arch/x86/kernel/olpc.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

arch/x86/kernel/olpc.c +4 -0

index 11670be..3e66722 100644
@@ -211,6 +211,10 @@ static int __init olpc_init(void)
 {
 	unsigned char *romsig;
 
+	/* The ioremap check is dangerous; limit what we run it on */
+	if (!is_geode() || geode_has_vsa2())
+		return 0;
+
 	spin_lock_init(&ec_lock);
 
 	romsig = ioremap(0xffffffc0, 16);

Signature · 5 lines

1.5.4.4


-- 
Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Jordan Crouse

Acked-by on v1 1/1

2008-04-21 21:10 UTC #86 lore raw

Re: OLPC: only check for OFW signature on VSA-less Geodes

Quoted 27 lines · Andres Salomon

Quoted 8 lines · On Mon, 21 Apr 2008 21:25:17 +0100 David Woodhouse <[email protected]> wrote:
Quoted 3 lines · Andres Salomon
Well, no, it wasn't already working; that's the reason this whole thread started. It was crashing someone's machine. That's why the OFW interface, as imperfect as it is, is an _improvement_.
You're proposing a new interface between bootloader and kernel as a temporary hack just to work around that until we fix it properly?
Quoted 7 lines
That seems like overkill to me. I'd just go for is_geode() as you suggested, and maybe PCI configuration tricks to detect the lack of VSA so we can be _fairly_ sure it's OLPC before we poke at it?
Or why not try '!page_is_ram(0xffffffc0 >> PAGE_SHIFT)' if it's just to avoid that particular warning? :)
Okay, does anyone have a problem with this?

Quoted 6 lines

The OFW sig check requires an ioremap that is dangerous on non-OLPC systems. Long term, we should be getting the signature from the device tree (/openprom/model), but for right now just limit the check to only run on a subset of Geode (GX2/LX) systems.
Signed-off-by Andres Salomon <[email protected]>

Acked-by Jordan Crouse <[email protected]>

Quoted 22 lines

arch/x86/kernel/olpc.c +4 -0

index 11670be..3e66722 100644
@@ -211,6 +211,10 @@ static int __init olpc_init(void)
 {
 	unsigned char *romsig;
 
+	/* The ioremap check is dangerous; limit what we run it on */
+	if (!is_geode() || geode_has_vsa2())
+		return 0;
+
 	spin_lock_init(&ec_lock);
 
 	romsig = ioremap(0xffffffc0, 16);

Signature · 5 lines

1.5.4.4


-- 
Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Signature · 3 lines

Jordan Crouse
Systems Software Development Engineer 
Advanced Micro Devices, Inc.

David Woodhouse

Acked-by on v1 1/1

2008-04-21 21:18 UTC #87 lore raw

Re: [PATCH] OLPC: only check for OFW signature on VSA-less Geodes

Quoted 4 lines · Andres Salomon

The OFW sig check requires an ioremap that is dangerous on non-OLPC systems. Long term, we should be getting the signature from the device tree (/openprom/model), but for right now just limit the check to only run on a subset of Geode (GX2/LX) systems.

That looks saner to me for now.

Acked-by David Woodhouse <[email protected]>

Signature · 1 line

dwmw2

Andrew Morton

author maintainer

2008-04-29 03:07 UTC #88 lore raw

Re: [PATCH] OLPC: only check for OFW signature on VSA-less Geodes

arch/x86/kernel/olpc.c static int __init olpc_init(void) · line 215

Quoted 1 line · On Mon, 21 Apr 2008 17:02:30 -0400 Andres Salomon <[email protected]> wrote:

arch/x86/kernel/olpc.c +1 -0
+	if (!is_geode() || geode_has_vsa2())

geode_has_vsa2() is a fairly expensive-looking function and afacit only needs to be evaluated once per boot. Perhaps we should cache it somewhere?

Andres Salomon

2008-04-29 05:29 UTC #89 lore raw

PATCH 1/1 x86: GEODE: cache results from geode_has_vsa2() and uninline

Quoted 7 lines · On Mon, 28 Apr 2008 20:06:51 -0700 Andrew Morton <[email protected]>…

Review
Quoted 1 line · On Mon, 21 Apr 2008 17:02:30 -0400 Andres Salomon <[email protected]> wrote:
(quoted diff) +1 -0
+	if (!is_geode() || geode_has_vsa2())
geode_has_vsa2() is a fairly expensive-looking function and afacit only needs to be evaluated once per boot. Perhaps we should cache it somewhere?

How about this?

This moves geode_has_vsa2 into a .c file, caches the result we get from the VSA virtual registers, and causes the function to no longer be inline.

In reply to David Woodhouse

---
 arch/x86/kernel/geode_32.c |   19 +++++++++++++++++++
 include/asm-x86/geode.h    |   11 +----------
 2 files changed, 20 insertions(+), 10 deletions(-)

arch/x86/kernel/geode_32.c +19 -0

index 9dad6ca..1cb8225 100644
@@ -161,6 +161,25 @@ void geode_gpio_setup_event(unsigned int gpio, int pair, int pme)
 }
 EXPORT_SYMBOL_GPL(geode_gpio_setup_event);
 
+static int has_vsa2 = -1;
+
+int geode_has_vsa2(void)
+{
+	if (has_vsa2 == -1) {
+		/*
+		 * The VSA has virtual registers that we can query for a
+		 * signature.
+		 */
+		outw(VSA_VR_UNLOCK, VSA_VRC_INDEX);
+		outw(VSA_VR_SIGNATURE, VSA_VRC_INDEX);
+
+		has_vsa2 = (inw(VSA_VRC_DATA) == VSA_SIG);
+	}
+
+	return has_vsa2;
+}
+EXPORT_SYMBOL_GPL(geode_has_vsa2);
+
 static int __init geode_southbridge_init(void)
 {
 	if (!is_geode())

include/asm-x86/geode.h +1 ~~-10~~

index 7154dc4..8a53bc8 100644
@@ -185,16 +185,7 @@ static inline int is_geode(void)
 	return (is_geode_gx() || is_geode_lx());
 }
 
-/*
- * The VSA has virtual registers that we can query for a signature.
- */
-static inline int geode_has_vsa2(void)
-{
-	outw(VSA_VR_UNLOCK, VSA_VRC_INDEX);
-	outw(VSA_VR_SIGNATURE, VSA_VRC_INDEX);
-
-	return (inw(VSA_VRC_DATA) == VSA_SIG);
-}
+extern int geode_has_vsa2(void);
 
 /* MFGPTs */

Signature · 1 line

1.5.5

Andrew Morton

author maintainer

2008-04-29 20:41 UTC #90 lore raw

Re: [PATCH] x86: GEODE: cache results from geode_has_vsa2() and uninline

Quoted 5 lines · On Tue, 29 Apr 2008 01:32:13 -0400 Andres Salomon <[email protected]> wrote:

Quoted 2 lines · On Mon, 28 Apr 2008 20:06:51 -0700 Andrew Morton <[email protected]>…
On Mon, 21 Apr 2008 17:02:30 -0400 Andres Salomon <[email protected]> wrote:

Quoted 8 lines

Quoted 5 lines
Review
Quoted 1 line
(quoted diff) +1 -0
+	if (!is_geode() || geode_has_vsa2())
geode_has_vsa2() is a fairly expensive-looking function and afacit only needs to be evaluated once per boot. Perhaps we should cache it somewhere?
How about this?

Looks sane. Although one wonders if it should be cached as one of the standard x86 feature bit thingies which show up in /proc/cpuinfo's 'flags' field.

Quoted 8 lines

(quoted diff) +8 -0

+static int has_vsa2 = -1;
+
+int geode_has_vsa2(void)
+{
+	if (has_vsa2 == -1) {
+		/*
+		 * The VSA has virtual registers that we can query for a
+		 * signature.

Review

Quoted 10 lines

(quoted diff) +10 -0

+		 */
+		outw(VSA_VR_UNLOCK, VSA_VRC_INDEX);
+		outw(VSA_VR_SIGNATURE, VSA_VRC_INDEX);
+
+		has_vsa2 = (inw(VSA_VRC_DATA) == VSA_SIG);
+	}
+
+	return has_vsa2;
+}
+EXPORT_SYMBOL_GPL(geode_has_vsa2);

nit:

arch/x86/kernel/geode_32.c +2 -2

@@ -161,10 +161,10 @@ void geode_gpio_setup_event(unsigned int
 }
 EXPORT_SYMBOL_GPL(geode_gpio_setup_event);
 
-static int has_vsa2 = -1;
-
 int geode_has_vsa2(void)
 {
+	static int has_vsa2 = -1;
+
 	if (has_vsa2 == -1) {
 		/*
 		 * The VSA has virtual registers that we can query for a

Andres Salomon

Acked-by

2008-04-29 20:54 UTC #91 lore raw

Re: [PATCH] x86: GEODE: cache results from geode_has_vsa2() and uninline

Quoted 14 lines · On Tue, 29 Apr 2008 13:35:12 -0700 Andrew Morton <[email protected]>…

Quoted 11 lines · On Tue, 29 Apr 2008 01:32:13 -0400 Andres Salomon <[email protected]> wrote:
Quoted 7 lines · On Mon, 28 Apr 2008 20:06:51 -0700 Andrew Morton <[email protected]>…
Review
Quoted 1 line · On Mon, 21 Apr 2008 17:02:30 -0400 Andres Salomon <[email protected]> wrote:
(quoted diff) +1 -0
+	if (!is_geode() || geode_has_vsa2())
geode_has_vsa2() is a fairly expensive-looking function and afacit only needs to be evaluated once per boot. Perhaps we should cache it somewhere?

Quoted 7 lines

Quoted 2 lines
How about this?
Looks sane. Although one wonders if it should be cached as one of the standard x86 feature bit thingies which show up in /proc/cpuinfo's 'flags' field.

The VSA lives in a weird place between hardware and BIOS. I'm not really sure whether it's appropriate for it to be an x86_cap_flags (it hadn't occurred to me), but I think of it more as BIOS. Jordan, what do you think?

Quoted 24 lines

Quoted 8 lines

(quoted diff) +8 -0

+static int has_vsa2 = -1;
+
+int geode_has_vsa2(void)
+{
+	if (has_vsa2 == -1) {
+		/*
+		 * The VSA has virtual registers that we can query for a
+		 * signature.

Review

Quoted 10 lines

(quoted diff) +10 -0

+		 */
+		outw(VSA_VR_UNLOCK, VSA_VRC_INDEX);
+		outw(VSA_VR_SIGNATURE, VSA_VRC_INDEX);
+
+		has_vsa2 = (inw(VSA_VRC_DATA) == VSA_SIG);
+	}
+
+	return has_vsa2;
+}
+EXPORT_SYMBOL_GPL(geode_has_vsa2);

nit:

arch/x86/kernel/geode_32.c void geode_gpio_setup_event(unsigned int · line 167

Quoted 9 lines

arch/x86/kernel/geode_32.c +2 -2

@@ -161,10 +161,10 @@ void geode_gpio_setup_event(unsigned int
 }
 EXPORT_SYMBOL_GPL(geode_gpio_setup_event);
 
-static int has_vsa2 = -1;
-
 int geode_has_vsa2(void)
 {
+	static int has_vsa2 = -1;
+

Looks good.

Acked-by Andres Salomon <[email protected]>

Quoted 4 lines

arch/x86/kernel/geode_32.c +0 -0

 	if (has_vsa2 == -1) {
 		/*
 		 * The VSA has virtual registers that we can query for a

H. Peter Anvin

reviewer

2008-04-21 17:05 UTC #92 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 9 lines · David Woodhouse

Quoted 3 lines
However, until we clean up the promfs stuff, there's no chance of getting an OFW device tree upstream.
I see no reason why we shouldn't be able to create a 'flattened' device-tree during early boot, like the PowerPC kernel does. And use it thereafter, having quiesced OpenFirmware. Haven't we already been working on unifying this between SPARC and PowerPC kernels?

Quoted 5 lines

I definitely don't think we need to play these tricks to keep OpenFirmware resident while the kernel is running. Take a look at your second patch -- it's _all_ just lookups in the device-tree, and you're inventing a new way to do it instead of using the existing one.

If so, would this apply to OLPC as well?

-hpa

David Woodhouse

2008-04-21 18:54 UTC #93 lore raw

Re: [PATCH 1/2] OLPC: Add support for calling into Open Firmware

Quoted 10 lines · H. Peter Anvin

Quoted 9 lines · David Woodhouse
Quoted 3 lines
However, until we clean up the promfs stuff, there's no chance of getting an OFW device tree upstream.
I see no reason why we shouldn't be able to create a 'flattened' device-tree during early boot, like the PowerPC kernel does. And use it thereafter, having quiesced OpenFirmware. Haven't we already been working on unifying this between SPARC and PowerPC kernels?

Quoted 7 lines

Quoted 5 lines
I definitely don't think we need to play these tricks to keep OpenFirmware resident while the kernel is running. Take a look at your second patch -- it's _all_ just lookups in the device-tree, and you're inventing a new way to do it instead of using the existing one.
If so, would this apply to OLPC as well?

Yes. The 'second patch' to which I refer is the one which makes OLPC platform code use the calls in OpenFirmware... all of them gratuitous.

Signature · 1 line

dwmw2

Andres Salomon

2008-04-19 17:37 UTC #94 lore raw

PATCH 2/2 OLPC: drop pre-OpenFirmware workarounds

Prior to including OFW kernel support, we had to work around the lack of OFW. Once OFW support is added, we can switch to using it. This cleans up some pre-OFW model detection and OFW signature detection.

Note: this should be a bit nicer to non-OLPC hardware.

---
 arch/x86/kernel/olpc.c |   43 +++++++++++++++++++++++++++++--------------
 1 files changed, 29 insertions(+), 14 deletions(-)

arch/x86/kernel/olpc.c +29 ~~-14~~

index 11670be..3a05683 100644
@@ -190,11 +190,11 @@ EXPORT_SYMBOL_GPL(olpc_ec_cmd);
 static void __init platform_detect(void)
 {
 	size_t propsize;
-	u32 rev;
+	uint32_t rev;
 
 	if (ofw("getprop", 4, 1, NULL, "board-revision-int", &rev, 4,
 			&propsize) || propsize != 4) {
-		printk(KERN_ERR "ofw: getprop call failed!\n");
+		printk(KERN_ERR "olpc:  ofw getprop call failed!\n");
 		rev = 0;
 	}
 	olpc_platform_info.boardrev = be32_to_cpu(rev);

@@ -207,26 +207,43 @@ static void __init platform_detect(void)
 }
 #endif
 
-static int __init olpc_init(void)
+static int __init ofw_detect(void)
 {
-	unsigned char *romsig;
+	size_t propsize;
+	char romsig[20];
+	ofw_phandle phandle;
 
-	spin_lock_init(&ec_lock);
+	/* Fetch /openprom/model */
+	if (ofw("finddevice", 1, 1, "/openprom", &phandle) || phandle == ~0)
+		return -ENODEV;
 
-	romsig = ioremap(0xffffffc0, 16);
-	if (!romsig)
-		return 0;
+	if (ofw("getprop", 4, 1, phandle, "model", &romsig, sizeof(romsig),
+			&propsize) || propsize < 7)
+		return -ENODEV;
 
+	/* String should look something like "CL1   Q2D08  Q2D" */
 	if (strncmp(romsig, "CL1   Q", 7))
-		goto unmap;
+		return -ENODEV;
 	if (strncmp(romsig+6, romsig+13, 3)) {
-		printk(KERN_INFO "OLPC BIOS signature looks invalid.  "
+		printk(KERN_INFO "olpc:  BIOS signature looks invalid.  "
 				"Assuming not OLPC\n");
-		goto unmap;
+		return -ENODEV;
 	}
 
-	printk(KERN_INFO "OLPC board with OpenFirmware %.16s\n", romsig);
+	/* Looks like we have OLPC's OFW */
 	olpc_platform_info.flags |= OLPC_F_PRESENT;
+	printk(KERN_INFO "olpc:  board with OpenFirmware %.16s\n", romsig);
+
+	return 0;
+}
+
+static int __init olpc_init(void)
+{
+	spin_lock_init(&ec_lock);
+
+	/* ensure OFW is available */
+	if (ofw_detect())
+		return 0;
 
 	/* get the platform revision */
 	platform_detect();

@@ -248,8 +265,6 @@ static int __init olpc_init(void)
 			olpc_platform_info.boardrev >> 4,
 			olpc_platform_info.ecver);
 
-unmap:
-	iounmap(romsig);
 	return 0;
 }

Signature · 6 lines

1.5.4.4



-- 
Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Andrew Morton

author maintainer

2008-04-19 17:38 UTC #95 lore raw

Quoted 86 lines · On Sat, 19 Apr 2008 09:25:44 -0400 Andres Salomon <[email protected]> wrote:

Quoted 69 lines · On Fri, 18 Apr 2008 20:29:25 -0700 Andrew Morton <[email protected]>…
Quoted 9 lines · On Fri, 18 Apr 2008 23:10:24 -0400 Joseph Fannin <[email protected]> wrote:
Quoted 2 lines · Andrew Morton
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/
New, in 2.6.25-mm1 is a hang I'm seeing, just after the kernel prints:
"[ 0.160375] NET: Registered protocol family 16"
The hang lasts about five minutes, and then boot continues.
Please add initcall_debug to the kernel boot command line - that should narrow it down.
Quoted 35 lines
Just
after that, a backtrace is printed; I don't know if it's related. The backtrace will follow.
This does not occur in mainline. It seems it might be related to OLPC support -- I enabled all those options -- but that's not good behavior, and I see no warning of thus in the help.
I'm sending a number or reports against 2.6.25-mm1, so I've put my dmesg and .config on a server:
http://home.columbus.rr.com/jfannin3/dmesg.txt
http://home.columbus.rr.com/jfannin3/config-2.6.25-mm1.txt
[    0.160375] NET: Registered protocol family 16
[  400.782683] ------------[ cut here ]------------
[  400.782832] WARNING: at arch/x86/mm/ioremap.c:158 __ioremap_caller+0x27d/0x2e0()
[  400.783022] Modules linked in:
[  400.783169] Pid: 1, comm: swapper Not tainted 2.6.25-mm1 #7
[  400.783300]  [<c0130fa9>] warn_on_slowpath+0x59/0x80
[  400.783480]  [<c0106c2e>] ? profile_pc+0x3e/0x50
[  400.783682]  [<c01374ee>] ? irq_exit+0x4e/0xa0
[  400.783879]  [<c0115aec>] ? smp_apic_timer_interrupt+0x5c/0x90
[  400.784087]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784298]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.784506]  [<c024314c>] ? trace_hardirqs_on_thunk+0xc/0x10
[  400.784706]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.784906]  [<c011d0e6>] ? page_is_ram+0xa6/0xd0
[  400.785059]  [<c011d4ed>] __ioremap_caller+0x27d/0x2e0
[  400.785221]  [<c03569d8>] ? _spin_unlock_irqrestore+0x48/0x80
[  400.785421]  [<c017f4cd>] ? ftrace_record_ip+0x7d/0x250
[  400.785621]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.785817]  [<c011d59f>] ioremap_nocache+0x1f/0x30
[  400.785976]  [<c0474801>] ? olpc_init+0x31/0x140
[  400.786165]  [<c0474801>] olpc_init+0x31/0x140
Quoted 8 lines
[  400.786318]  [<c0464992>] kernel_init+0x142/0x2d0
[  400.786479]  [<c01552cd>] ? trace_hardirqs_on_caller+0xcd/0x150
[  400.786680]  [<c010416c>] ? restore_nocheck_notrace+0x0/0xe
[  400.786879]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787069]  [<c0464850>] ? kernel_init+0x0/0x2d0
[  400.787260]  [<c0104d9b>] kernel_thread_helper+0x7/0x10
[  400.787422]  =======================
[  400.787727] ---[ end trace 4eaa2a86a8e2da22 ]---
<looks at this again>
That's
                WARN_ON_ONCE(is_ram);
the changelog for the patch which added that warning is information-free and there's no code comment explaining what went wrong, which makes things rather harder than they ought to be.
Quoted 7 lines
Yes it's due to the new OLPC code. olpc_init() has
	romsig = ioremap(0xffffffc0, 16);
which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed.
Hm. We could either protect that code with an:
if (!is_geode())
return;

Quoted 7 lines

Or I could add the OpenFirmware patches which would allow us to get rid of this code, and instead check for the existence of OFW using that.
The former is quick and easy; the latter is (imo) nicer, so long as people don't have problems w/ the OFW code. :)

Do both ;)

The quick-n-easy version sounds suitable for now.

Andres Salomon

2008-04-19 17:48 UTC #96 lore raw

Quoted 7 lines · On Sat, 19 Apr 2008 10:38:33 -0700 Andrew Morton <[email protected]>…

Quoted 7 lines · On Sat, 19 Apr 2008 09:25:44 -0400 Andres Salomon <[email protected]> wrote:
Quoted 3 lines · On Fri, 18 Apr 2008 20:29:25 -0700 Andrew Morton <[email protected]>…
Quoted 1 line · On Fri, 18 Apr 2008 23:10:24 -0400 Joseph Fannin <[email protected]> wrote:
On Fri, Apr 18, 2008 at 01:47:57AM -0700, Andrew Morton wrote:

[...]

Quoted 14 lines

Quoted 14 lines
Quoted 4 lines
which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed.
Hm. We could either protect that code with an:
if (!is_geode())
return;
Or I could add the OpenFirmware patches which would allow us to get rid of this code, and instead check for the existence of OFW using that.

Quoted 7 lines

Quoted 3 lines
The former is quick and easy; the latter is (imo) nicer, so long as people don't have problems w/ the OFW code. :)
Do both ;)
The quick-n-easy version sounds suitable for now.

Heh, I already had sent the nicer version. If people have some fundamental problem w/ it, I can send the quick-n-easy version.

Signature · 1 line

Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Jordan Crouse

2008-04-21 14:53 UTC #97 lore raw

Quoted 29 lines · Andres Salomon

Quoted 7 lines · On Sat, 19 Apr 2008 10:38:33 -0700 Andrew Morton <[email protected]>…
Quoted 7 lines · On Sat, 19 Apr 2008 09:25:44 -0400 Andres Salomon <[email protected]> wrote:
Quoted 3 lines · On Fri, 18 Apr 2008 20:29:25 -0700 Andrew Morton <[email protected]>…
Quoted 1 line · On Fri, 18 Apr 2008 23:10:24 -0400 Joseph Fannin <[email protected]> wrote:
On Fri, Apr 18, 2008 at 01:47:57AM -0700, Andrew Morton wrote:
[...]
Quoted 18 lines
Quoted 17 lines
Quoted 4 lines
which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed.
Hm. We could either protect that code with an:
if (!is_geode())
return;
Or I could add the OpenFirmware patches which would allow us to get rid of this code, and instead check for the existence of OFW using that.
The former is quick and easy; the latter is (imo) nicer, so long as people don't have problems w/ the OFW code. :)

Quoted 6 lines

Quoted 3 lines
Do both ;)
The quick-n-easy version sounds suitable for now.
Heh, I already had sent the nicer version. If people have some fundamental problem w/ it, I can send the quick-n-easy version.

I prefer the nicer version. It is not a good policy IMHO to wrap OLPC specfic code with is_geode() and friends. Even by Geode standards, we've abused the code greatly for the benefit of the Geode, and few of those abuses would translate very well even to the general Geode community. I would prefer that we use the is_olpc() and #ifdef wrappers to ensure that the code that is exclusively OLPC stays exclusively OLPC.

Thanks,
Jordan

Signature · 3 lines

Jordan Crouse
Systems Software Development Engineer 
Advanced Micro Devices, Inc.

Andres Salomon

2008-04-21 15:03 UTC #98 lore raw

Quoted 44 lines · On Mon, 21 Apr 2008 08:56:19 -0600 Jordan Crouse <[email protected]> wrote:

Quoted 29 lines · Andres Salomon
Quoted 7 lines · On Sat, 19 Apr 2008 10:38:33 -0700 Andrew Morton <[email protected]>…
Quoted 7 lines · On Sat, 19 Apr 2008 09:25:44 -0400 Andres Salomon <[email protected]> wrote:
Quoted 3 lines · On Fri, 18 Apr 2008 20:29:25 -0700 Andrew Morton <[email protected]>…
Quoted 1 line · On Fri, 18 Apr 2008 23:10:24 -0400 Joseph Fannin <[email protected]> wrote:
On Fri, Apr 18, 2008 at 01:47:57AM -0700, Andrew Morton wrote:
[...]
Quoted 18 lines
Quoted 17 lines
Quoted 4 lines
which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed.
Hm. We could either protect that code with an:
if (!is_geode())
return;
Or I could add the OpenFirmware patches which would allow us to get rid of this code, and instead check for the existence of OFW using that.
The former is quick and easy; the latter is (imo) nicer, so long as people don't have problems w/ the OFW code. :)
Quoted 6 lines
Quoted 3 lines
Do both ;)
The quick-n-easy version sounds suitable for now.
Heh, I already had sent the nicer version. If people have some fundamental problem w/ it, I can send the quick-n-easy version.
I prefer the nicer version. It is not a good policy IMHO to wrap OLPC specfic code with is_geode() and friends. Even by Geode standards, we've abused the code greatly for the benefit of the Geode, and few of those abuses would translate very well even to the general Geode community. I would prefer that we use the is_olpc() and #ifdef wrappers to ensure that the code that is exclusively OLPC stays exclusively OLPC.

Quoted 3 lines

Thanks,
Jordan

Yeah, like I said; the nicer version is the _correct_ way to do things. I just fear that the OFW code isn't ready for merging (see hpa's concerns).

The code is already #ifdef'd (the original reporter had enabled CONFIG_OLPC), and the code in question is what determines what is_olpc() should return. is_geode() is just to narrow the scope of what hardware the check runs on.

Signature · 1 line

Need a kernel or Debian developer?  Contact me, I'm looking for contracts.

Jordan Crouse

2008-04-21 15:06 UTC #99 lore raw

Quoted 51 lines · Andres Salomon

Quoted 44 lines · On Mon, 21 Apr 2008 08:56:19 -0600 Jordan Crouse <[email protected]> wrote:
Quoted 29 lines · Andres Salomon
Quoted 7 lines · On Sat, 19 Apr 2008 10:38:33 -0700 Andrew Morton <[email protected]>…
Quoted 7 lines · On Sat, 19 Apr 2008 09:25:44 -0400 Andres Salomon <[email protected]> wrote:
Quoted 3 lines · On Fri, 18 Apr 2008 20:29:25 -0700 Andrew Morton <[email protected]>…
Quoted 1 line · On Fri, 18 Apr 2008 23:10:24 -0400 Joseph Fannin <[email protected]> wrote:
On Fri, Apr 18, 2008 at 01:47:57AM -0700, Andrew Morton wrote:
[...]
Quoted 18 lines
Quoted 17 lines
Quoted 4 lines
which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed.
Hm. We could either protect that code with an:
if (!is_geode())
return;
Or I could add the OpenFirmware patches which would allow us to get rid of this code, and instead check for the existence of OFW using that.
The former is quick and easy; the latter is (imo) nicer, so long as people don't have problems w/ the OFW code. :)
Quoted 6 lines
Quoted 3 lines
Do both ;)
The quick-n-easy version sounds suitable for now.
Heh, I already had sent the nicer version. If people have some fundamental problem w/ it, I can send the quick-n-easy version.
I prefer the nicer version. It is not a good policy IMHO to wrap OLPC specfic code with is_geode() and friends. Even by Geode standards, we've abused the code greatly for the benefit of the Geode, and few of those abuses would translate very well even to the general Geode community. I would prefer that we use the is_olpc() and #ifdef wrappers to ensure that the code that is exclusively OLPC stays exclusively OLPC.
Quoted 3 lines
Thanks,
Jordan

Quoted 7 lines

Yeah, like I said; the nicer version is the _correct_ way to do things. I just fear that the OFW code isn't ready for merging (see hpa's concerns).
The code is already #ifdef'd (the original reporter had enabled CONFIG_OLPC), and the code in question is what determines what is_olpc() should return. is_geode() is just to narrow the scope of what hardware the check runs on.

My bad, I missed the key points. This still is dangerous on a generic Geode, but at least if they encounter the problem, we can loudly proclaim "Don't do that".

Jordan

Signature · 3 lines

Jordan Crouse
Systems Software Development Engineer 
Advanced Micro Devices, Inc.

Arjan van de Ven

2008-04-19 18:21 UTC #100 lore raw

Quoted 6 lines · On Fri, 18 Apr 2008 20:29:25 -0700 Andrew Morton <[email protected]>…

That's
                WARN_ON_ONCE(is_ram);
the changelog for the patch which added that warning is information-free

it got added with a full changelog, then temporary removed and then added back ;(

Quoted 11 lines

and there's no code comment explaining what went wrong, which makes things rather harder than they ought to be.
Yes it's due to the new OLPC code. olpc_init() has
	romsig = ioremap(0xffffffc0, 16);
which we probably just shouldn't do this at all unless we're running on the OLPC hardware. But we need to do this to find out if we're running on the OLPC hardware! Perhaps the warning should just be removed. --

calling ioremap() on something which COULD be ram is... REALLY nasty. The kernel has to mark that page uncached, for all users and mappings of that memory. A second hard case then is to find out when the last ioremap() user has released that memory (since there's several cases where different parts of the same 4K page can be ioremapped) before it can map it cached again. The good news is that until this olpc patch got in, there were no users of this capability.... Instead of outright forbidding it though we added a warn_on to find out if the assumption of no users was correct...
seems it caught some new code which is trying to do this here.

this code should probably be a lot more careful and check that
1) there is no actual kernel memory or something else at this region (what if there's some other device there? this code could blow up)
2) the machine won't tripple fault or otherwise throw tantrums if this hardcoded value is accessed (not automatic on x86!!)
3) it only runs if there's a really high degree of confidence that this really is an OLPC device.
or maybe
4) get this address from some other table or system provided resource

Signature · 3 lines

If you want to reach me at my work email, use [email protected]
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org

Jiri Slaby

2008-04-20 11:29 UTC #101 lore raw

internal compiler error: SIGSEGV [Was: 2.6.25-mm1]

Quoted 1 line · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.25/2.6.25-mm1/

Hi, I'm not sure by what was this caused.

LANG=en strace -fo strace_gcc.txt gcc -Wp,-MD,drivers/usb/class/.usblp.o.d -nostdinc -isystem /usr/lib64/gcc/x86_64-suse-linux/4.3/include -D__KERNEL__ -Iinclude -Iinclude2 -I/home/l/latest/xxx/include -include include/linux/autoconf.h -I/home/l/latest/xxx/drivers/usb/class -Idrivers/usb/class -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing -fno-common -Werror-implicit-function-declaration -O2 -fno-stack-protector -m64 -march=core2 -mno-red-zone -mcmodel=kernel -funit-at-a-time -maccumulate-outgoing-args -DCONFIG_AS_CFI=1 -DCONFIG_AS_CFI_SIGNAL_FRAME=1 -pipe -Wno-sign-compare -fno-asynchronous-unwind-tables -mno-sse -mno-mmx -mno-sse2 -mno-3dnow -I/home/l/latest/xxx/include/asm-x86/mach-default -Iinclude/asm-x86/mach-default -fno-omit-frame-pointer -fno-optimize-sibling-calls -g -Wdeclaration-after-statement -Wno-pointer-sign -DMODULE -D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_STR(usblp)" -D"KBUILD_MODNAME=KBUILD_STR(usblp)" /home/l/latest/xxx/drivers/usb/class/usblp.c -S -o usblp.s /home/l/latest/xxx/drivers/usb/class/usblp.c: In function 'usblp_submit_read':
/home/l/latest/xxx/drivers/usb/class/usblp.c:977: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://bugs.opensuse.org/> for instructions.

strace_gcc.txt:
http://www.fi.muni.cz/~xslaby/sklad/strace_gcc.txt

preprocessor output available here:
http://www.fi.muni.cz/~xslaby/sklad/usblp.E

Reboot fixed it. It happened after few suspend/resume cycles. The preproc output differs in no way from after the reboot. Now, the strace looks like:

5341  mmap(NULL, 32768, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f362e004000
5341  mmap(NULL, 1048576, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f362df04000
5341  brk(0x1964000)                    = 0x1964000
5341  brk(0x194c000)                    = 0x194c000
5341  brk(0x196d000)                    = 0x196d000
5341  brk(0x195a000)                    = 0x195a000
5341  mmap(NULL, 143360, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f362dee1000
5341  munmap(0x7f362dee1000, 143360)    = 0
5341  brk(0x1981000)                    = 0x1981000
5341  brk(0x196b000)                    = 0x196b000
5341  brk(0x1966000)                    = 0x1966000
5341  mmap(NULL, 32768, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f362defc000
5341  brk(0x1988000)                    = 0x1988000

at that sigsegv place.

Some kind of random-brk gcc (gcc-4.3-30) non-readiness?

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

Jiri Slaby

2008-04-21 08:31 UTC #102 lore raw

[Was: 2.6.25-mm1]