Skip to contentSkip to conversation

New: Kernel bug in mm/bootmem.c on Sparc machines

Bug 11046 Andrew Morton 3 participants 8 messages Jul 6, 2008 – Jul 24, 2008 Memory management

lore
Share & tools
lore

Andrew Morton

author maintainer
#0 lore raw

(switched to email. Please respond via emailed reply-to-all, not via the bugzilla web interface).

Quoted 81 lines · On Sun, 6 Jul 2008 13:02:28 -0700 (PDT) [email protected] wrote:

http://bugzilla.kernel.org/show_bug.cgi?id046

Summary: Kernel bug in mm/bootmem.c on Sparc machines
Product: Memory Management
Version: 2.5
KernelVersion: 2.6.25.10
Platform: All
OS/Version: Linux
Tree: Mainline
Status: NEW
Severity: blocking
Priority: P1
Component: Other
AssignedTo: [email protected]
ReportedBy: [email protected]

Latest working kernel version:
Earliest failing kernel version:
Distribution:
Hardware Environment: Sparc Blade B100s ( Ultra Sparc IIe 650 Mhz ) Software Environment:
Problem Description: Kernel Bug

Steps to reproduce: Create default kernel config for Sparc. Change this config that the kernel can do rarpd and change the network driver to be included in kernel (Cassini). Well set the kernel to be loaded over the net via tftpboot. I di this on my other Sparc machines and it worked there.

Here is the BUG:

[    0.000000] PROMLIB: Sun IEEE Boot Prom 'OBP 4.11.5 2003/11/12 10:40'
[    0.000000] PROMLIB: Root node compatible: 
[    0.000000] Linux version 2.6.25.10 (root@sparc1) (gcc version 4.1.2
20061115 (prerelease) (Debian 4.1.1-21)) #5 SMP Sun Jul 6 21:05:42 CEST 2008
[    0.000000] console [earlyprom0] enabled
[    0.000000] ARCH: SUN4U
[    0.000000] Ethernet address: 00:03:ba:7a:f3:d6
[    0.000000] Kernel: Using 2 locked TLB entries for main kernel image.
[    0.000000] Remapping the kernel... done.
[    0.000000] kernel BUG at mm/bootmem.c:125!
[    0.000000]               \|/ ____ \|/
[    0.000000]               "@'/ .. \`@"
[    0.000000]               /_| \__/ |_\
[    0.000000]                  \__U_/
[    0.000000] swapper(0): Kernel bad sw trap 5 [#1]
[    0.000000] TSTATE: 0000000080f01604 TPC: 00000000007ae2c4 TNPC:
00000000007ae2c8 Y: 00000000    Not tainted
[    0.000000] TPC: <reserve_bootmem_core+0x38/0xd8>
[    0.000000] g0: 0000000000000000 g1: 0000000000000001 g2: 000000000075ac00
g3: 000000000075afa8
[    0.000000] g4: 00000000007563c0 g5: 0000000000000000 g6: 00000000007523c0
g7: 0000000000000000
[    0.000000] o0: 0000000000000032 o1: 00000000007044e0 o2: 000000000000007d
o3: fffff80000438000
[    0.000000] o4: 0000000000000000 o5: 000000000075ef80 sp: 00000000007557c1
ret_pc: 00000000007ae2bc
[    0.000000] RPC: <reserve_bootmem_core+0x30/0xd8>
[    0.000000] l0: 000000000075ef50 l1: 0000000000000030 l2: 0000000000000000
l3: 0000000000000010
[    0.000000] l4: 0000000000000000 l5: 0000000000000010 l6: 0000000000000000
l7: 0000000000000000
[    0.000000] i0: 000000000081c5c0 i1: 0000000000438000 i2: 0000000000000000
i3: 0000000000000000
[    0.000000] i4: 0000000000000000 i5: 0000000000438000 i6: 0000000000755881
i7: 00000000007ab7c8
[    0.000000] I7: <paging_init+0xd50/0xee8>
[    0.000000] Caller[00000000007ab7c8]: paging_init+0xd50/0xee8
[    0.000000] Caller[00000000007a3da0]: setup_arch+0x2d8/0x2e0
[    0.000000] Caller[00000000007a0810]: start_kernel+0x7c/0x300
[    0.000000] Caller[00000000006860f0]: auxio_probe+0x0/0xd0
[    0.000000] Caller[0000000000000000]: 0x8
[    0.000000] Instruction DUMP: 11001c11  7ff1ee23  901220e0 <91d02005>
8406401a  b5307033  8200801a  8330700d  80a04003 
[    0.000000] Kernel panic - not syncing: Attempted to kill the idle task!
[    0.000000] Press Stop-A (L1-A) to return to the boot prom

Any help would be nice. If you need further information please tell me. This is my firs bug report and i hope i did it right.

#1 lore raw

From: Andrew Morton <[email protected]>
Date: Sun, 6 Jul 2008 13:20:49 -0700

Quoted 3 lines · On Sun, 6 Jul 2008 13:02:28 -0700 (PDT) [email protected] wrote:
Quoted 2 lines

http://bugzilla.kernel.org/show_bug.cgi?id046

...

Quoted 12 lines
Quoted 12 lines

Here is the BUG:

[    0.000000] PROMLIB: Sun IEEE Boot Prom 'OBP 4.11.5 2003/11/12 10:40'
[    0.000000] PROMLIB: Root node compatible: 
[    0.000000] Linux version 2.6.25.10 (root@sparc1) (gcc version 4.1.2
20061115 (prerelease) (Debian 4.1.1-21)) #5 SMP Sun Jul 6 21:05:42 CEST 2008
[    0.000000] console [earlyprom0] enabled
[    0.000000] ARCH: SUN4U
[    0.000000] Ethernet address: 00:03:ba:7a:f3:d6
[    0.000000] Kernel: Using 2 locked TLB entries for main kernel image.
[    0.000000] Remapping the kernel... done.
[    0.000000] kernel BUG at mm/bootmem.c:125!

This can only happen if you attach a zero-sized initrd to the kernel.

I see platforms like x86 sometimes have explicit checks for a zero size to guard reserve_bootmem() and similar calls, but if that's what callers are all going to do doesn't it make better sense for reserve_bootmem_core() to just return instead of BUG on a zero size argument?

author maintainer
#2 lore raw
Quoted 20 lines · On Wed, 23 Jul 2008 20:25:33 -0700 (PDT) David Miller <[email protected]>…

From: Andrew Morton <[email protected]>
Date: Sun, 6 Jul 2008 13:20:49 -0700

Quoted 3 lines · On Sun, 6 Jul 2008 13:02:28 -0700 (PDT) [email protected] wrote:
Quoted 2 lines

http://bugzilla.kernel.org/show_bug.cgi?id046

...

Quoted 12 lines
Quoted 12 lines

Here is the BUG:

[    0.000000] PROMLIB: Sun IEEE Boot Prom 'OBP 4.11.5 2003/11/12 10:40'
[    0.000000] PROMLIB: Root node compatible: 
[    0.000000] Linux version 2.6.25.10 (root@sparc1) (gcc version 4.1.2
20061115 (prerelease) (Debian 4.1.1-21)) #5 SMP Sun Jul 6 21:05:42 CEST 2008
[    0.000000] console [earlyprom0] enabled
[    0.000000] ARCH: SUN4U
[    0.000000] Ethernet address: 00:03:ba:7a:f3:d6
[    0.000000] Kernel: Using 2 locked TLB entries for main kernel image.
[    0.000000] Remapping the kernel... done.
[    0.000000] kernel BUG at mm/bootmem.c:125!
Quoted 7 lines

This can only happen if you attach a zero-sized initrd to the kernel.

I see platforms like x86 sometimes have explicit checks for a zero size to guard reserve_bootmem() and similar calls, but if that's what callers are all going to do doesn't it make better sense for reserve_bootmem_core() to just return instead of BUG on a zero size argument?

Sounds logical.

Johannes just rewrote the bootmem code, but from a quick read it appears that this behaviour has been retained.

So if we're going to change it in 2.6.26, we'll need a separate patch.

#3 lore raw

From: Andrew Morton <[email protected]>
Date: Wed, 23 Jul 2008 20:38:36 -0700

Quoted 1 line

So if we're going to change it in 2.6.26, we'll need a separate patch.

Here is the 2.6.26 version:

bootmem: Allow zero length reserve and free.

It's either this or all the call sites explicitly check when such a case is possible and sometimes expected.

mm/bootmem.c +4 -2
index 8d9f60e..e540f7a 100644
@@ -153,7 +153,8 @@ static void __init reserve_bootmem_core(bootmem_data_t *bdata,
 	unsigned long sidx, eidx;
 	unsigned long i;
 
-	BUG_ON(!size);
+	if (!size)
+		return;
 
 	/* out of range */
 	if (addr + size < bdata->node_boot_start ||
@@ -187,7 +188,8 @@ static void __init free_bootmem_core(bootmem_data_t *bdata, unsigned long addr,
 	unsigned long sidx, eidx;
 	unsigned long i;
 
-	BUG_ON(!size);
+	if (!size)
+		return;
 
 	/* out range */
 	if (addr + size < bdata->node_boot_start ||
#4 lore raw

Hi,

Quoted 29 lines · David Miller

From: Andrew Morton <[email protected]>
Date: Wed, 23 Jul 2008 20:38:36 -0700

Quoted 1 line

So if we're going to change it in 2.6.26, we'll need a separate patch.

Here is the 2.6.26 version:

bootmem: Allow zero length reserve and free.

It's either this or all the call sites explicitly check when such a case is possible and sometimes expected.

mm/bootmem.c +2 -1
index 8d9f60e..e540f7a 100644
@@ -153,7 +153,8 @@ static void __init reserve_bootmem_core(bootmem_data_t *bdata,
 	unsigned long sidx, eidx;
 	unsigned long i;
 
-	BUG_ON(!size);
+	if (!size)
+		return;
 
 	/* out of range */
 	if (addr + size < bdata->node_boot_start ||

mm/bootmem.c static void __init free_bootmem_core(bootmem_data_t *bdata, unsigned long addr, · line 195

Quoted 9 lines
mm/bootmem.c +2 -1
@@ -187,7 +188,8 @@ static void __init free_bootmem_core(bootmem_data_t *bdata, unsigned long addr,
 	unsigned long sidx, eidx;
 	unsigned long i;
 
-	BUG_ON(!size);
+	if (!size)
+		return;
 
 	/* out range */
 	if (addr + size < bdata->node_boot_start ||

Sorry, Dave, I missed that before: there is still the BUG_ON() in can_reserve_bootmem_core(), which should just return 0 instead.

Other than that, yes, Andrew, this introduces the same behaviour the bootmem rewrite.

Hannes

#5 lore raw

From: Johannes Weiner <[email protected]>
Date: Thu, 24 Jul 2008 23:32:06 +0200

Quoted 5 lines

Sorry, Dave, I missed that before: there is still the BUG_ON() in can_reserve_bootmem_core(), which should just return 0 instead.

Other than that, yes, Andrew, this introduces the same behaviour the bootmem rewrite.

Thanks, here is an updated version of the patch:

bootmem: Allow zero length reserve and free.

It's either this or all the call sites explicitly check when such a case is possible and sometimes expected.

mm/bootmem.c +6 -3
index 8d9f60e..5e3fab8 100644
@@ -117,7 +117,8 @@ static int __init can_reserve_bootmem_core(bootmem_data_t *bdata,
 	unsigned long sidx, eidx;
 	unsigned long i;
 
-	BUG_ON(!size);
+	if (!size)
+		return 0;
 
 	/* out of range, don't hold other */
 	if (addr + size < bdata->node_boot_start ||
@@ -153,7 +154,8 @@ static void __init reserve_bootmem_core(bootmem_data_t *bdata,
 	unsigned long sidx, eidx;
 	unsigned long i;
 
-	BUG_ON(!size);
+	if (!size)
+		return;
 
 	/* out of range */
 	if (addr + size < bdata->node_boot_start ||
@@ -187,7 +189,8 @@ static void __init free_bootmem_core(bootmem_data_t *bdata, unsigned long addr,
 	unsigned long sidx, eidx;
 	unsigned long i;
 
-	BUG_ON(!size);
+	if (!size)
+		return;
 
 	/* out range */
 	if (addr + size < bdata->node_boot_start ||
#6 lore raw
In reply to Andrew Morton

Hi,

Quoted 30 lines · Andrew Morton
Quoted 20 lines · On Wed, 23 Jul 2008 20:25:33 -0700 (PDT) David Miller <[email protected]>…

From: Andrew Morton <[email protected]>
Date: Sun, 6 Jul 2008 13:20:49 -0700

Quoted 3 lines · On Sun, 6 Jul 2008 13:02:28 -0700 (PDT) [email protected] wrote:
Quoted 2 lines

http://bugzilla.kernel.org/show_bug.cgi?id046

...

Quoted 12 lines
Quoted 12 lines

Here is the BUG:

[    0.000000] PROMLIB: Sun IEEE Boot Prom 'OBP 4.11.5 2003/11/12 10:40'
[    0.000000] PROMLIB: Root node compatible: 
[    0.000000] Linux version 2.6.25.10 (root@sparc1) (gcc version 4.1.2
20061115 (prerelease) (Debian 4.1.1-21)) #5 SMP Sun Jul 6 21:05:42 CEST 2008
[    0.000000] console [earlyprom0] enabled
[    0.000000] ARCH: SUN4U
[    0.000000] Ethernet address: 00:03:ba:7a:f3:d6
[    0.000000] Kernel: Using 2 locked TLB entries for main kernel image.
[    0.000000] Remapping the kernel... done.
[    0.000000] kernel BUG at mm/bootmem.c:125!
Quoted 7 lines

This can only happen if you attach a zero-sized initrd to the kernel.

I see platforms like x86 sometimes have explicit checks for a zero size to guard reserve_bootmem() and similar calls, but if that's what callers are all going to do doesn't it make better sense for reserve_bootmem_core() to just return instead of BUG on a zero size argument?

Quoted 4 lines

Sounds logical.

Johannes just rewrote the bootmem code, but from a quick read it appears that this behaviour has been retained.

In the new version, zero sized ranges are okay for reservation and freeing. It still bugs on allocation, though.

Quoted 1 line

So if we're going to change it in 2.6.26, we'll need a separate patch.

Hannes

author maintainer
#7 lore raw
Quoted 34 lines · On Thu, 24 Jul 2008 14:09:38 +0200 Johannes Weiner <[email protected]> wrote:

Hi,

Quoted 30 lines · Andrew Morton
Quoted 27 lines · On Wed, 23 Jul 2008 20:25:33 -0700 (PDT) David Miller <[email protected]>…

From: Andrew Morton <[email protected]>
Date: Sun, 6 Jul 2008 13:20:49 -0700

Quoted 3 lines · On Sun, 6 Jul 2008 13:02:28 -0700 (PDT) [email protected] wrote:
Quoted 2 lines

http://bugzilla.kernel.org/show_bug.cgi?id046

...

Quoted 12 lines
Quoted 12 lines

Here is the BUG:

[    0.000000] PROMLIB: Sun IEEE Boot Prom 'OBP 4.11.5 2003/11/12 10:40'
[    0.000000] PROMLIB: Root node compatible: 
[    0.000000] Linux version 2.6.25.10 (root@sparc1) (gcc version 4.1.2
20061115 (prerelease) (Debian 4.1.1-21)) #5 SMP Sun Jul 6 21:05:42 CEST 2008
[    0.000000] console [earlyprom0] enabled
[    0.000000] ARCH: SUN4U
[    0.000000] Ethernet address: 00:03:ba:7a:f3:d6
[    0.000000] Kernel: Using 2 locked TLB entries for main kernel image.
[    0.000000] Remapping the kernel... done.
[    0.000000] kernel BUG at mm/bootmem.c:125!

This can only happen if you attach a zero-sized initrd to the kernel.

I see platforms like x86 sometimes have explicit checks for a zero size to guard reserve_bootmem() and similar calls, but if that's what callers are all going to do doesn't it make better sense for reserve_bootmem_core() to just return instead of BUG on a zero size argument?

Quoted 8 lines
Quoted 4 lines

Sounds logical.

Johannes just rewrote the bootmem code, but from a quick read it appears that this behaviour has been retained.

In the new version, zero sized ranges are okay for reservation and freeing. It still bugs on allocation, though.

Interesting. So from Dave's patch (which changes only reserve_bootmem_core() and free_bootmem_core()), it sounds like we have already fixed 2.6.27?

In which case David's 2.6.26 patch is a "minimal backport".