Patchwork [RFC,powerpc] try secondary hash before BUG in kernel_map_linear_page()

login
register
mail settings
Submitter Li Zhong
Date Feb. 25, 2013, 9:29 a.m.
Message ID <1361784575.3001.5.camel@ThinkPad-T5421.cn.ibm.com>
Download mbox | patch
Permalink /patch/222876/
State Changes Requested
Delegated to: Michael Ellerman
Headers show

Comments

Li Zhong - Feb. 25, 2013, 9:29 a.m.
This patch tries to fix following issue when CONFIG_DEBUG_PAGEALLOC
is enabled:

[  543.075675] ------------[ cut here ]------------
[  543.075701] kernel BUG at arch/powerpc/mm/hash_utils_64.c:1239!
[  543.075714] Oops: Exception in kernel mode, sig: 5 [#1]
[  543.075722] PREEMPT SMP NR_CPUS=16 DEBUG_PAGEALLOC NUMA pSeries
[  543.075741] Modules linked in: binfmt_misc ehea
[  543.075759] NIP: c000000000036eb0 LR: c000000000036ea4 CTR: c00000000005a594
[  543.075771] REGS: c0000000a90832c0 TRAP: 0700   Not tainted  (3.8.0-next-20130222)
[  543.075781] MSR: 8000000000029032 <SF,EE,ME,IR,DR,RI>  CR: 22224482  XER: 00000000
[  543.075816] SOFTE: 0
[  543.075823] CFAR: c00000000004c200
[  543.075830] TASK = c0000000e506b750[23934] 'cc1' THREAD: c0000000a9080000 CPU: 1
GPR00: 0000000000000001 c0000000a9083540 c000000000c600a8 ffffffffffffffff
GPR04: 0000000000000050 fffffffffffffffa c0000000a90834e0 00000000004ff594
GPR08: 0000000000000001 0000000000000000 000000009592d4d8 c000000000c86854
GPR12: 0000000000000002 c000000006ead300 0000000000a51000 0000000000000001
GPR16: f000000003354380 ffffffffffffffff ffffffffffffff80 0000000000000000
GPR20: 0000000000000001 c000000000c600a8 0000000000000001 0000000000000001
GPR24: 0000000003354380 c000000000000000 0000000000000000 c000000000b65950
GPR28: 0000002000000000 00000000000cd50e 0000000000bf50d9 c000000000c7c230
[  543.076005] NIP [c000000000036eb0] .kernel_map_pages+0x1e0/0x3f8
[  543.076016] LR [c000000000036ea4] .kernel_map_pages+0x1d4/0x3f8
[  543.076025] Call Trace:
[  543.076033] [c0000000a9083540] [c000000000036ea4] .kernel_map_pages+0x1d4/0x3f8 (unreliable)
[  543.076053] [c0000000a9083640] [c000000000167638] .get_page_from_freelist+0x6cc/0x8dc
[  543.076067] [c0000000a9083800] [c000000000167a48] .__alloc_pages_nodemask+0x200/0x96c
[  543.076082] [c0000000a90839c0] [c0000000001ade44] .alloc_pages_vma+0x160/0x1e4
[  543.076098] [c0000000a9083a80] [c00000000018ce04] .handle_pte_fault+0x1b0/0x7e8
[  543.076113] [c0000000a9083b50] [c00000000018d5a8] .handle_mm_fault+0x16c/0x1a0
[  543.076129] [c0000000a9083c00] [c0000000007bf1dc] .do_page_fault+0x4d0/0x7a4
[  543.076144] [c0000000a9083e30] [c0000000000090e8] handle_page_fault+0x10/0x30
[  543.076155] Instruction dump:
[  543.076163] 7c630038 78631d88 e80a0000 f8410028 7c0903a6 e91f01de e96a0010 e84a0008
[  543.076192] 4e800421 e8410028 7c7107b4 7a200fe0 <0b000000> 7f63db78 48785781 60000000
[  543.076224] ---[ end trace bd5807e8d6ae186b ]---

The code is borrowed from that in __hash_page_huge().

Signed-off-by: Li Zhong <zhong@linux.vnet.ibm.com>
---
 arch/powerpc/mm/hash_utils_64.c |   17 +++++++++++++++++
 1 file changed, 17 insertions(+)
Michael Ellerman - April 10, 2013, 2:21 a.m.
On Mon, Feb 25, 2013 at 05:29:35PM +0800, Li Zhong wrote:
> This patch tries to fix following issue when CONFIG_DEBUG_PAGEALLOC
> is enabled:
> 
> [  543.075675] ------------[ cut here ]------------
> [  543.075701] kernel BUG at arch/powerpc/mm/hash_utils_64.c:1239!
> [  543.075714] Oops: Exception in kernel mode, sig: 5 [#1]

So the issue is that kernel_map_linear_page() doesn't try the secondary
hash slot.

> The code is borrowed from that in __hash_page_huge().

It is, and in fact there is another copy in hash_low_64.S - in assembler.

So I think we should at least try and keep ourselves to two
implementations, one in asm and one in C. So can you split it out into a
helper routine called by both kernel_map_linear_page() and
__hash_page_huge() ?

cheers
Li Zhong - April 10, 2013, 5:29 a.m.
On Wed, 2013-04-10 at 12:21 +1000, Michael Ellerman wrote:
> On Mon, Feb 25, 2013 at 05:29:35PM +0800, Li Zhong wrote:
> > This patch tries to fix following issue when CONFIG_DEBUG_PAGEALLOC
> > is enabled:
> > 
> > [  543.075675] ------------[ cut here ]------------
> > [  543.075701] kernel BUG at arch/powerpc/mm/hash_utils_64.c:1239!
> > [  543.075714] Oops: Exception in kernel mode, sig: 5 [#1]

Hi Michael, 

Thanks for the review.

> So the issue is that kernel_map_linear_page() doesn't try the secondary
> hash slot.

It seems so.

> > The code is borrowed from that in __hash_page_huge().
> 
> It is, and in fact there is another copy in hash_low_64.S - in assembler.
> 
> So I think we should at least try and keep ourselves to two
> implementations, one in asm and one in C. So can you split it out into a
> helper routine called by both kernel_map_linear_page() and
> __hash_page_huge() ?

OK, I'll try to update it as you suggested. 

Thanks, Zhong

> cheers
>

Patch

diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c
index 1b6e127..31c7924 100644
--- a/arch/powerpc/mm/hash_utils_64.c
+++ b/arch/powerpc/mm/hash_utils_64.c
@@ -1231,11 +1231,28 @@  static void kernel_map_linear_page(unsigned long vaddr, unsigned long lmi)
 	int ret;
 
 	hash = hpt_hash(vpn, PAGE_SHIFT, mmu_kernel_ssize);
+
+repeat:
 	hpteg = ((hash & htab_hash_mask) * HPTES_PER_GROUP);
 
 	ret = ppc_md.hpte_insert(hpteg, vpn, __pa(vaddr),
 				 mode, HPTE_V_BOLTED,
 				 mmu_linear_psize, mmu_kernel_ssize);
+
+	if (unlikely(ret == -1)) {
+		hpteg = (~hash & htab_hash_mask) * HPTES_PER_GROUP;
+		ret = ppc_md.hpte_insert(hpteg, vpn, __pa(vaddr), mode,
+					 HPTE_V_SECONDARY,
+					 mmu_linear_psize, mmu_kernel_ssize);
+		if (ret == -1) {
+			if (mftb() & 0x1)
+				hpteg = (hash & htab_hash_mask) *
+					 HPTES_PER_GROUP;
+			ppc_md.hpte_remove(hpteg);
+			goto repeat;
+		}
+	}
+
 	BUG_ON (ret < 0);
 	spin_lock(&linear_map_hash_lock);
 	BUG_ON(linear_map_hash_slots[lmi] & 0x80);