Skip to contentSkip to conversation

FAILED: patch "[PATCH] mm: hwpoison: disable memory error handling on 1GB hugepage" failed to apply to 4.14-stable tree

PATCH [email protected] 2 participants 3 messages Apr 19, 2018 – Jul 10, 2018

Subsystems (4)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

Quoted 1 line

From 31286a8484a85e8b4e91ddb0f5415aee8a416827 Mon Sep 17 00:00:00 2001

From: Naoya Horiguchi <[email protected]>
Date: Thu, 5 Apr 2018 16:23:05 -0700
Subject: [PATCH] mm: hwpoison: disable memory error handling on 1GB hugepage

Recently the following BUG was reported:

    Injecting memory failure for pfn 0x3c0000 at process virtual address 0x7fe300000000
    Memory failure: 0x3c0000: recovery action for huge page: Recovered
    BUG: unable to handle kernel paging request at ffff8dfcc0003000
    IP: gup_pgd_range+0x1f0/0xc20
    PGD 17ae72067 P4D 17ae72067 PUD 0
    Oops: 0000 [#1] SMP PTI
    ...
    CPU: 3 PID: 5467 Comm: hugetlb_1gb Not tainted 4.15.0-rc8-mm1-abc+ #3
    Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.3-1.fc25 04/01/2014

You can easily reproduce this by calling madvise(MADV_HWPOISON) twice on a 1GB hugepage. This happens because get_user_pages_fast() is not aware of a migration entry on pud that was created in the 1st madvise() event.

I think that conversion to pud-aligned migration entry is working, but other MM code walking over page table isn't prepared for it. We need some time and effort to make all this work properly, so this patch avoids the reported bug by just disabling error handling for 1GB hugepage.

include/linux/mm.h +1 -0
index 2e40a44a1fae..2e2be527642a 100644
@@ -2613,6 +2613,7 @@ enum mf_action_page_type {
 	MF_MSG_POISONED_HUGE,
 	MF_MSG_HUGE,
 	MF_MSG_FREE_HUGE,
+	MF_MSG_NON_PMD_HUGE,
 	MF_MSG_UNMAP_FAILED,
 	MF_MSG_DIRTY_SWAPCACHE,
 	MF_MSG_CLEAN_SWAPCACHE,
mm/memory-failure.c +16 -0
index 8291b75f42c8..2d4bf647cf01 100644
@@ -502,6 +502,7 @@ static const char * const action_page_types[] = {
 	[MF_MSG_POISONED_HUGE]		= "huge page already hardware poisoned",
 	[MF_MSG_HUGE]			= "huge page",
 	[MF_MSG_FREE_HUGE]		= "free huge page",
+	[MF_MSG_NON_PMD_HUGE]		= "non-pmd-sized huge page",
 	[MF_MSG_UNMAP_FAILED]		= "unmapping failed page",
 	[MF_MSG_DIRTY_SWAPCACHE]	= "dirty swapcache page",
 	[MF_MSG_CLEAN_SWAPCACHE]	= "clean swapcache page",
@@ -1084,6 +1085,21 @@ static int memory_failure_hugetlb(unsigned long pfn, int flags)
 		return 0;
 	}
 
+	/*
+	 * TODO: hwpoison for pud-sized hugetlb doesn't work right now, so
+	 * simply disable it. In order to make it work properly, we need
+	 * make sure that:
+	 *  - conversion of a pud that maps an error hugetlb into hwpoison
+	 *    entry properly works, and
+	 *  - other mm code walking over page table is aware of pud-aligned
+	 *    hwpoison entries.
+	 */
+	if (huge_page_size(page_hstate(head)) > PMD_SIZE) {
+		action_result(pfn, MF_MSG_NON_PMD_HUGE, MF_IGNORED);
+		res = -EBUSY;
+		goto out;
+	}
+
 	if (!hwpoison_user_mappings(p, pfn, flags, &head)) {
 		action_result(pfn, MF_MSG_UNMAP_FAILED, MF_IGNORED);
 		res = -EBUSY;
Acked-by Michal Hocko Reviewed-by Andrew Morton Reviewed-by Mike Kravetz Acked-by Punit Agrawal Tested-by Michael Ellerman
#1 lore raw

Hi Greg,

Quoted 5 lines · On Thu, Apr 19, 2018 at 05:24:22PM +0200, [email protected] wrote:

The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

The attached backported patch should apply to v4.14-stable.

--
Regards
Sudip

Quoted 1 line

From 1339325714109713ead30ec83e4b5cad186626ba Mon Sep 17 00:00:00 2001

From: Naoya Horiguchi <[email protected]>
Date: Thu, 5 Apr 2018 16:23:05 -0700
Subject: [PATCH] mm: hwpoison: disable memory error handling on 1GB hugepage

commit 31286a8484a85e8b4e91ddb0f5415aee8a416827 upstream

Recently the following BUG was reported:

    Injecting memory failure for pfn 0x3c0000 at process virtual address 0x7fe300000000
    Memory failure: 0x3c0000: recovery action for huge page: Recovered
    BUG: unable to handle kernel paging request at ffff8dfcc0003000
    IP: gup_pgd_range+0x1f0/0xc20
    PGD 17ae72067 P4D 17ae72067 PUD 0
    Oops: 0000 [#1] SMP PTI
    ...
    CPU: 3 PID: 5467 Comm: hugetlb_1gb Not tainted 4.15.0-rc8-mm1-abc+ #3
    Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.3-1.fc25 04/01/2014

You can easily reproduce this by calling madvise(MADV_HWPOISON) twice on a 1GB hugepage. This happens because get_user_pages_fast() is not aware of a migration entry on pud that was created in the 1st madvise() event.

I think that conversion to pud-aligned migration entry is working, but other MM code walking over page table isn't prepared for it. We need some time and effort to make all this work properly, so this patch avoids the reported bug by just disabling error handling for 1GB hugepage.

---
 include/linux/mm.h  |  1 +
 mm/memory-failure.c | 16 ++++++++++++++++
 2 files changed, 17 insertions(+)
include/linux/mm.h +1 -0
index f23215854c80..a26cf767407e 100644
@@ -2549,6 +2549,7 @@ enum mf_action_page_type {
 	MF_MSG_POISONED_HUGE,
 	MF_MSG_HUGE,
 	MF_MSG_FREE_HUGE,
+	MF_MSG_NON_PMD_HUGE,
 	MF_MSG_UNMAP_FAILED,
 	MF_MSG_DIRTY_SWAPCACHE,
 	MF_MSG_CLEAN_SWAPCACHE,
mm/memory-failure.c +16 -0
index 1cd3b3569af8..345e69d88b37 100644
@@ -508,6 +508,7 @@ static const char * const action_page_types[] = {
 	[MF_MSG_POISONED_HUGE]		= "huge page already hardware poisoned",
 	[MF_MSG_HUGE]			= "huge page",
 	[MF_MSG_FREE_HUGE]		= "free huge page",
+	[MF_MSG_NON_PMD_HUGE]		= "non-pmd-sized huge page",
 	[MF_MSG_UNMAP_FAILED]		= "unmapping failed page",
 	[MF_MSG_DIRTY_SWAPCACHE]	= "dirty swapcache page",
 	[MF_MSG_CLEAN_SWAPCACHE]	= "clean swapcache page",
@@ -1090,6 +1091,21 @@ static int memory_failure_hugetlb(unsigned long pfn, int trapno, int flags)
 		return 0;
 	}
 
+	/*
+	 * TODO: hwpoison for pud-sized hugetlb doesn't work right now, so
+	 * simply disable it. In order to make it work properly, we need
+	 * make sure that:
+	 *  - conversion of a pud that maps an error hugetlb into hwpoison
+	 *    entry properly works, and
+	 *  - other mm code walking over page table is aware of pud-aligned
+	 *    hwpoison entries.
+	 */
+	if (huge_page_size(page_hstate(head)) > PMD_SIZE) {
+		action_result(pfn, MF_MSG_NON_PMD_HUGE, MF_IGNORED);
+		res = -EBUSY;
+		goto out;
+	}
+
 	if (!hwpoison_user_mappings(p, pfn, trapno, flags, &head)) {
 		action_result(pfn, MF_MSG_UNMAP_FAILED, MF_IGNORED);
 		res = -EBUSY;
Signature · 1 line
2.11.0
author maintainer
#2 lore raw
Quoted 10 lines · Sudip Mukherjee

Hi Greg,

Quoted 5 lines · On Thu, Apr 19, 2018 at 05:24:22PM +0200, [email protected] wrote:

The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

The attached backported patch should apply to v4.14-stable.

Now applied, thanks.

greg k-h