From: Hugh Dickins <hugh@veritas.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>,
Dave Jones <davej@redhat.com>,
Arjan van de Ven <arjan@infradead.org>,
linux-kernel@vger.kernel.org, linux-mm@kvack.org
Subject: [PATCH 2/8] badpage: keep any bad page out of circulation
Date: Mon, 1 Dec 2008 00:41:39 +0000 (GMT) [thread overview]
Message-ID: <Pine.LNX.4.64.0812010040330.11401@blonde.site> (raw)
In-Reply-To: <Pine.LNX.4.64.0812010032210.10131@blonde.site>
Until now the bad_page() checkers have special-cased PageReserved, keeping
those pages out of circulation thereafter. Now extend the special case to
all: we want to keep ANY page with bad state out of circulation - the
"free" page may well be in use by something.
Leave the bad state of those pages untouched, for examination by debuggers;
except for PageBuddy - leaving that set would risk bringing the page back.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
---
mm/page_alloc.c | 52 +++++++++++++++++++++-------------------------
1 file changed, 24 insertions(+), 28 deletions(-)
--- badpage1/mm/page_alloc.c 2008-11-28 20:40:33.000000000 +0000
+++ badpage2/mm/page_alloc.c 2008-11-28 20:40:36.000000000 +0000
@@ -231,9 +231,9 @@ static void bad_page(struct page *page)
printk(KERN_EMERG "Trying to fix it up, but a reboot is needed\n"
KERN_EMERG "Backtrace:\n");
dump_stack();
- set_page_count(page, 0);
- reset_page_mapcount(page);
- page->mapping = NULL;
+
+ /* Leave bad fields for debug, except PageBuddy could make trouble */
+ __ClearPageBuddy(page);
add_taint(TAINT_BAD_PAGE);
}
@@ -290,25 +290,31 @@ void prep_compound_gigantic_page(struct
}
#endif
-static void destroy_compound_page(struct page *page, unsigned long order)
+static int destroy_compound_page(struct page *page, unsigned long order)
{
int i;
int nr_pages = 1 << order;
+ int bad = 0;
- if (unlikely(compound_order(page) != order))
+ if (unlikely(compound_order(page) != order) ||
+ unlikely(!PageHead(page))) {
bad_page(page);
+ bad++;
+ }
- if (unlikely(!PageHead(page)))
- bad_page(page);
__ClearPageHead(page);
+
for (i = 1; i < nr_pages; i++) {
struct page *p = page + i;
- if (unlikely(!PageTail(p) |
- (p->first_page != page)))
+ if (unlikely(!PageTail(p) | (p->first_page != page))) {
bad_page(page);
+ bad++;
+ }
__ClearPageTail(p);
}
+
+ return bad;
}
static inline void prep_zero_page(struct page *page, int order, gfp_t gfp_flags)
@@ -428,7 +434,8 @@ static inline void __free_one_page(struc
int migratetype = get_pageblock_migratetype(page);
if (unlikely(PageCompound(page)))
- destroy_compound_page(page, order);
+ if (unlikely(destroy_compound_page(page, order)))
+ return;
page_idx = page_to_pfn(page) & ((1 << MAX_ORDER) - 1);
@@ -465,15 +472,10 @@ static inline int free_pages_check(struc
if (unlikely(page_mapcount(page) |
(page->mapping != NULL) |
(page_count(page) != 0) |
- (page->flags & PAGE_FLAGS_CHECK_AT_FREE)))
+ (page->flags & PAGE_FLAGS_CHECK_AT_FREE))) {
bad_page(page);
- /*
- * For now, we report if PG_reserved was found set, but do not
- * clear it, and do not free the page. But we shall soon need
- * to do more, for when the ZERO_PAGE count wraps negative.
- */
- if (PageReserved(page))
return 1;
+ }
if (page->flags & PAGE_FLAGS_CHECK_AT_PREP)
page->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
return 0;
@@ -521,11 +523,11 @@ static void __free_pages_ok(struct page
{
unsigned long flags;
int i;
- int reserved = 0;
+ int bad = 0;
for (i = 0 ; i < (1 << order) ; ++i)
- reserved += free_pages_check(page + i);
- if (reserved)
+ bad += free_pages_check(page + i);
+ if (bad)
return;
if (!PageHighMem(page)) {
@@ -610,17 +612,11 @@ static int prep_new_page(struct page *pa
if (unlikely(page_mapcount(page) |
(page->mapping != NULL) |
(page_count(page) != 0) |
- (page->flags & PAGE_FLAGS_CHECK_AT_PREP)))
+ (page->flags & PAGE_FLAGS_CHECK_AT_PREP))) {
bad_page(page);
-
- /*
- * For now, we report if PG_reserved was found set, but do not
- * clear it, and do not allocate the page: as a safety net.
- */
- if (PageReserved(page))
return 1;
+ }
- page->flags &= ~PAGE_FLAGS_CHECK_AT_PREP;
set_page_private(page, 0);
set_page_refcounted(page);
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2008-12-01 0:41 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-12-01 0:37 [PATCH 0/8] badpage: more resilient bad page pte and rmap Hugh Dickins
2008-12-01 0:40 ` [PATCH 1/8] badpage: simplify page_alloc flag check+clear Hugh Dickins
2008-12-01 14:47 ` Christoph Lameter
2008-12-01 23:50 ` Hugh Dickins
2008-12-02 2:21 ` Christoph Lameter
2008-12-02 10:39 ` Hugh Dickins
2008-12-02 13:12 ` Christoph Lameter
2008-12-02 14:12 ` Hugh Dickins
2008-12-03 0:57 ` Andrew Morton
2008-12-01 0:41 ` Hugh Dickins [this message]
2008-12-01 14:49 ` [PATCH 2/8] badpage: keep any bad page out of circulation Christoph Lameter
2008-12-01 23:19 ` Hugh Dickins
2008-12-01 0:42 ` [PATCH 3/8] badpage: replace page_remove_rmap Eeek and BUG Hugh Dickins
2008-12-01 0:43 ` [PATCH 4/8] badpage: vm_normal_page use print_bad_pte Hugh Dickins
2008-12-01 0:44 ` [PATCH 5/8] badpage: zap print_bad_pte on swap and file Hugh Dickins
2008-12-01 0:45 ` [PATCH 6/8] badpage: remove vma from page_remove_rmap Hugh Dickins
2008-12-01 0:46 ` [PATCH 7/8] badpage: ratelimit print_bad_pte and bad_page Hugh Dickins
2008-12-03 0:56 ` Andrew Morton
2008-12-03 13:04 ` Hugh Dickins
2008-12-01 0:48 ` [PATCH 8/8] badpage: KERN_ALERT BUG instead of KERN_EMERG Hugh Dickins
2008-12-01 14:40 ` Christoph Lameter
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Pine.LNX.4.64.0812010040330.11401@blonde.site \
--to=hugh@veritas.com \
--cc=akpm@linux-foundation.org \
--cc=arjan@infradead.org \
--cc=davej@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=nickpiggin@yahoo.com.au \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox