From: Michal Hocko <mhocko@kernel.org>
To: Wei Yang <richardw.yang@linux.intel.com>
Cc: akpm@linux-foundation.org, linux-kernel@vger.kernel.org,
linux-mm@kvack.org, rientjes@google.com
Subject: Re: [Patch v2 3/4] mm/page_alloc.c: pass all bad reasons to bad_page()
Date: Mon, 20 Jan 2020 11:22:00 +0100 [thread overview]
Message-ID: <20200120102200.GW18451@dhcp22.suse.cz> (raw)
In-Reply-To: <20200120030415.15925-4-richardw.yang@linux.intel.com>
On Mon 20-01-20 11:04:14, Wei Yang wrote:
> Now we can pass all bad reasons to __dump_page().
And we do we want to do that? The dump of the page will tell us the
whole story so a single and the most important reason sounds like a
better implementation. The code is also more subtle because each caller
of the function has to be aware of how many reasons there might be.
Not to mention that you need a room for 5 pointers on the stack and this
and page allocator might be called from deeper call chains.
> Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
> ---
> mm/page_alloc.c | 52 ++++++++++++++++++++++++++-----------------------
> 1 file changed, 28 insertions(+), 24 deletions(-)
>
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index a43b9d2482f2..a7b793c739fc 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -609,7 +609,7 @@ static inline int __maybe_unused bad_range(struct zone *zone, struct page *page)
> }
> #endif
>
> -static void bad_page(struct page *page, const char *reason,
> +static void bad_page(struct page *page, int nr, const char **reason,
> unsigned long bad_flags)
> {
> static unsigned long resume;
> @@ -638,7 +638,7 @@ static void bad_page(struct page *page, const char *reason,
>
> pr_alert("BUG: Bad page state in process %s pfn:%05lx\n",
> current->comm, page_to_pfn(page));
> - __dump_page(page, 1, &reason);
> + __dump_page(page, nr, reason);
> bad_flags &= page->flags;
> if (bad_flags)
> pr_alert("bad because of flags: %#lx(%pGp)\n",
> @@ -1027,27 +1027,25 @@ static inline bool page_expected_state(struct page *page,
>
> static void free_pages_check_bad(struct page *page)
> {
> - const char *bad_reason;
> - unsigned long bad_flags;
> -
> - bad_reason = NULL;
> - bad_flags = 0;
> + const char *bad_reason[5];
> + unsigned long bad_flags = 0;
> + int nr = 0;
>
> if (unlikely(atomic_read(&page->_mapcount) != -1))
> - bad_reason = "nonzero mapcount";
> + bad_reason[nr++] = "nonzero mapcount";
> if (unlikely(page->mapping != NULL))
> - bad_reason = "non-NULL mapping";
> + bad_reason[nr++] = "non-NULL mapping";
> if (unlikely(page_ref_count(page) != 0))
> - bad_reason = "nonzero _refcount";
> + bad_reason[nr++] = "nonzero _refcount";
> if (unlikely(page->flags & PAGE_FLAGS_CHECK_AT_FREE)) {
> - bad_reason = "PAGE_FLAGS_CHECK_AT_FREE flag(s) set";
> + bad_reason[nr++] = "PAGE_FLAGS_CHECK_AT_FREE flag(s) set";
> bad_flags = PAGE_FLAGS_CHECK_AT_FREE;
> }
> #ifdef CONFIG_MEMCG
> if (unlikely(page->mem_cgroup))
> - bad_reason = "page still charged to cgroup";
> + bad_reason[nr++] = "page still charged to cgroup";
> #endif
> - bad_page(page, bad_reason, bad_flags);
> + bad_page(page, nr, bad_reason, bad_flags);
> }
>
> static inline int free_pages_check(struct page *page)
> @@ -1062,6 +1060,7 @@ static inline int free_pages_check(struct page *page)
>
> static int free_tail_pages_check(struct page *head_page, struct page *page)
> {
> + const char *reason;
> int ret = 1;
>
> /*
> @@ -1078,7 +1077,8 @@ static int free_tail_pages_check(struct page *head_page, struct page *page)
> case 1:
> /* the first tail page: ->mapping may be compound_mapcount() */
> if (unlikely(compound_mapcount(page))) {
> - bad_page(page, "nonzero compound_mapcount", 0);
> + reason = "nonzero compound_mapcount";
> + bad_page(page, 1, &reason, 0);
> goto out;
> }
> break;
> @@ -1090,17 +1090,20 @@ static int free_tail_pages_check(struct page *head_page, struct page *page)
> break;
> default:
> if (page->mapping != TAIL_MAPPING) {
> - bad_page(page, "corrupted mapping in tail page", 0);
> + reason = "corrupted mapping in tail page";
> + bad_page(page, 1, &reason, 0);
> goto out;
> }
> break;
> }
> if (unlikely(!PageTail(page))) {
> - bad_page(page, "PageTail not set", 0);
> + reason = "PageTail not set";
> + bad_page(page, 1, &reason, 0);
> goto out;
> }
> if (unlikely(compound_head(page) != head_page)) {
> - bad_page(page, "compound_head not consistent", 0);
> + reason = "compound_head not consistent";
> + bad_page(page, 1, &reason, 0);
> goto out;
> }
> ret = 0;
> @@ -2041,29 +2044,30 @@ static inline void expand(struct zone *zone, struct page *page,
>
> static void check_new_page_bad(struct page *page)
> {
> - const char *bad_reason = NULL;
> + const char *bad_reason[5];
> unsigned long bad_flags = 0;
> + int nr = 0;
>
> if (unlikely(atomic_read(&page->_mapcount) != -1))
> - bad_reason = "nonzero mapcount";
> + bad_reason[nr++] = "nonzero mapcount";
> if (unlikely(page->mapping != NULL))
> - bad_reason = "non-NULL mapping";
> + bad_reason[nr++] = "non-NULL mapping";
> if (unlikely(page_ref_count(page) != 0))
> - bad_reason = "nonzero _refcount";
> + bad_reason[nr++] = "nonzero _refcount";
> if (unlikely(page->flags & __PG_HWPOISON)) {
> /* Don't complain about hwpoisoned pages */
> page_mapcount_reset(page); /* remove PageBuddy */
> return;
> }
> if (unlikely(page->flags & PAGE_FLAGS_CHECK_AT_PREP)) {
> - bad_reason = "PAGE_FLAGS_CHECK_AT_PREP flag set";
> + bad_reason[nr++] = "PAGE_FLAGS_CHECK_AT_PREP flag set";
> bad_flags = PAGE_FLAGS_CHECK_AT_PREP;
> }
> #ifdef CONFIG_MEMCG
> if (unlikely(page->mem_cgroup))
> - bad_reason = "page still charged to cgroup";
> + bad_reason[nr++] = "page still charged to cgroup";
> #endif
> - bad_page(page, bad_reason, bad_flags);
> + bad_page(page, 1, bad_reason, bad_flags);
> }
>
> /*
> --
> 2.17.1
>
--
Michal Hocko
SUSE Labs
next prev parent reply other threads:[~2020-01-20 10:22 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-20 3:04 [Patch v2 0/4] mm/page_alloc.c: cleanup on check page Wei Yang
2020-01-20 3:04 ` [Patch v2 1/4] mm: enable dump several reasons for __dump_page() Wei Yang
2020-01-20 6:12 ` Anshuman Khandual
2020-01-20 8:55 ` Wei Yang
2020-01-21 5:20 ` Anshuman Khandual
2020-01-22 0:58 ` Wei Yang
2020-01-26 2:44 ` Wei Yang
2020-01-20 3:04 ` [Patch v2 2/4] mm/page_alloc.c: bad_[reason|flags] is not necessary when PageHWPoison Wei Yang
2020-01-20 6:28 ` Anshuman Khandual
2020-01-20 12:13 ` Wei Yang
2020-01-20 10:17 ` Michal Hocko
2020-01-20 12:20 ` David Hildenbrand
2020-01-20 3:04 ` [Patch v2 3/4] mm/page_alloc.c: pass all bad reasons to bad_page() Wei Yang
2020-01-20 6:33 ` Anshuman Khandual
2020-01-20 12:33 ` Wei Yang
2020-01-20 10:22 ` Michal Hocko [this message]
2020-01-20 12:19 ` David Hildenbrand
2020-01-21 1:49 ` Wei Yang
2020-01-21 6:08 ` Anshuman Khandual
2020-01-21 8:47 ` Michal Hocko
2020-01-20 3:04 ` [Patch v2 4/4] mm/page_alloc.c: extract commom part to check page Wei Yang
2020-01-20 6:43 ` Anshuman Khandual
2020-01-20 12:36 ` Wei Yang
2020-01-21 4:49 ` Anshuman Khandual
2020-01-22 1:00 ` Wei Yang
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=20200120102200.GW18451@dhcp22.suse.cz \
--to=mhocko@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=richardw.yang@linux.intel.com \
--cc=rientjes@google.com \
/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;
as well as URLs for NNTP newsgroup(s).