linux-kernel - Re: [PATCH V5] mm: fix use-after free of page

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20220826201242.bbf4948e5221c883dadfc2d6@linux-foundation.org>
Date:   Fri, 26 Aug 2022 20:12:42 -0700
From:   Andrew Morton <akpm@...ux-foundation.org>
To:     Michal Hocko <mhocko@...e.com>
Cc:     Charan Teja Kalla <quic_charante@...cinc.com>, david@...hat.com,
        vbabka@...e.cz, pasha.tatashin@...een.com, shakeelb@...gle.com,
        sieberf@...zon.com, sjpark@...zon.de, william.kucharski@...cle.com,
        willy@...radead.org, quic_pkondeti@...cinc.com, minchan@...gle.com,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [PATCH V5] mm: fix use-after free of page_ext after race with
 memory-offline

On Fri, 26 Aug 2022 09:40:52 +0200 Michal Hocko <mhocko@...e.com> wrote:

> > Changes in V5:
> >    o Used the loop variable name as ext_put_continue -- David
> 
> If this is the only change since the last version (I didn't have much
> time to read through the whole patch) then my ack still applies.

yup.

--- a/mm/page_owner.c~mm-fix-use-after-free-of-page_ext-after-race-with-memory-offline-v5
+++ a/mm/page_owner.c
@@ -326,7 +326,7 @@ void pagetypeinfo_showmixedcount_print(s
 				continue;
 
 			if (!test_bit(PAGE_EXT_OWNER_ALLOCATED, &page_ext->flags))
-				goto loop;
+				goto ext_put_continue;
 
 			page_owner = get_page_owner(page_ext);
 			page_mt = gfp_migratetype(page_owner->gfp_mask);
@@ -341,7 +341,7 @@ void pagetypeinfo_showmixedcount_print(s
 				break;
 			}
 			pfn += (1UL << page_owner->order) - 1;
-loop:
+ext_put_continue:
 			page_ext_put(page_ext);
 		}
 	}
@@ -561,14 +561,14 @@ read_page_owner(struct file *file, char
 		 * because we don't hold the zone lock.
 		 */
 		if (!test_bit(PAGE_EXT_OWNER, &page_ext->flags))
-			goto loop;
+			goto ext_put_continue;
 
 		/*
 		 * Although we do have the info about past allocation of free
 		 * pages, it's not relevant for current memory usage.
 		 */
 		if (!test_bit(PAGE_EXT_OWNER_ALLOCATED, &page_ext->flags))
-			goto loop;
+			goto ext_put_continue;
 
 		page_owner = get_page_owner(page_ext);
 
@@ -577,7 +577,7 @@ read_page_owner(struct file *file, char
 		 * would inflate the stats.
 		 */
 		if (!IS_ALIGNED(pfn, 1 << page_owner->order))
-			goto loop;
+			goto ext_put_continue;
 
 		/*
 		 * Access to page_ext->handle isn't synchronous so we should
@@ -585,7 +585,7 @@ read_page_owner(struct file *file, char
 		 */
 		handle = READ_ONCE(page_owner->handle);
 		if (!handle)
-			goto loop;
+			goto ext_put_continue;
 
 		/* Record the next PFN to read in the file offset */
 		*ppos = (pfn - min_low_pfn) + 1;
@@ -594,7 +594,7 @@ read_page_owner(struct file *file, char
 		page_ext_put(page_ext);
 		return print_page_owner(buf, count, pfn, page,
 				&page_owner_tmp, handle);
-loop:
+ext_put_continue:
 		page_ext_put(page_ext);
 	}
 
@@ -654,13 +654,13 @@ static void init_pages_in_zone(pg_data_t
 
 			/* Maybe overlapping zone */
 			if (test_bit(PAGE_EXT_OWNER, &page_ext->flags))
-				goto loop;
+				goto ext_put_continue;
 
 			/* Found early allocated page */
 			__set_page_owner_handle(page_ext, early_handle,
 						0, 0);
 			count++;
-loop:
+ext_put_continue:
 			page_ext_put(page_ext);
 		}
 		cond_resched();
_