All of lore.kernel.org
 help / color / mirror / Atom feed
From: akpm@linux-foundation.org
To: mm-commits@vger.kernel.org, will@kernel.org,
	gerald.schaefer@linux.ibm.com, chuhu@redhat.com,
	christophe.leroy@csgroup.eu, catalin.marinas@arm.com, cai@lca.pw,
	anshuman.khandual@arm.com, aneesh.kumar@linux.ibm.com,
	gshan@redhat.com
Subject: + mm-debug_vm_pgtable-fix-corrupted-page-flag.patch added to -mm tree
Date: Tue, 27 Jul 2021 13:53:03 -0700	[thread overview]
Message-ID: <20210727205303.fB02K%akpm@linux-foundation.org> (raw)


The patch titled
     Subject: mm/debug_vm_pgtable: fix corrupted page flag
has been added to the -mm tree.  Its filename is
     mm-debug_vm_pgtable-fix-corrupted-page-flag.patch

This patch should soon appear at
    https://ozlabs.org/~akpm/mmots/broken-out/mm-debug_vm_pgtable-fix-corrupted-page-flag.patch
and later at
    https://ozlabs.org/~akpm/mmotm/broken-out/mm-debug_vm_pgtable-fix-corrupted-page-flag.patch

Before you just go and hit "reply", please:
   a) Consider who else should be cc'ed
   b) Prefer to cc a suitable mailing list as well
   c) Ideally: find the original patch on the mailing list and do a
      reply-to-all to that, adding suitable additional cc's

*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***

The -mm tree is included into linux-next and is updated
there every 3-4 working days

------------------------------------------------------
From: Gavin Shan <gshan@redhat.com>
Subject: mm/debug_vm_pgtable: fix corrupted page flag

In page table entry modifying tests, set_xxx_at() are used to populate the
page table entries.  On ARM64, PG_arch_1 (PG_dcache_clean) flag is set to
the target page flag if execution permission is given.  The logic exits
since commit 4f04d8f00545 ("arm64: MMU definitions").  The page flag is
kept when the page is free'd to buddy's free area list.  However, it will
trigger page checking failure when it's pulled from the buddy's free area
list, as the following warning messages indicate.

   BUG: Bad page state in process memhog  pfn:08000
   page:0000000015c0a628 refcount:0 mapcount:0 \
        mapping:0000000000000000 index:0x1 pfn:0x8000
   flags: 0x7ffff8000000800(arch_1|node=0|zone=0|lastcpupid=0xfffff)
   raw: 07ffff8000000800 dead000000000100 dead000000000122 0000000000000000
   raw: 0000000000000001 0000000000000000 00000000ffffffff 0000000000000000
   page dumped because: PAGE_FLAGS_CHECK_AT_PREP flag(s) set

This fixes the issue by clearing PG_arch_1 through flush_dcache_page()
after set_xxx_at() is called.  For architectures other than ARM64, the
unexpected overhead of cache flushing is acceptable.

Link: https://lkml.kernel.org/r/20210727061401.592616-13-gshan@redhat.com
Signed-off-by: Gavin Shan <gshan@redhat.com>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Chunyu Hu <chuhu@redhat.com>
Cc: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Cc: Qian Cai <cai@lca.pw>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 mm/debug_vm_pgtable.c |   55 +++++++++++++++++++++++++++++++++++++---
 1 file changed, 51 insertions(+), 4 deletions(-)

--- a/mm/debug_vm_pgtable.c~mm-debug_vm_pgtable-fix-corrupted-page-flag
+++ a/mm/debug_vm_pgtable.c
@@ -29,6 +29,8 @@
 #include <linux/start_kernel.h>
 #include <linux/sched/mm.h>
 #include <linux/io.h>
+
+#include <asm/cacheflush.h>
 #include <asm/pgalloc.h>
 #include <asm/tlbflush.h>
 
@@ -119,19 +121,28 @@ static void __init pte_basic_tests(struc
 
 static void __init pte_advanced_tests(struct pgtable_debug_args *args)
 {
+	struct page *page;
 	pte_t pte;
 
 	/*
 	 * Architectures optimize set_pte_at by avoiding TLB flush.
 	 * This requires set_pte_at to be not used to update an
 	 * existing pte entry. Clear pte before we do set_pte_at
+	 *
+	 * flush_dcache_page() is called after set_pte_at() to clear
+	 * PG_arch_1 for the page on ARM64. The page flag isn't cleared
+	 * when it's released and page allocation check will fail when
+	 * the page is allocated again. For architectures other than ARM64,
+	 * the unexpected overhead of cache flushing is acceptable.
 	 */
-	if (args->pte_pfn == ULONG_MAX)
+	page = (args->pte_pfn != ULONG_MAX) ? pfn_to_page(args->pte_pfn) : NULL;
+	if (!page)
 		return;
 
 	pr_debug("Validating PTE advanced\n");
 	pte = pfn_pte(args->pte_pfn, args->page_prot);
 	set_pte_at(args->mm, args->vaddr, args->ptep, pte);
+	flush_dcache_page(page);
 	ptep_set_wrprotect(args->mm, args->vaddr, args->ptep);
 	pte = ptep_get(args->ptep);
 	WARN_ON(pte_write(pte));
@@ -143,6 +154,7 @@ static void __init pte_advanced_tests(st
 	pte = pte_wrprotect(pte);
 	pte = pte_mkclean(pte);
 	set_pte_at(args->mm, args->vaddr, args->ptep, pte);
+	flush_dcache_page(page);
 	pte = pte_mkwrite(pte);
 	pte = pte_mkdirty(pte);
 	ptep_set_access_flags(args->vma, args->vaddr, args->ptep, pte, 1);
@@ -155,6 +167,7 @@ static void __init pte_advanced_tests(st
 	pte = pfn_pte(args->pte_pfn, args->page_prot);
 	pte = pte_mkyoung(pte);
 	set_pte_at(args->mm, args->vaddr, args->ptep, pte);
+	flush_dcache_page(page);
 	ptep_test_and_clear_young(args->vma, args->vaddr, args->ptep);
 	pte = ptep_get(args->ptep);
 	WARN_ON(pte_young(pte));
@@ -213,15 +226,24 @@ static void __init pmd_basic_tests(struc
 
 static void __init pmd_advanced_tests(struct pgtable_debug_args *args)
 {
+	struct page *page;
 	pmd_t pmd;
 	unsigned long vaddr = args->vaddr;
 
 	if (!has_transparent_hugepage())
 		return;
 
-	if (args->pmd_pfn == ULONG_MAX)
+	page = (args->pmd_pfn != ULONG_MAX) ? pfn_to_page(args->pmd_pfn) : NULL;
+	if (!page)
 		return;
 
+	/*
+	 * flush_dcache_page() is called after set_pmd_at() to clear
+	 * PG_arch_1 for the page on ARM64. The page flag isn't cleared
+	 * when it's released and page allocation check will fail when
+	 * the page is allocated again. For architectures other than ARM64,
+	 * the unexpected overhead of cache flushing is acceptable.
+	 */
 	pr_debug("Validating PMD advanced\n");
 	/* Align the address wrt HPAGE_PMD_SIZE */
 	vaddr &= HPAGE_PMD_MASK;
@@ -230,6 +252,7 @@ static void __init pmd_advanced_tests(st
 
 	pmd = pfn_pmd(args->pmd_pfn, args->page_prot);
 	set_pmd_at(args->mm, vaddr, args->pmdp, pmd);
+	flush_dcache_page(page);
 	pmdp_set_wrprotect(args->mm, vaddr, args->pmdp);
 	pmd = READ_ONCE(*args->pmdp);
 	WARN_ON(pmd_write(pmd));
@@ -241,6 +264,7 @@ static void __init pmd_advanced_tests(st
 	pmd = pmd_wrprotect(pmd);
 	pmd = pmd_mkclean(pmd);
 	set_pmd_at(args->mm, vaddr, args->pmdp, pmd);
+	flush_dcache_page(page);
 	pmd = pmd_mkwrite(pmd);
 	pmd = pmd_mkdirty(pmd);
 	pmdp_set_access_flags(args->vma, vaddr, args->pmdp, pmd, 1);
@@ -253,6 +277,7 @@ static void __init pmd_advanced_tests(st
 	pmd = pmd_mkhuge(pfn_pmd(args->pmd_pfn, args->page_prot));
 	pmd = pmd_mkyoung(pmd);
 	set_pmd_at(args->mm, vaddr, args->pmdp, pmd);
+	flush_dcache_page(page);
 	pmdp_test_and_clear_young(args->vma, vaddr, args->pmdp);
 	pmd = READ_ONCE(*args->pmdp);
 	WARN_ON(pmd_young(pmd));
@@ -339,21 +364,31 @@ static void __init pud_basic_tests(struc
 
 static void __init pud_advanced_tests(struct pgtable_debug_args *args)
 {
+	struct page *page;
 	unsigned long vaddr = args->vaddr;
 	pud_t pud;
 
 	if (!has_transparent_hugepage())
 		return;
 
-	if (args->pud_pfn == ULONG_MAX)
+	page = (args->pud_pfn != ULONG_MAX) ? pfn_to_page(args->pud_pfn) : NULL;
+	if (!page)
 		return;
 
+	/*
+	 * flush_dcache_page() is called after set_pud_at() to clear
+	 * PG_arch_1 for the page on ARM64. The page flag isn't cleared
+	 * when it's released and page allocation check will fail when
+	 * the page is allocated again. For architectures other than ARM64,
+	 * the unexpected overhead of cache flushing is acceptable.
+	 */
 	pr_debug("Validating PUD advanced\n");
 	/* Align the address wrt HPAGE_PUD_SIZE */
 	vaddr &= HPAGE_PUD_MASK;
 
 	pud = pfn_pud(args->pud_pfn, args->page_prot);
 	set_pud_at(args->mm, vaddr, args->pudp, pud);
+	flush_dcache_page(page);
 	pudp_set_wrprotect(args->mm, vaddr, args->pudp);
 	pud = READ_ONCE(*args->pudp);
 	WARN_ON(pud_write(pud));
@@ -367,6 +402,7 @@ static void __init pud_advanced_tests(st
 	pud = pud_wrprotect(pud);
 	pud = pud_mkclean(pud);
 	set_pud_at(args->mm, vaddr, args->pudp, pud);
+	flush_dcache_page(page);
 	pud = pud_mkwrite(pud);
 	pud = pud_mkdirty(pud);
 	pudp_set_access_flags(args->vma, vaddr, args->pudp, pud, 1);
@@ -382,6 +418,7 @@ static void __init pud_advanced_tests(st
 	pud = pfn_pud(args->pud_pfn, args->page_prot);
 	pud = pud_mkyoung(pud);
 	set_pud_at(args->mm, vaddr, args->pudp, pud);
+	flush_dcache_page(page);
 	pudp_test_and_clear_young(args->vma, vaddr, args->pudp);
 	pud = READ_ONCE(*args->pudp);
 	WARN_ON(pud_young(pud));
@@ -594,16 +631,26 @@ static void __init pgd_populate_tests(st
 
 static void __init pte_clear_tests(struct pgtable_debug_args *args)
 {
+	struct page *page;
 	pte_t pte = pfn_pte(args->pte_pfn, args->page_prot);
 
-	if (args->pte_pfn == ULONG_MAX)
+	page = (args->pte_pfn != ULONG_MAX) ? pfn_to_page(args->pte_pfn) : NULL;
+	if (!page)
 		return;
 
+	/*
+	 * flush_dcache_page() is called after set_pte_at() to clear
+	 * PG_arch_1 for the page on ARM64. The page flag isn't cleared
+	 * when it's released and page allocation check will fail when
+	 * the page is allocated again. For architectures other than ARM64,
+	 * the unexpected overhead of cache flushing is acceptable.
+	 */
 	pr_debug("Validating PTE clear\n");
 #ifndef CONFIG_RISCV
 	pte = __pte(pte_val(pte) | RANDOM_ORVALUE);
 #endif
 	set_pte_at(args->mm, args->vaddr, args->ptep, pte);
+	flush_dcache_page(page);
 	barrier();
 	pte_clear(args->mm, args->vaddr, args->ptep);
 	pte = ptep_get(args->ptep);
_

Patches currently in -mm which might be from gshan@redhat.com are

mm-debug_vm_pgtable-introduce-struct-pgtable_debug_args.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-basic-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-leaf-and-savewrite-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-protnone-and-devmap-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-soft_dirty-and-swap-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-migration-and-thp-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-pte-modifying-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-pmd-modifying-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-pud-modifying-tests.patch
mm-debug_vm_pgtable-use-struct-pgtable_debug_args-in-pgd-and-p4d-modifying-tests.patch
mm-debug_vm_pgtable-remove-unused-code.patch
mm-debug_vm_pgtable-fix-corrupted-page-flag.patch


             reply	other threads:[~2021-07-27 20:53 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-27 20:53 akpm [this message]
  -- strict thread matches above, loose matches on Subject: below --
2021-08-09 18:07 + mm-debug_vm_pgtable-fix-corrupted-page-flag.patch added to -mm tree akpm
2021-08-02 20:59 akpm
2021-07-20 22:14 akpm
2021-07-10  3:41 akpm

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=20210727205303.fB02K%akpm@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=aneesh.kumar@linux.ibm.com \
    --cc=anshuman.khandual@arm.com \
    --cc=cai@lca.pw \
    --cc=catalin.marinas@arm.com \
    --cc=christophe.leroy@csgroup.eu \
    --cc=chuhu@redhat.com \
    --cc=gerald.schaefer@linux.ibm.com \
    --cc=gshan@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mm-commits@vger.kernel.org \
    --cc=will@kernel.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.