[lkml]   [2018]   [Jul]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
Patch in this message
Subject[tip:x86/pti] perf/core: Make sure the ring-buffer is mapped in all page-tables
Commit-ID:  cdbaf0a372db2bc3c3127e8b63fd15bd6e6757ee
Author: Joerg Roedel <>
AuthorDate: Fri, 20 Jul 2018 18:22:22 +0200
Committer: Thomas Gleixner <>
CommitDate: Fri, 20 Jul 2018 21:32:08 +0200

perf/core: Make sure the ring-buffer is mapped in all page-tables

The ring-buffer is accessed in the NMI handler, so it's better to avoid
faulting on it. Sync the vmalloc range with all page-tables in system to
make sure everyone has it mapped.

This fixes a WARN_ON_ONCE() that can be triggered with PTI enabled on

WARNING: CPU: 4 PID: 0 at arch/x86/mm/fault.c:320 vmalloc_fault+0x220/0x230

This triggers because with PTI enabled on an PAE kernel the PMDs are no
longer shared between the page-tables, so the vmalloc changes do not
propagate automatically.

Note: Andy said rightfully that we should try to fix the vmalloc code for
that case, but that's not a hot fix for the issue at hand.

Fixes: 7757d607c6b3 ("x86/pti: Allow CONFIG_PAGE_TABLE_ISOLATION for x86_32")
Signed-off-by: Joerg Roedel <>
Signed-off-by: Thomas Gleixner <>
Cc: "H . Peter Anvin" <>
Cc: Linus Torvalds <>
Cc: Andy Lutomirski <>
Cc: Dave Hansen <>
Cc: Josh Poimboeuf <>
Cc: Juergen Gross <>
Cc: Peter Zijlstra <>
Cc: Borislav Petkov <>
Cc: Jiri Kosina <>
Cc: Boris Ostrovsky <>
Cc: Brian Gerst <>
Cc: David Laight <>
Cc: Denys Vlasenko <>
Cc: Eduardo Valentin <>
Cc: Greg KH <>
Cc: Will Deacon <>
Cc: Andrea Arcangeli <>
Cc: Waiman Long <>
Cc: Pavel Machek <>
Cc: "David H . Gutteridge" <>
Cc: Arnaldo Carvalho de Melo <>
Cc: Alexander Shishkin <>
Cc: Jiri Olsa <>
Cc: Namhyung Kim <>

kernel/events/ring_buffer.c | 10 ++++++++++
1 file changed, 10 insertions(+)

diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c
index 5d3cf407e374..7b0e9aafafdf 100644
--- a/kernel/events/ring_buffer.c
+++ b/kernel/events/ring_buffer.c
@@ -814,6 +814,9 @@ static void rb_free_work(struct work_struct *work)

+ /* Make sure buffer is unmapped in all page-tables */
+ vmalloc_sync_all();

void rb_free(struct ring_buffer *rb)
@@ -840,6 +843,13 @@ struct ring_buffer *rb_alloc(int nr_pages, long watermark, int cpu, int flags)
if (!all_buf)
goto fail_all_buf;

+ /*
+ * The buffer is accessed in NMI handlers, make sure it is
+ * mapped in all page-tables in the system so that we don't
+ * fault on the range in an NMI handler.
+ */
+ vmalloc_sync_all();
rb->user_page = all_buf;
rb->data_pages[0] = all_buf + PAGE_SIZE;
if (nr_pages) {
 \ /
  Last update: 2018-07-20 21:39    [W:0.208 / U:3.644 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site