[tip: perf/urgent] perf/core: Avoid put_page() when GUP fails

[tip-bot2 for Greg Thelen]

The following commit has been merged into the perf/urgent branch of tip:

Commit-ID:     4716023a8f6a0f4a28047f14dd7ebdc319606b84
Gitweb:        https://git.kernel.org/tip/4716023a8f6a0f4a28047f14dd7ebdc319606b84
Author:        Greg Thelen <gthelen@xxxxxxxxxx>
AuthorDate:    Wed, 10 Nov 2021 18:18:14 -08:00
Committer:     Peter Zijlstra <peterz@xxxxxxxxxxxxx>
CommitterDate: Thu, 11 Nov 2021 13:09:34 +01:00

perf/core: Avoid put_page() when GUP fails

PEBS PERF_SAMPLE_PHYS_ADDR events use perf_virt_to_phys() to convert PMU
sampled virtual addresses to physical using get_user_page_fast_only()
and page_to_phys().

Some get_user_page_fast_only() error cases return false, indicating no
page reference, but still initialize the output page pointer with an
unreferenced page. In these error cases perf_virt_to_phys() calls
put_page(). This causes page reference count underflow, which can lead
to unintentional page sharing.

Fix perf_virt_to_phys() to only put_page() if get_user_page_fast_only()
returns a referenced page.

Fixes: fc7ce9c74c3ad ("perf/core, x86: Add PERF_SAMPLE_PHYS_ADDR")
Signed-off-by: Greg Thelen <gthelen@xxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Link: https://lkml.kernel.org/r/20211111021814.757086-1-gthelen@xxxxxxxxxx
---
 kernel/events/core.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index f2253ea..523106a 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -7154,7 +7154,6 @@ void perf_output_sample(struct perf_output_handle *handle,
 static u64 perf_virt_to_phys(u64 virt)
 {
 	u64 phys_addr = 0;
-	struct page *p = NULL;
 
 	if (!virt)
 		return 0;
@@ -7173,14 +7172,15 @@ static u64 perf_virt_to_phys(u64 virt)
 		 * If failed, leave phys_addr as 0.
 		 */
 		if (current->mm != NULL) {
+			struct page *p;
+
 			pagefault_disable();
-			if (get_user_page_fast_only(virt, 0, &p))
+			if (get_user_page_fast_only(virt, 0, &p)) {
 				phys_addr = page_to_phys(p) + virt % PAGE_SIZE;
+				put_page(p);
+			}
 			pagefault_enable();
 		}
-
-		if (p)
-			put_page(p);
 	}
 
 	return phys_addr;