[tip:,perf/core] perf/core: Don't allow grouping events from different hw pmus

Message ID 166929138240.4906.17980748237008458473.tip-bot2@tip-bot2
State New
Headers
Series [tip:,perf/core] perf/core: Don't allow grouping events from different hw pmus |

Commit Message

tip-bot2 for Thomas Gleixner Nov. 24, 2022, 12:03 p.m. UTC
  The following commit has been merged into the perf/core branch of tip:

Commit-ID:     bf480f9385667309a8866e65833f261d4773df9a
Gitweb:        https://git.kernel.org/tip/bf480f9385667309a8866e65833f261d4773df9a
Author:        Ravi Bangoria <ravi.bangoria@amd.com>
AuthorDate:    Tue, 22 Nov 2022 13:33:26 +05:30
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Thu, 24 Nov 2022 11:09:19 +01:00

perf/core: Don't allow grouping events from different hw pmus

Event group from different hw pmus does not make sense and thus perf
has never allowed it. However, with recent rewrite that restriction
has been inadvertently removed. Fix it.

Fixes: bd2756811766 ("perf: Rewrite core context handling")
Signed-off-by: Ravi Bangoria <ravi.bangoria@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20221122080326.228-1-ravi.bangoria@amd.com
---
 kernel/events/core.c | 24 +++++++++++++++---------
 1 file changed, 15 insertions(+), 9 deletions(-)
  

Patch

diff --git a/kernel/events/core.c b/kernel/events/core.c
index f2b865c..65e20c5 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -12503,15 +12503,21 @@  SYSCALL_DEFINE5(perf_event_open,
 			 * perf_event_pmu_context.
 			 */
 			pmu = group_leader->pmu_ctx->pmu;
-		} else if (!is_software_event(event) &&
-			is_software_event(group_leader) &&
-			(group_leader->group_caps & PERF_EV_CAP_SOFTWARE)) {
-			/*
-			 * In case the group is a pure software group, and we
-			 * try to add a hardware event, move the whole group to
-			 * the hardware context.
-			 */
-			move_group = 1;
+		} else if (!is_software_event(event)) {
+			if (is_software_event(group_leader) &&
+			    (group_leader->group_caps & PERF_EV_CAP_SOFTWARE)) {
+				/*
+				 * In case the group is a pure software group, and we
+				 * try to add a hardware event, move the whole group to
+				 * the hardware context.
+				 */
+				move_group = 1;
+			}
+
+			/* Don't allow group of multiple hw events from different pmus */
+			if (!in_software_context(group_leader) &&
+			    group_leader->pmu_ctx->pmu != pmu)
+				goto err_locked;
 		}
 	}