From: Kyle Huey me@kylehuey.com
The regressing commit is new in 6.10. It assumed that anytime event->prog is set bpf_overflow_handler() should be invoked to execute the attached bpf program. This assumption is false for tracing events, and as a result the regressing commit broke bpftrace by invoking the bpf handler with garbage inputs on overflow.
Prior to the regression the overflow handlers formed a chain (of length 0, 1, or 2) and perf_event_set_bpf_handler() (the !tracing case) added bpf_overflow_handler() to that chain, while perf_event_attach_bpf_prog() (the tracing case) did not. Both set event->prog. The chain of overflow handlers was replaced by a single overflow handler slot and a fixed call to bpf_overflow_handler() when appropriate. This modifies the condition there to check event->prog->type == BPF_PROG_TYPE_PERF_EVENT, restoring the previous behavior and fixing bpftrace.
Signed-off-by: Kyle Huey khuey@kylehuey.com Suggested-by: Andrii Nakryiko andrii.nakryiko@gmail.com Reported-by: Joe Damato jdamato@fastly.com Closes: https://lore.kernel.org/lkml/ZpFfocvyF3KHaSzF@LQ3V64L9R2/ Fixes: f11f10bfa1ca ("perf/bpf: Call BPF handler directly, not through overflow machinery") Cc: stable@vger.kernel.org Tested-by: Joe Damato jdamato@fastly.com # bpftrace --- v2: - Update patch based on Andrii's suggestion - Update commit message
kernel/events/core.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/kernel/events/core.c b/kernel/events/core.c index aa3450bdc227..c973e3c11e03 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -9706,7 +9706,8 @@ static int __perf_event_overflow(struct perf_event *event,
ret = __perf_event_account_interrupt(event, throttle);
- if (event->prog && !bpf_overflow_handler(event, data, regs)) + if (event->prog && event->prog->type == BPF_PROG_TYPE_PERF_EVENT && + !bpf_overflow_handler(event, data, regs)) return ret;
/*
On Tue, Aug 13, 2024 at 8:17 AM Joe Damato jdamato@fastly.com wrote:
From: Kyle Huey me@kylehuey.com
The regressing commit is new in 6.10. It assumed that anytime event->prog is set bpf_overflow_handler() should be invoked to execute the attached bpf program. This assumption is false for tracing events, and as a result the regressing commit broke bpftrace by invoking the bpf handler with garbage inputs on overflow.
Prior to the regression the overflow handlers formed a chain (of length 0, 1, or 2) and perf_event_set_bpf_handler() (the !tracing case) added bpf_overflow_handler() to that chain, while perf_event_attach_bpf_prog() (the tracing case) did not. Both set event->prog. The chain of overflow handlers was replaced by a single overflow handler slot and a fixed call to bpf_overflow_handler() when appropriate. This modifies the condition there to check event->prog->type == BPF_PROG_TYPE_PERF_EVENT, restoring the previous behavior and fixing bpftrace.
Signed-off-by: Kyle Huey khuey@kylehuey.com Suggested-by: Andrii Nakryiko andrii.nakryiko@gmail.com Reported-by: Joe Damato jdamato@fastly.com Closes: https://lore.kernel.org/lkml/ZpFfocvyF3KHaSzF@LQ3V64L9R2/ Fixes: f11f10bfa1ca ("perf/bpf: Call BPF handler directly, not through overflow machinery") Cc: stable@vger.kernel.org Tested-by: Joe Damato jdamato@fastly.com # bpftrace
v2:
- Update patch based on Andrii's suggestion
- Update commit message
LGTM, thanks.
Acked-by: Andrii Nakryiko andrii@kernel.org
kernel/events/core.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/kernel/events/core.c b/kernel/events/core.c index aa3450bdc227..c973e3c11e03 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -9706,7 +9706,8 @@ static int __perf_event_overflow(struct perf_event *event,
ret = __perf_event_account_interrupt(event, throttle);
if (event->prog && !bpf_overflow_handler(event, data, regs))
if (event->prog && event->prog->type == BPF_PROG_TYPE_PERF_EVENT &&!bpf_overflow_handler(event, data, regs)) return ret; /*-- 2.25.1
Hello:
This patch was applied to bpf/bpf.git (master) by Alexei Starovoitov ast@kernel.org:
On Tue, 13 Aug 2024 15:17:27 +0000 you wrote:
From: Kyle Huey me@kylehuey.com
The regressing commit is new in 6.10. It assumed that anytime event->prog is set bpf_overflow_handler() should be invoked to execute the attached bpf program. This assumption is false for tracing events, and as a result the regressing commit broke bpftrace by invoking the bpf handler with garbage inputs on overflow.
[...]
Here is the summary with links: - [v2] perf/bpf: Don't call bpf_overflow_handler() for tracing events https://git.kernel.org/bpf/bpf/c/100bff23818e
You are awesome, thank you!
linux-stable-mirror@lists.linaro.org
-
Andrii Nakryiko -
Joe Damato -
patchwork-bot+netdevbpf＠kernel.org