Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Sumanth Korikkar <sumanthk@linux.ibm.com>
To:     tip-bot2@linutronix.de
Cc:     linux-kernel@vger.kernel.org, linux-tip-commits@vger.kernel.org,
        namhyung@kernel.org, peterz@infradead.org, x86@kernel.org,
        iii@linux.ibm.com, gor@linux.ibm.com, hca@linux.ibm.com,
        svens@linux.ibm.com, tmricht@linux.ibm.com, bpf@vger.kernel.org,
        Sumanth Korikkar <sumanthk@linux.ibm.com>
Subject: [PATCH] Re: [tip: perf/core] perf: Use sample_flags for raw_data
Date:   Thu,  6 Oct 2022 18:00:44 +0200
Message-Id: <20221006160044.3397237-1-sumanthk@linux.ibm.com>
In-Reply-To: <166434824149.401.4361243714612738808.tip-bot2@tip-bot2>
References: <166434824149.401.4361243714612738808.tip-bot2@tip-bot2>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi,

This causes segfaults.

Steps to recreate:
*  Run ./samples/bpf/trace_output
BUG pid 9 cookie 1001000000004 sized 4
BUG pid 9 cookie 1001000000004 sized 4
BUG pid 9 cookie 1001000000004 sized 4
Segmentation fault (core dumped)

Problem:
* The following commit sets data->raw to NULL, when the raw data is not filled
by PMU driver. This leads to stale data.
   
* raw data could also be filled by bpf_perf_event_output(), bpf_event_output()
...
 686         perf_sample_data_init(sd, 0, 0);
 687         sd->raw = &raw;
 688
 689         err = __bpf_perf_event_output(regs, map, flags, sd);
...

* The below patch eliminates segfaults. However, contradicts with
the description mentioned in this commit (Filled by only PMU driver).
  
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 49fb9ec8366d..1ed08967fb97 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -687,6 +687,7 @@ BPF_CALL_5(bpf_perf_event_output, struct pt_regs *, regs, struct bpf_map *, map,
 
        perf_sample_data_init(sd, 0, 0);
        sd->raw = &raw;
+       sd->sample_flags |= PERF_SAMPLE_RAW;
 
        err = __bpf_perf_event_output(regs, map, flags, sd);
 
@@ -745,6 +746,7 @@ u64 bpf_event_output(struct bpf_map *map, u64 flags, void *meta, u64 meta_size,
        perf_fetch_caller_regs(regs);
        perf_sample_data_init(sd, 0, 0);
        sd->raw = &raw;
+       sd->sample_flags |= PERF_SAMPLE_RAW;
 
        ret = __bpf_perf_event_output(regs, map, flags, sd);
 out:
  
--
Thanks,
Sumanth