> On Jun 26, 2020, at 8:40 AM, Yonghong Song <y...@fb.com> wrote:
>
>
>
> On 6/25/20 5:13 PM, Song Liu wrote:
>> Introduce helper bpf_get_task_stack(), which dumps stack trace of given
>> task. This is different to bpf_get_stack(), which gets stack track of
>> current task. One potential use case of bpf_get_task_stack() is to call
>> it from bpf_iter__task and dump all /proc/<pid>/stack to a seq_file.
>> bpf_get_task_stack() uses stack_trace_save_tsk() instead of
>> get_perf_callchain() for kernel stack. The benefit of this choice is that
>> stack_trace_save_tsk() doesn't require changes in arch/. The downside of
>> using stack_trace_save_tsk() is that stack_trace_save_tsk() dumps the
>> stack trace to unsigned long array. For 32-bit systems, we need to
>> translate it to u64 array.
>> Signed-off-by: Song Liu <songliubrav...@fb.com>
>>
[...]
>> +++ b/include/uapi/linux/bpf.h
>> @@ -3252,6 +3252,38 @@ union bpf_attr {
>> * case of **BPF_CSUM_LEVEL_QUERY**, the current skb->csum_level
>> * is returned or the error code -EACCES in case the skb is not
>> * subject to CHECKSUM_UNNECESSARY.
>> + *
>> + * int bpf_get_task_stack(struct task_struct *task, void *buf, u32 size,
>> u64 flags)
>
> Andrii's recent patch changed the return type to 'long' to align with
> kernel u64 return type for better llvm code generation.
>
> Please rebase and you will see the new convention.
Will fix.
>
>> + * Description
>>
[...]
>> +static struct perf_callchain_entry *
>> +get_callchain_entry_for_task(struct task_struct *task, u32 init_nr)
>> +{
>> + struct perf_callchain_entry *entry;
>> + int rctx;
>> +
>> + entry = get_callchain_entry(&rctx);
>> +
>> + if (rctx == -1)
>> + return NULL;
>
> Is this needed? Should be below !entry enough?
It is needed before Peter's suggestion. After applying Peter's patch,
this is no longer needed.
Thanks,
Song
