On 03/09/2025 18:22, Alexei Starovoitov wrote:
On Wed, Sep 3, 2025 at 9:20 AM Lecomte, Arnaud <contact@xxxxxxxxxxxxxx> wrote:
On 03/09/2025 18:12, Alexei Starovoitov wrote:
On Wed, Sep 3, 2025 at 6:52 AM Arnaud Lecomte <contact@xxxxxxxxxxxxxx> wrote:
A new helper function stack_map_calculate_max_depth() that
computes the max depth for a stackmap.
Changes in v2:
- Removed the checking 'map_size % map_elem_size' from
stack_map_calculate_max_depth
- Changed stack_map_calculate_max_depth params name to be more generic
Changes in v3:
- Changed map size param to size in max depth helper
Changes in v4:
- Fixed indentation in max depth helper for args
Changes in v5:
- Bound back trace_nr to num_elem in __bpf_get_stack
- Make a copy of sysctl_perf_event_max_stack
in stack_map_calculate_max_depth
Changes in v6:
- Restrained max_depth computation only when required
- Additional cleanup from Song in __bpf_get_stack
This is not a refactor anymore.
Pls don't squash different things into one patch.
Keep refactor as patch 1, and another cleanup as patch 2.
The main problem is that patch 2 is not a cleanup too. It is a bug fix
so it doesn't really
fit either.
We could maybe split this patch into 2 new patches but I don't really
like this idea.
If we decide to stick to 2 patches format, I don't have any preference
which patch's scope
should be extended.
I wasn't proposing to squash cleanup into patch 2.
Make 3 patches where each one is doing one thing.
I've sent it:
https://lore.kernel.org/all/20250903233910.29431-1-contact@xxxxxxxxxxxxxx/
Thanks !
Arnaud
