diff options
| author | Benjamin Tissoires <bentiss@kernel.org> | 2026-09-04 14:53:00 +0200 |
|---|---|---|
| committer | Benjamin Tissoires <bentiss@kernel.org> | 2026-09-14 18:06:03 +0200 |
| commit | c4afa4862b878d56e0cc1021298794ac1b45bc49 (patch) | |
| tree | 21c7280627e70884f5ab3d9db35fc771e1f13073 /tools/perf/scripts/python/task-analyzer.py | |
| parent | 9d1e523b92d8cb11a4a7e5a2655d6824c2d0f6e3 (diff) | |
| download | linux-next-c4afa4862b878d56e0cc1021298794ac1b45bc49.tar.gz linux-next-c4afa4862b878d56e0cc1021298794ac1b45bc49.zip | |
HID: bpf: fix __hid_bpf_hw_check_params report length
Turns out that USB, I2C and other transport drivers (except uhid which
just passes the data) still need to have the report ID in the first
byte.
Because they expect the first byte to be the report ID or 0, when the
report ID is 0, they strip that first byte before forwarding to the
device. This means that the transport layer forwards a buffer of size
N-1 to the device, which gets rejected.
Fixes: 5599f8019661 ("HID: bpf: export hid_hw_output_report as a BPF kfunc")
Signed-off-by: Benjamin Tissoires <bentiss@kernel.org>
Diffstat (limited to 'tools/perf/scripts/python/task-analyzer.py')
0 files changed, 0 insertions, 0 deletions