mirror of
https://github.com/zephyrproject-rtos/zephyr
synced 2025-08-08 02:06:29 +00:00
ca7b3b3703
- differentiating actual "fragment" to buffer element from a net_pkt - instead of copying data (and thus allocating buffer) let's just take the fragment buffer and put it into result packet. - fixing compilation issue in relevant test It's more efficient that way, as we use already allocated fragment buffer instead of reallocating/deallocating after each fragment reception. A possible optimization would be to calculate the actual size + header difference if only the actual size is close to the target size. It would avoid to get a dry run of the header decompression for each fragment received. The difficulty being in finding the sweet spot when it is relevant to calculate the header difference. Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com> |
||
|---|---|---|
| .. | ||
| crypto | ||
| fragment | ||
| l2 | ||