|SUCCESS| net/gve: check the Rx scatter MTU at device start
dpdklab at iol.unh.edu
dpdklab at iol.unh.edu
Thu Oct 22 10:15:27 CEST 2026
Test-Label: ai-code-review
Test-Status: SUCCESS
https://dpdk.org/patch/171827
_AI Code Review_
=== Patch Review: series.patch (via Claude) ===
Review date: 2026-10-22
# DPDK Patch Review
## ERRORS
None identified.
## WARNINGS
None identified.
## INFO
**Inconsistent return variable naming**
The patch uses both `ret` and `err` as return value variables in the same file context. In `gve_dev_start()`, the existing code uses `ret`, but the new code assigns the result of `gve_rx_check_scatter()` to `ret`. In `gve_dev_mtu_set()`, the code uses `err`. While both patterns work, consistency within a file is preferred per DPDK style guidelines. Consider using `ret` consistently, or following the existing pattern in each function.
```c
/* In gve_dev_mtu_set(), could use: */
ret = gve_rx_check_scatter(dev, mtu);
if (ret != 0)
return ret;
ret = process_priv->ctrl_ops->set_mtu(priv, mtu);
if (ret) {
/* ... */
}
```
This is a minor style preference rather than a correctness issue.
---
**Positive findings:**
- The patch correctly moves the scatter check to a common function, eliminating code duplication
- Error handling is appropriate - the check returns `-EINVAL` with a descriptive log message
- The new `gve_rx_check_scatter()` function properly handles the case when `rxq` is NULL
- The check correctly identifies that this only applies to GQI mode and when scatter is not enabled
- The fix addresses the correctness issue where MTU changes via `rte_eth_dev_configure()` bypassed validation
- NULL check on `rxq` before dereferencing is correct
- The function follows DPDK naming conventions with `gve_` prefix
- Comment in `gve_dev_start()` accurately explains why the recheck is needed
Review-Result: CLEAN
More information about the test-report
mailing list