|WARNING| net/bnxt: fix ctx_dma_arr allocation size

[dpdklab at iol.unh.edu]

Test-Label: ai-code-review
Test-Status: WARNING
https://dpdk.org/patch/169570

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-09-15

# Review of net/bnxt: fix ctx_dma_arr allocation size

## Summary
This patch corrects an allocation size calculation error where `sizeof(rte_iova_t *)` (pointer size, 8 bytes on 64-bit) was used instead of `sizeof(rte_iova_t)` (typically `uint64_t`, also 8 bytes). While the bug happens to be harmless on common 64-bit platforms where both sizes are equal, the fix is correct and improves code clarity.

---

## Errors

None.

---

## Warnings

**1. Missing context for harmlessness in commit message**

The commit message does not mention that `rte_iova_t` is `uint64_t` (8 bytes) and `rte_iova_t *` is also 8 bytes on 64-bit systems, making this a latent bug that doesn't manifest in practice. While the fix is correct, noting this in the commit message would help reviewers understand why the bug wasn't caught by testing.

Suggested addition to commit message:
```
On 64-bit systems where sizeof(rte_iova_t) == sizeof(rte_iova_t *) == 8,
this bug is harmless. However, the code should use the correct element type
for clarity and correctness.
```

---

## Info

**1. Consider adding a comment explaining the array contents**

The allocation is storing DMA addresses (`rte_iova_t` values), not pointers. A brief comment near the allocation would improve readability:

```c
/* Allocate array to store DMA addresses (rte_iova_t) for each page */
ctx_pg->ctx_dma_arr = rte_zmalloc(name, sizeof(rte_iova_t) * rmem->nr_pages, 0);
```

**2. Verify ctx_dma_arr usage matches the fix**

The patch assumes `ctx_dma_arr` is used to store `rte_iova_t` values, not pointers. Confirm that all code accessing `ctx_dma_arr` treats it as `rte_iova_t[]` rather than `rte_iova_t *[]`. This is outside the scope of this patch but should be verified during integration.

---

## Conclusion

The patch is **correct** and should be accepted. The allocation size fix improves code correctness even though the bug is latent on 64-bit platforms. The warnings are minor suggestions for documentation improvements.