Bug 1378
| Summary: | Coverity defect reported in rte_mempool.h : Out-of-bounds access | ||
|---|---|---|---|
| Product: | DPDK | Reporter: | dmckie |
| Component: | core | Assignee: | dev |
| Status: | UNCONFIRMED --- | ||
| Severity: | normal | CC: | dmckie, stephen |
| Priority: | Normal | ||
| Version: | 22.11 | ||
| Target Milestone: | --- | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Attachments: |
rte_mempool coverity finding
Coverity formatted output |
||
Please add link to the coverity report. Created attachment 273 [details]
rte_mempool coverity finding
Our Coverity is a local instance in our company and not accessible via public internet. I attached an image of the finding as best I could The attached image was from a scan of 21.11-1.el8 However, I updated to 21.11-3 and the same finding was reported during coverity static analysis. If this is induced by our application I haven't figured out how as of yet. I see nothing obviously wrong there. Created attachment 274 [details]
Coverity formatted output
I grabbed the html formatted coverity output for this finding. It does include a little more detail than my previous image but I still don't see anything obvious. If this doesn't reveal anything I will assume this is a false positive coverity finding for now. We have seen no related operational issues or crashes in our product testing. Thank you for your time looking at this.
|
I have searched unsuccessfully for information on this potential issue reported by coverity. Is this a known issue? or a false-positive coverity finding? Category: Memory - corruptions file: dpdk/include/rte_mempool.h static __rte_always_inline void rte_mempool_put(struct rte_mempool *mp, void *obj) { Out-of-bounds access (OVERRUN) overrun-buffer-val: Overrunning buffer pointed to by &obj of 8 bytes by passing it to a function which accesses it at byte offset 463. rte_mempool_put_bulk(mp, &obj, 1);