Messages in this thread | | | Date | Fri, 5 Apr 2024 10:45:05 -0700 | From | Dan Williams <> | Subject | Re: [PATCH v2 1/1] cxl/core: Fix initialization of mbox_cmd.size_out in get event |
| |
Alison Schofield wrote: [..] > My thought was 2) device bug. Bad device is returning payload length > that exceeds what pci/cxl-driver can consume, so gets ignored. Am I > worrying about debugging the hardware needlessly?
I would not go so far as to say "needlessly", but the number of fun and interesting ways that hardware can violate software expectations is myriad, so it will always be game of after-the-fact quirks and fixups.
A payload truncation would seem to fail safely in the sense of no buffer overrun and no stale data exposure. Still a bug, but no riskier than all the other potential hardware bugs / spec violations.
| |