[Bug 297777] geom_part: access leaked when a resized table is spoiled, wedging all USB enumeration
| Newsgroups | gmane.os.freebsd.bugs |
|---|---|
| Message-ID | <[email protected]/bugzilla/> |
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=297777 Warner Losh <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|New |In Progress --- Comment #7 from Warner Losh <[email protected]> --- OK. I've looked at the code, and agree the proposed fix is good. I took some time to study g_part.c, and also asked claude to validate your bug report as a sanity check. I then took some time to create a patch, that claude verified, that I've attached. It implements your idea of moving the checks into g_part_wither(), which I think should prevent future problems better. I was nervous about just doing a limited patch since there's so many paths to wither, and I prefer consolidating reference count things like this as much as possible because whack-a-mole often leads to bugs later. -- You are receiving this mail because: You are the assignee for the bug.