rpcgen: Don't free() a pointer after realloc().
ClosedPublic
Actions

Authored by jhb on Nov 29 2022, 7:54 PM.

Details

Reviewers

emaste
brooks
imp

Commits

rGa7b1263ef066: rpcgen: Don't free() a pointer after realloc().
rGf0b58b190107: rpcgen: Don't free() a pointer after realloc().

Summary

A successful realloc() already frees the old pointer.

Reported by: GCC -Wuse-after-free

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Skipped

Unit

Tests Skipped

Build Status

Buildable 48541
Build 45427: arc lint + arc unit

Event Timeline

jhb created this revision.Nov 29 2022, 7:54 PM

Herald added a subscriber: imp. · View Herald TranscriptNov 29 2022, 7:54 PM

jhb requested review of this revision.Nov 29 2022, 7:54 PM

Harbormaster completed remote builds in B48541: Diff 113627.Nov 29 2022, 7:54 PM

brooks accepted this revision.Nov 29 2022, 7:57 PM

This revision is now accepted and ready to land.Nov 29 2022, 7:57 PM

I wonder how this ever worked.

I wonder how this ever worked.

Maybe in practice the first allocation setting argmax to 32 ends up being sufficient?

In D37540#853610, @emaste wrote:

I wonder how this ever worked.

Maybe in practice the first allocation setting argmax to 32 ends up being sufficient?

jemalloc doesn't detect most misuse of free() so likely it wasn't used as you speculate or future allocations didn't use the same size bucket enough to end up with aliasing allocations.

Closed by commit rGf0b58b190107: rpcgen: Don't free() a pointer after realloc(). (authored by jhb). · Explain WhyDec 5 2022, 12:33 AM

This revision was automatically updated to reflect the committed changes.

jhb added a commit: rGf0b58b190107: rpcgen: Don't free() a pointer after realloc()..

jhb added a commit: rGa7b1263ef066: rpcgen: Don't free() a pointer after realloc()..May 3 2023, 12:29 AM