Draft: Circumvent gcc 12.2.0 segfault #23950
Draft
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
gcc 12.2.0 previously segfaulted trying to inline a bunch of code. This
can be reproduced prior to these patches with:
```bash
$ make clean && ./configure && CFLAGS=-O2 make -j 4
```
which produces:
```
during RTL pass: expand
p/pickle/plugin.c: In function ‘pickle_encode’:
p/pickle/plugin.c:632:17: internal compiler error: Segmentation fault
632 | switch (ob) {
| ^~~~~~
0x7fc4bd56604f ???
./signal/../sysdeps/unix/sysv/linux/x86_64/libc_sigaction.c:0
0x7fc4bd697f7a __memset_avx512_unaligned_erms
../sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S:236
Please submit a full bug report, with preprocessed source (by using -freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-12/README.Bugs> for instructions.
```
gcc 14.2.0 doesn't segfault, but tries and fails to allocate 32G of
memory:
```
cc1: out of memory allocating 34359738352 bytes after a total of 3411968 bytes
```
```
╭─vagrant@testing ~/scratch/radare2 (master)
╰─➤ sudo dmesg | tail -3
[ 4137.211190] __vm_enough_memory: pid: 90750, comm: cc1, bytes: 34359742464 not enough memory for the allocation
[ 4137.211210] __vm_enough_memory: pid: 90750, comm: cc1, bytes: 34359635968 not enough memory for the allocation
[ 4137.211217] __vm_enough_memory: pid: 90750, comm: cc1, bytes: 34359873536 not enough memory for the allocation
```
The root cause requires additional debugging and should possibly be
reported upstream to the gcc bug tracker. However, these patches fix
compilation for now with gcc 12.2.0 using `CLFAGS=-O2`.
|
This compiled no problem yesterday night, but segfaults again today. I don't think these patches are stable and fix the issue at hand. Pushing a Draft PR so that there's some documentation of what's been tried and how to reproduce this somewhere else than on just my machine. I'm totally unreachable for the next week and open to anyone else taking a stab at this. I'll continue wrestling with this when I'm back. |
|
Started failing again after rebasing. Pulling more stuff out of the switch might be worth a try. Horrible hack, but might fix the segfault. ;) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
gcc 12.2.0 previously segfaulted trying to inline a bunch of code. This can be reproduced prior to these patches with:
which produces:
gcc 14.2.0 doesn't segfault, but tries and fails to allocate 32G of memory:
The root cause requires additional debugging and should possibly be reported upstream to the gcc bug tracker. However, these patches fix compilation for now with gcc 12.2.0 using
CLFAGS=-O2.