Skip to content

Attempt to fix a probable bug introduced by GPR#156.#279

Merged
alainfrisch merged 1 commit intoocaml:trunkfrom
alainfrisch:fix_frametables
Nov 5, 2015
Merged

Attempt to fix a probable bug introduced by GPR#156.#279
alainfrisch merged 1 commit intoocaml:trunkfrom
alainfrisch:fix_frametables

Conversation

@alainfrisch
Copy link
Copy Markdown
Contributor

See my last note on #156.

It would be interesting to re-run benchmarks done when accepting #156 to see if the gains are confirmed after this fix.

alainfrisch added a commit that referenced this pull request Nov 5, 2015
Attempt to fix a probable bug introduced by GPR#156.
@alainfrisch alainfrisch merged commit 3cc3932 into ocaml:trunk Nov 5, 2015
@alainfrisch
Copy link
Copy Markdown
Contributor Author

We've got confirmation that this patch fixes the segfaults on several tests for our customer, so I've merged this PR.

I confirm this patch is the cause of some segfaults; we've seen this on Coq with flambda at high inlining > levels.

@mshinwell @chambart Can you confirm that the patch proposed here (now merged) fixes those segfaults?

@chambart
Copy link
Copy Markdown
Contributor

chambart commented Nov 6, 2015

@alainfrisch The fix seems correct. In fact it was missing only missing

   frametables = new_frametables

There is no need to move the last expressions into the if branches (frametables is correctly NULL at that point in the first branch)

Sorry for the bug, it seems that I introduced it while renaming things in the patch chambart@3fdaf02#diff-3c7717e893a739a54e5fbc92ec4b7100L148

@alainfrisch
Copy link
Copy Markdown
Contributor Author

frametables is correctly NULL at that point in the first branch

I don't understand. Since frametables is NULL, tail->next = frametables will set tail->next to NULL, but tail still points to the last new frametable; this will in effect remove previous frametables from the linked list. Or did I miss something?

@chambart
Copy link
Copy Markdown
Contributor

chambart commented Nov 6, 2015

Effectively, sorry...

@xavierleroy
Copy link
Copy Markdown
Contributor

Is this solved to everyone's satisfaction? If so please close this PR as well as PR#156.

@alainfrisch
Copy link
Copy Markdown
Contributor Author

I closed #156, but for some reason, I don't see a "Close" button on this one.

@gasche
Copy link
Copy Markdown
Member

gasche commented Nov 13, 2015

It is automatically considered closed (in fact "Merged") since you merged the corresponding patch -- this PR does not appear in the list of open PRs anymore.

@alainfrisch
Copy link
Copy Markdown
Contributor Author

Ah thanks, I was confused by Xavier's request.

@xavierleroy
Copy link
Copy Markdown
Contributor

My bad, I was probably looking at an outdated Web page generated by Github.

stedolan pushed a commit to stedolan/ocaml that referenced this pull request Feb 20, 2020
mshinwell pushed a commit to mshinwell/ocaml that referenced this pull request Jan 6, 2021
* Various improvements related to unboxing and meet

* Review: Catch invalid field reads earlier

* Add comment on Row_like.For_blocks.get_field
chambart pushed a commit to chambart/ocaml-1 that referenced this pull request Jan 4, 2022
EmileTrotignon pushed a commit to EmileTrotignon/ocaml that referenced this pull request Jan 12, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants