Neil Horman comments

Results 475 comments of


                                            Neil Horman

Data races: there is wrong memory order in refcount.h

I see what that is saying, but I can't seem to reconcile it with what the compiler produces. on aarch64, CRYPTO_DOWN_REF emits this: ``` static inline int CRYPTO_DOWN_REF(CRYPTO_REF_COUNT *refcnt, int...

Data races: there is wrong memory order in refcount.h

I'm not sure I follow you there @kroeckx , its the only call in that function to the atomic fetch_add routine there, so I'm not sure how it could be...

Data races: there is wrong memory order in refcount.h

yeah, thats the issue, heres the full output: ``` void ThreadProc(atomic& counter) { 400c44: a9bc7bfd stp x29, x30, [sp, #-64]! 400c48: 910003fd mov x29, sp 400c4c: f9000fe0 str x0, [sp,...

Data races: there is wrong memory order in refcount.h

@MBkkt can you test with the associated PR please, and confirm that it fixes the issues you've encountered?

`check-format-commit.sh`: make `awk` patterns more robust; further improvements

I'm fine with the piggy-back CI errors look unrelated

The OCSP responder listens only on IPv4

Marking as inactive, to be closed at the end of 3.4 dev barring further input

The OCSP responder listens only on IPv4

the oscp responder should work @andrejpodzimek as far as I know. If you are having independent issues with it, please open a separate issue and we will investigate

Fix potential NULL dereference and use-after-free in ossl_qtx_write_pkt

can we add a test for this in quicapitest, in which the mdpl is increased, so we're sure the bug is fixed?

fix: restore missing --help in Configure, clean up output and show enabled/disabled flags, warn user of possible incorrect usage

Looks like you need a rebasee, but otherwise, LGTM

sslc s_server command prompts for password even when it is provided in the command line itself using -pass.

@levitte could you please inviestigate this and propose a solution to the issue as described?