[PATCH] Cygwin: setjmp/longjmp: decrement incyg after signal handling

[Corinna Vinschen]

On Dec  5 22:40, Takashi Yano wrote:
> Hi Corinna,
> 
> On Thu, 5 Dec 2024 11:51:17 +0100
> Corinna Vinschen wrote:
> > That was not the intention.  It was just something I found while looking
> > into the assembler code.  This looks like a long-standing bug, which,
> > if my description above is right, might result in threads missing
> > signals, too.  Or at least, signals being defered, because user-space
> > is running partially with the "iscyg" flag being set.
> > 
> > It would be nice if you could check this, too, to be sure I'm not
> > totally wrong here.
> 
> It looks almost good to me and your commit message sounds reasonable.
> One question.
> There are three incyg <- 0 lines in gendef that are outside of stacklock.
> Is that safe?

Hmmm. Let me check.  As a first guess I'd say yes, but...


Thanks,
Corinna