Hi Finn,
Am 25.04.2023 um 14:32 schrieb Michael Schmitz:
Maybe we should try modifying get_sigframe() to increase the gap between
the signal and exception frames from 0-1 long words up to 64-65 long
words.
It turns out that doing so (patch below) does make the problem go away.
Was the exception frame getting clobbered?
Might happen, if the frame gap isn't actually equal to the exception
frame extra size anymore? Aligning the start of the signal frame to the
next lower quadword boundary increases the gap size.
When setting up the sigframe, the extra is copied to the correct
location (right past struct sigframe, or into uc_filler). When moving
that exception frame into place, the assumption is that the gap is the
extra size, not more.
I'll try dropping the quadword alignment constraint - the return
trampoline still ought to remain longword aligned.
No luck - still stack corruption.
Cheers,
Michael
