This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] (analysis) bootstrap failure on armv5l-linux
Diego Novillo wrote:
On Fri, 2003-06-13 at 17:45, Jeroen Dobbelaere wrote:
Hi,
I've been looking to the problem, and this is what I see happening :
In toplev.c, we do optimization of 'basic block reordening' (bbro) :
Jeroen,
Just to make sure, this failure doesn't occur in mainline, right? The
Yes, that's correct.
failure is not in the tree optimizers, though I seem to remember you
mentioning that it only happened in tree-ssa.
It is possible that this just shows a latent bug in 'merge_blocks',
depending on the exact definition
of 'live at entry', 'live at exit' and how this data should be updated
after a merger.
In this specific case, it seems to be assumed that the liveness data
stays identical (which isn't).
Diego.
Greetings,
Jeroen Dobbelaere