Re: [PATCH] middle-end/104786 - ICE with asm and VLA

2022-03-09 Thread Jakub Jelinek via Gcc-patches
On Wed, Mar 09, 2022 at 12:30:08PM +0100, Richard Biener wrote: > The following fixes an ICE observed with a MEM_REF allows_mem asm > operand referencing a VLA. The following makes sure to not attempt > to go the temporary creation way when we cannot. > > 2022-03-09 Richard Biener > > P

Re: [PATCH] middle-end/104786 - ICE with asm and VLA

2022-03-09 Thread Richard Biener via Gcc-patches
Yes, and that's good enough for the testcase at hand. assign_temp also handles some cases through max_int_size_in_bytes. I'm going to leave the optimization above for a followup and will test the following for now. Richard. >From 90066f8dc0353103b9e7276b7a3e18e925509ded Mon Sep 17 00:00:00 2001 From: Richard Biener Dat

Re: [PATCH] middle-end/104786 - ICE with asm and VLA

2022-03-09 Thread Jakub Jelinek via Gcc-patches
On Wed, Mar 09, 2022 at 11:07:21AM +0100, Richard Biener wrote: > The following fixes an ICE observed with a MEM_REF allows_mem asm > operand. There's code expecting INDIRECT_REFs that are now never > going to appear. The following simply treats all tcc_reference > operands the same. The INDIREC

[PATCH] middle-end/104786 - ICE with asm and VLA

2022-03-09 Thread Richard Biener via Gcc-patches
The following fixes an ICE observed with a MEM_REF allows_mem asm operand. There's code expecting INDIRECT_REFs that are now never going to appear. The following simply treats all tcc_reference operands the same. A more conservative approach would be to use INDIRECT_REF || MEM_REF which would fi