Re: [PATCH] Var-tracking initialization fix (PR debug/63623)

2014-10-23 Thread Jeff Law
On 10/23/14 01:30, Jakub Jelinek wrote: Hi! As I wrote in the PR, vt_stack_adjustments can often compute wrong offsets, because it never considers pops with autoinc addressing, which can lead either to wrong debug info, or turning off -fvar-tracking altogether for a function on which that issue

Re: [PATCH] Var-tracking initialization fix (PR debug/63623)

2014-10-23 Thread Richard Biener
On Thu, Oct 23, 2014 at 9:30 AM, Jakub Jelinek wrote: > Hi! > > As I wrote in the PR, vt_stack_adjustments can often compute wrong offsets, > because it never considers pops with autoinc addressing, which can lead > either to wrong debug info, or turning off -fvar-tracking altogether for > a funct

[PATCH] Var-tracking initialization fix (PR debug/63623)

2014-10-23 Thread Jakub Jelinek
Hi! As I wrote in the PR, vt_stack_adjustments can often compute wrong offsets, because it never considers pops with autoinc addressing, which can lead either to wrong debug info, or turning off -fvar-tracking altogether for a function on which that issue resulted in stack depth inconsistencies on