Now that the LIR stack is a bit bigger (thanks to bug 473769), lirasm --random should be able to test LIR_alloc more thoroughly. This is good because there is all sorts of special-casing for LIR_alloc. It'll need to be a bit careful because the stack still isn't all that big -- eg. we can't have the amount of stack space used increase proportionally with the block size. As part of this, the scratch space in the stack used by --random for loads and stores should be made fixed-size, not proportional to the stack size.
Created attachment 418763 [details] [diff] [review] patch The patch uses LIR_alloc a lot more. I've removed the scratch space that was being used previously, now all loads and stores are to randomly chosen LIR_alloc'd blocks.
Attachment #418763 - Flags: review?(stejohns)
Comment on attachment 418763 [details] [diff] [review] patch nit: maxExplicitlyUsedtackSzB -> maxExplicitlyUsedStackSzB
Attachment #418763 - Flags: review?(stejohns) → review+
Bustage fix for WinNT: http://hg.mozilla.org/projects/nanojit-central/rev/427799622bd9
Whiteboard: fixed-in-nanojit → fixed-in-nanojit, fixed-in-tracemonkey
Status: ASSIGNED → RESOLVED
Last Resolved: 9 years ago
Resolution: --- → FIXED
Component: Nanojit → Nanojit
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.