If the IA32 code generator also cleans the deferred code stubs on stack overflows, this LGTM.
On Wed, Mar 11, 2009 at 11:14 AM, Kevin Millikin <[email protected]> wrote: > 1. Deferred code: there is a problem. I have added a call to clear the > deferred code if we hit a stack overflow during compilation. > 2. Currently, we unwind all the recursive calls to the code generator by > returning, and a stack overflow just makes subsequent calls to the code > generator return immediately. So it looks like with the current stack > overflow behavior we're OK. > New change uploaded. > > On Wed, Mar 11, 2009 at 02:39, Kasper Lund <[email protected]> wrote: >> >> I'm not sure if this is good enough yet. What happens if we get an >> exception (stack overflow) while compiling code, but we already have a >> bunch of deferred code stubs that needs processing? Shouldn't we be >> clearing those as well? How about the JumpTargets initialized in the >> code generator - can we ever get a stack overflow or something and not >> get to the Unuse call? >> >> On Wed, Mar 11, 2009 at 10:31 AM, <[email protected]> wrote: >> > Reviewers: Kasper Lund, >> > >> > Description: >> > Fix issue 268 by explicitly calling Unuse (to deallocate all contained >> > virtual frames) on zone-allocated jump targets. These include jump >> > targets in AST nodes and the entry and exit targets of deferred code. >> > >> > See http://code.google.com/p/v8/issues/detail?id=268 >> > >> > >> > Please review this at http://codereview.chromium.org/42067 >> > >> > SVN Base: http://v8.googlecode.com/svn/branches/bleeding_edge/ >> > >> > Affected files: >> > M src/codegen-arm.cc >> > M src/codegen-ia32.cc >> > M src/codegen.h >> > M src/codegen.cc >> > >> > >> > > > --~--~---------~--~----~------------~-------~--~----~ v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev -~----------~----~----~----~------~----~------~--~---
