Bug 69532

Summary:	Generated code for short-lifetime blocks should be self-checking
Product:	[Developer tools] valgrind	Reporter:	Jeremy Fitzhardinge <jeremy>
Component:	general	Assignee:	Julian Seward <jseward>
Status:	RESOLVED DUPLICATE
Severity:	normal	CC:	lkindness
Priority:	NOR
Version:	2.1 CVS
Target Milestone:	---
Platform:	Unlisted Binaries
OS:	Linux
Latest Commit:		Version Fixed In:

Description Jeremy Fitzhardinge 2003-12-02 23:40:33 UTC

Some basic blocks are very short lived, like a stack thunk generated by gcc as
part of its nested function calling feature.

We should recognize that these basic blocks are short lived, and generate code
into them which check whether the cached generated code is still valid, or if
the original code has changed.  This could be something as simple as a checksum.

We could do this by marking certain segments, such as stack segments, as
containing highly volatile code, and therefore the codegen should build checks
into them.

Assuming that short-lived code is not performance-critical, it should be cheaper
for the code to be self-checking, than checking every other memory-modifying
operation to see if it hits some cached code.

Comment 1 Dirk Mueller 2003-12-05 17:41:56 UTC


*** This bug has been marked as a duplicate of 69511 ***

Comment 2 Jeremy Fitzhardinge 2003-12-05 19:16:52 UTC

Subject: Re:  Generated code for short-lifetime blocks should be
	self-checking

On Fri, 2003-12-05 at 08:41, Dirk Mueller wrote:
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.
>      
> http://bugs.kde.org/show_bug.cgi?id=69532     
> mueller@kde.org changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>              Status|UNCONFIRMED                 |RESOLVED
>          Resolution|                            |DUPLICATE
> 
> 
> 
> ------- Additional Comments From mueller@kde.org  2003-12-05 17:41 -------
> 
> 
> *** This bug has been marked as a duplicate of 69511 ***

Hm, does the KDE bugzilla not implement the bug dependency stuff?  I was
going to mark 69511 as depending on 69532.

	J