Here's the fix, I'm just validating before I push:
--- old-ghc-testing/rts/Updates.cmm 2010-05-10 16:20:22.000000000 +0100
+++ new-ghc-testing/rts/Updates.cmm 2010-05-10 16:20:25.000000000 +0100
@@ -95,5 +95,5 @@
// high watermark.
INFO_TABLE_RET (stg_bh_upd_frame, UPDATE_FRAME, UPD_FRAME_PARAMS)
{
- jump stg_marked_upd_frame_info;
+ jump RET_LBL(stg_marked_upd_frame);
}
On 10/05/2010 15:08, Simon Marlow wrote:
Thanks for narrowing it down, unfortunately I can't see any obvious problems
with that bit of code. I'll need to start a new build and look into it.
Cheers,
Simon
-----Original Message-----
From: David Terei [mailto:[email protected]]
Sent: 10 May 2010 14:58
To: Andy Georges; [email protected]; Simon Marlow
Subject: Re: GHC Head broken when TNTC not used
Hi all and Simon,
OK this is the patch that breaks GHC when TablesNextToCode isn't used:
Tue Mar 30 01:44:56 EST 2010 Simon Marlow<[email protected]>
* New implementation of BLACKHOLEs
Specifically this patch causes anything compiled by the ghc-stage1
compiler to segfault in 'stg_bh_upd_frame_info ()'.
Cheers,
David
On 10 May 2010 12:46, David Terei<[email protected]> wrote:
Hi all,
Sometime in the last month or so a change in GHC head broke the
compiler when 'TablesNextToCode' isn't used. E.g, build GHC with this
line in your build.mk:
GhcEnableTablesNextToCode = NO
When compiling ghc, the first stage will build but then anything
compiled with ghc-stage1 simply segfaults when run. I'll start trying
to figure out the patch that introduced this bug but feel free to help
out :).
Cheers,
David
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
