Date: Mon, 1 Dec 2003 23:45:43 +0100 (MET) From: Adriaan de Groot <adridg@cs.kun.nl> To: freebsd-amd64@freebsd.org Subject: Re: Varargs issues Message-ID: <Pine.GSO.4.44.0312012331250.24115-100000@odin.cs.kun.nl> In-Reply-To: <200312012208.hB1M8V1O064328@bigtex.jrv.org>
next in thread | previous in thread | raw e-mail | index | archive | help
'avond, James,
On Mon, 1 Dec 2003, James Van Artsdalen wrote:
> I don't have my Opteron hardware yet but I did build a gcc
> cross-compiler to look at your earlier bug report.
Ah, thanks.
> point values passed in registers as opposed to on the stack.
> The function prologue writes those registers to stack slots so
> var_args can refer to them later.
OK, that's pretty sure to be it. You can see in the disassembly in the PR
that it _is_ a movaps instruction causing the bus error.
> The trick is that gcc 3.3 doesn't seem to try to keep the stack
> aligned to 16-bytes, so on entry to a function the stack may be 8-byte
> aligned, 16-byte aligned, etc. If the AMD book is not out-of-date
> then a fault is bound to happen on MOVAPS at some point.
The $rbp and $rsp are both 8-aligned when it bus errors. In the main
thread, the stack is 16-aligned at every call to PrintF (I'm working with
the third example program, maybe I should attach iot somewhere).
I'll get the hard numbers for you in a minute, next message.
--
Adriaan de Groot adridg@cs.kun.nl Kamer A6020 024-3652272
GPG Key Fingerprint 934E 31AA 80A7 723F 54F9 50ED 76AC EE01 FEA2 A3FE
http://www.cs.kun.nl/~adridg/research/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.4.44.0312012331250.24115-100000>