[lkml]   [1998]   [Mar]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: egcs 1.0.1 miscompiles Linux 2.0.33

In message <199803091917.OAA22156@dcl.MIT.EDU>you write:
> Could you please explain the syntax of an earlyclobber (I never
> heard of it). I would undergo the task of patching Linux-Kernel's
> asms, since I'd like to stop the debate about an EGCS/Kernel
> mismatch. EGCS is a valuable project a nd compiling a kernel should
> not delay using it.
> Is someone in the EGCS project working on clarifying the documentation?
> In the past few days, I've heard at least one EGCS developer admit that
> the documentation is "ambiguous", and Linus's original position was that
> he would gladly fix the Linux sources to match the _documented_
> behaviour of GCC, and he then listed a few reasons why he thought the
> existing GCC documentation supported his interpretation.
I will update the egcs docs as soon as we've agreed on the technical
details about how to handle these situations.

I think we're close, but I don't want to start changing the docs then
find out y'all don't agree with the direction I'm heading.

> Some of the mail messages in this thread look like they would be fine
> fodder for clarifying EGCS's documentation (in particular the mail
> message I'm currently replying to --- please, put lots of detail and
> example in the docs!! This current message was good that way.)
I'd probably sit down a review all of this discussion before updating
the docs. The key things off the top of my head that need updating:

* Clarify that overlapping inputs and clobbers are wrong.

* Clarify that a clobbered reg can not be used for an input
operand or the address of any operand.

* Clarify that earlyclobbered outputs with a matching input
is the right way to handle this situation.

* Clarify issues relating to the maximum number of registers
that can be used in an asm.

* Doc new switch (assuming that's the direction we want to go).

> A conditional switch sounds like a good compromise. But IMHO getting
> the documentation updated is just as important....
Yup. It's the poor documentation that helped get us into this
situation to start with :-)


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to

 \ /
  Last update: 2005-03-22 13:41    [W:0.118 / U:0.260 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site