xfs
[Top] [All Lists]

Re: gcc 3.2

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: gcc 3.2
From: Steve Lord <lord@xxxxxxx>
Date: 29 Aug 2002 08:46:38 -0500
Cc: stefano sartini <stefano.sartini@xxxxxxxxx>, linux-xfs@xxxxxxxxxxx
In-reply-to: <20020829144501.A2608@xxxxxxxxxxxxx>
References: <3D6E233B.4090209@xxxxxxxxx> <1030628491.25030.4.camel@xxxxxxxxxxxxxxxxxxxx> <20020829144501.A2608@xxxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
On Thu, 2002-08-29 at 08:45, Christoph Hellwig wrote:
> On Thu, Aug 29, 2002 at 08:41:31AM -0500, Steve Lord wrote:
> > 
> > On Thu, 2002-08-29 at 08:35, stefano sartini wrote:
> > 
> > > 
> > > Compiling back with the 2.95.3 gcc gave no problem...
> > > Is a compiler issue?
> > 
> > 
> > Go to line 203 of page_buf_locking.c and replace the call to ffs
> > with a call to generic_ffs.
> > 
> > Let me know if this fixes it and I will check it in to the tree.
> 
> Please don't check it in.  The following patch that went into 2.4.20-pre
> should fix the root of the problem instead of the symptom:

Ah, interesting, well, we will leave it as a workaround for folks
with earlier kernels. It is not as if this is a function we need
to optimize.

Steve

> 
> 
> --- linux-2.4.19-rc3-ac4/include/asm-i386/bitops.h    Mon Jul 29 21:12:34 2002
> +++ linux/include/asm-i386/bitops.h   Mon Jul 29 21:38:44 2002
> @@ -456,7 +456,7 @@ static __inline__ int ffs(int x)
>       __asm__("bsfl %1,%0\n\t"
>               "jnz 1f\n\t"
>               "movl $-1,%0\n"
> -             "1:" : "=r" (r) : "g" (x));
> +             "1:" : "=r" (r) : "rm" (x));
>       return r+1;
>  }
>  
-- 

Steve Lord                                      voice: +1-651-683-3511
Principal Engineer, Filesystem Software         email: lord@xxxxxxx


<Prev in Thread] Current Thread [Next in Thread>