Mailing List Archive

Linux 2.2.0pre7ac2
I've put up 2.2.0pre7ac2 on
ftp://ftp.linux.org.uk/pub/linux/alan/2.2pre/...
Differences between 2.2.0pre7ac1 and 2.2.0pre7ac2
o Fixed Config help bug in PCI stuff
o NFS client over tcp might well work now
o IP automatic route addition can be disabled
o IPX bind to down interface sort of fixed.
o Small sound clean ups on the SB/ESS stuff
o Major bttv update - please test this hard
o Make xconfig ought to work - give it a hammering
o Updated RCPCI I2O lan driver
o AHA15xx config info update
o Typo fixes
o Added dcache doc
o APM update
o Old HD driver fixed
o Geometry fixes for IDE
o UM8886 IDE controllers irq detection fixed
o 3c59x byte counters
o Remove remaining cern_hippi bits
o ext2 fsync update
o isofs fixes
o de4x5 should compile ok now
Differences between 2.2.0pre7 and 2.2.0pre7ac1
o Config help for Apollo and for INITIO scsi
o Large fd arrays (100,000 per process!)
o Console font handling fixes in 512 char mode
o Lance allows compiling as a module
o EEPro100 has byte counters
o Some shaper bug work - still doesnt work in 2.2.0pre
o Symbios 53c416 support
o QlogicFC fibre channel controller support
o Problems with certain numbers of disks fixed
o PAS2 and SB sound fixes
o Swap/page size header stuff handled right
o NFS client knows about BSD style FIFO rule
o Memory locking fixes for /proc
o AVL tree VMA handling for speed when running debug tools
o Root-dev sysctl works for both endians
o IP masquerade bug fix for ICMP masq
o NFS over TCP mostly works
o Xconfig fixes
o IPI counter on SMP
o ver_linux script improvements
Oh yeah I left some I2O references in in error. Just say N to those 8)
Alan
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.0pre7ac2 [ In reply to ]
Alan Cox wrote:
>
> I've put up 2.2.0pre7ac2 on
> ftp://ftp.linux.org.uk/pub/linux/alan/2.2pre/...
[...]
> o isofs fixes
> o de4x5 should compile ok now
[...]
Yes, it compiles but the changes in asm/string.h make it oops
on first use. (using both gcc and egcs as per RH52 std.)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.0pre7ac2 [ In reply to ]
> > o de4x5 should compile ok now
> [...]
>
> Yes, it compiles but the changes in asm/string.h make it oops
> on first use. (using both gcc and egcs as per RH52 std.)
Ok. de4x5 will just have to stay not compiling with some egcs for 2.2.0,
sorry people.
Alan
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.0pre7ac2 [ In reply to ]
> > > o de4x5 should compile ok now
> > [...]
> >
> > Yes, it compiles but the changes in asm/string.h make it oops
> > on first use. (using both gcc and egcs as per RH52 std.)
>
Indeed. The fix tries to save the early-clobbered value of 'edi' in
another early-clobbered register ('edx') and is clearly wrong.
I suggest making 'strstr' an external function rather than an
inline. As it stands, it leaves no free general purpose registers for
the compiler to work with, so the optimization gain from inlining will
in any case be extremely limited.
Cheers,
Trond
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.0pre7ac2 [ In reply to ]
alan@lxorguk.ukuu.org.uk (Alan Cox) writes:
> > > o de4x5 should compile ok now
> > [...]
> >
> > Yes, it compiles but the changes in asm/string.h make it oops
> > on first use. (using both gcc and egcs as per RH52 std.)
>
> Ok. de4x5 will just have to stay not compiling with some egcs for 2.2.0,
> sorry people.
>
I can't believe it.
It was recently that de4x5 got fixed (wrt egcs) and now we have
problems again. How's that happened?
--
Zlatko
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
Re: Linux 2.2.0pre7ac2 [ In reply to ]
This is a multi-part message in MIME format.
--------------753F0053BCC3E36DD05A1D3A
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Zlatko Calusic wrote:
>
> alan@lxorguk.ukuu.org.uk (Alan Cox) writes:
>
> > > > o de4x5 should compile ok now
> > > [...]
> > >
> > > Yes, it compiles but the changes in asm/string.h make it oops
> > > on first use. (using both gcc and egcs as per RH52 std.)
> >
> > Ok. de4x5 will just have to stay not compiling with some egcs for 2.2.0,
> > sorry people.
> >
>
> I can't believe it.
>
> It was recently that de4x5 got fixed (wrt egcs) and now we have
> problems again. How's that happened?
> --
> Zlatko
There was a new patch for asm-i386/string.h proposed in the list
that was changing the last "g" operator with a "b" operator. I've
checked it and seems to work OK with both gcc and egcs...
OTOS I think that it is too a big fuss for such a little issue: strstr
does not seem to be a performance critical function anyway...I have used
the C library version and I have not appreciated any performance loss.
--------------753F0053BCC3E36DD05A1D3A
Content-Type: text/plain; charset=us-ascii;
name="asm_i386_string.h.diff"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="asm_i386_string.h.diff"
diff -u --recursive --new-file v2.2.0-pre7/linux/include/asm-i386/string.h linux/include/asm-i386/string.h
--- v2.2.0-pre7/linux/include/asm-i386/string.h Thu Dec 31 10:29:02 1998
+++ linux/include/asm-i386/string.h Thu Jan 14 10:31:41 1999
@@ -213,7 +213,7 @@
"jne 1b\n\t"
"xorl %%eax,%%eax\n\t"
"2:"
- :"=a" (__res), "=&c" (d0), "=&S" (d1), "=&d" (d2), "=&D" (d3) : "0" (0),"1" (0xffffffff),"2" (cs),"g" (ct));
+ :"=a" (__res), "=&c" (d0), "=&S" (d1), "=&d" (d2), "=&D" (d3) : "0" (0),"1" (0xffffffff),"2" (cs),"b" (ct));
return __res;
}

--------------753F0053BCC3E36DD05A1D3A--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/