Convert core 8xx drivers to use in_xxxbe/in_xxx macros instead of direct
memory references.

Other than making IO accesses explicit (which is a plus for readability), a
common set of macros provides a unified place for the volatile flag to
constraint compiler code reordering.

There are several unlucky places at the moment which lack the volatile
flag.
Signed-off-by: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

Depending on how GCC is built, GCC 4 may generate altivec instructions without
user explicitly requesting vector operations in the code.  Although this is a
performance booster for user applications, it is a problem for kernel.

This patch explicitly instruct GCC to NOT generate altivec instructions while
building the kernel.

Here are some test cases I ran.

(1) build gcc 4.0.1 with '--with-cpu=7450 --enable-altivec
    --enable-cxx-flags=-mcpu=7450', and use this gcc to build kernel WITHOUT
    this kernel patch.  Kernel fail to boot up on a 7450 board because of
    altivec instructions in kernel.

(2) build gcc 4.0.1 with "--with-cpu=7450 --enable-altivec
    --enable-cxx-flags=-mcpu=7450", and use this gcc to build kernel WITH this
    kernel patch.  Kernel boot up on a 7450 board without any problem.

(3) build gcc 4.0.1 with "--with-cpu=750 --enable-cxx-flags=-mcpu=750",
    and use this gcc to build kernel with or without this kernel patch.
    Kernel boot up on a 7450 board without any problem.

This patch should also work with GCC 3 or even earlier GCC 2.95.3.
Signed-off-by: Lee Nicks <allinux@gmail.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

The merged version follows the ppc64 version pretty closely mostly,
and in fact ARCH=ppc64 now uses the arch/powerpc/xmon version.
The main difference for ppc64 is that the 'p' command to call
show_state (which was always pretty dodgy) has been replaced by
the ppc32 'p' command, which calls a given procedure (so in fact
the old 'p' command behaviour can be achieved with 'p $show_state').
Signed-off-by: Paul Mackerras <paulus@samba.org>

The sc instruction emulation can't be done the same way on 32-bit
as 64-bit yet, but this should work OK.
Signed-off-by: Paul Mackerras <paulus@samba.org>

... for consistency with 64-bit.
Signed-off-by: Paul Mackerras <paulus@samba.org>

This patch moves the XICS interrupt controller code into the
platforms/pseries directory, since it only appears on pSeries
machines.  If it ever appears on some other machine we can move it to
sysdev, although xics.c itself will need a bunch of changes in that
case to remove pSeries specific assumptions.
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

and remove the same bits from ppc64/lib/string.S.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is identical to usercopy.c from ppc64/lib.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is identical to strcase.c from ppc64/lib.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is identical to sstep.c from ppc64/lib.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is identical to mempcy.S from ppc64/lib.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is effectively the same as locks.c from ppc64/lib.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is identical to copyuser.S from ppc64/lib.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is identical to copypage.S from ppc64/lib.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

since it is identical to e2a.c from ppc64/lib
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

as it is identical to checksum.S from ppc64.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

This allows us to also use entry_64.S from the merged tree and reverts
the setup_64.c part of fda262b8.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

We still had an old copy of i8259.h lying around; this gets rid of it
and corrects the callers of i8259_init and i8259_irq.
Signed-off-by: Paul Mackerras <paulus@samba.org>

Since I sent the patch to purge bootinfo.h from ARCH=powerpc and
ARCH=ppc64, setup-common.c has come into existence, and another
#include of bootinfo.h slipped in.  This patch removes it.  It also
removes include/asm-ppc64/bootinfo.h, which somehow survived the
previous patch which was supposed to remove it.
Signed-off-by: Paul Mackerras <paulus@samba.org>

For the current time idle_6xx only applies to 6xx ppc32 CPUs
Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

"Better late than never"

Don't try to access not-present CPUs.  Conservative governor will always
oops on SMP without this fix.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=4781Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Commit id 6142891a

Andi Kleen reports that it seems to break things for some people,
and since it's purely a small optimization, revert it for now.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

... which is needed now that ARCH=ppc64 is using the merged
setup_64.c.
Signed-off-by: Paul Mackerras <paulus@samba.org>

Signed-off-by: Paul Mackerras <paulus@samba.org>

This bug is responsible for causing the infamous "Treason uncloaked"
messages that's been popping up everywhere since the printk was added.
It has usually been blamed on foreign operating systems.  However,
some of those reports implicate Linux as both systems are running
Linux or the TCP connection is going across the loopback interface.

In fact, there really is a bug in the Linux TCP header prediction code
that's been there since at least 2.1.8.  This bug was tracked down with
help from Dale Blount.

The effect of this bug ranges from harmless "Treason uncloaked"
messages to hung/aborted TCP connections.  The details of the bug
and fix is as follows.

When snd_wnd is updated, we only update pred_flags if
tcp_fast_path_check succeeds.  When it fails (for example,
when our rcvbuf is used up), we will leave pred_flags with
an out-of-date snd_wnd value.

When the out-of-date pred_flags happens to match the next incoming
packet we will again hit the fast path and use the current snd_wnd
which will be wrong.

In the case of the treason messages, it just happens that the snd_wnd
cached in pred_flags is zero while tp->snd_wnd is non-zero.  Therefore
when a zero-window packet comes in we incorrectly conclude that the
window is non-zero.

In fact if the peer continues to send us zero-window pure ACKs we
will continue making the same mistake.  It's only when the peer
transmits a zero-window packet with data attached that we get a
chance to snap out of it.  This is what triggers the treason
message at the next retransmit timeout.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>

This just makes sure that a thread's expiry times can't get reset after
it clears them in do_exit.

This is what allowed us to re-introduce the stricter BUG_ON() check in
a362f463.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This reverts commit 3de463c7.

Roland has another patch that allows us to leave the BUG_ON() in place
by just making sure that the condition it tests for really is always
true.

That goes in next.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

My G5 was being reported as an OldWorld in /proc/cpuinfo, which is
obviously not right... :)
Signed-off-by: Paul Mackerras <paulus@samba.org>

On 32-bit platforms, these convert from kernel virtual addresses
to real (physical addresses), like tophys/tovirt but they use
the same register for the source and destination.  On 64-bit
platforms, they do nothing because the hardware ignores the top
two bits of the address in real mode.

These new macros are used in fpu.S now.
Signed-off-by: Paul Mackerras <paulus@samba.org>

Untested, but "should" work...  at least this way it compiles.
Signed-off-by: Paul Mackerras <paulus@samba.org>

Signed-off-by: Paul Mackerras <paulus@samba.org>

Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

do_dabr() is not relevant on 40x or Book-E processors so dont build it
Signed-off-by: Kumar K. Gala <kumar.gala@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

In readiness for 64k pages, when THREAD_SIZE will be less than
PAGE_SIZE, ppc64 uses kmalloc() rather than __get_free_pages() to
allocate kernel stacks, and since thread_info.h was merged, so does
ppc32.  However that adds some overhead which we don't really want
when PAGE_SIZE <= THREAD_SIZE (including all ppc32 machines), so this
patch avoids it.
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

Save for the header #define, ppc32 and ppc64 versions of parport.h are
identical.  This patch merges them.
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

With ARCH=powerpc we assume the presence of a device tree, so we don't
require any support for the old bi_recs method of passing boot
parameters.  Likewise, we've never needed it for ppc64, but we still
had an include/asm-ppc64/bootinfo.h from which nothing was used.  This
patch removes that file, and all references to it in arch/ppc64 and
arch/powerpc.  A related, unused variable 'boot_mem_size' is also
removed from setup_32.c.  The bootinfo stuff remains in ARCH=ppc for
the time being.

Built and booted on Power5 (ARCH=ppc64 and ARCH=powerpc), built for
32-bit powermac (ARCH=powerpc and ARCH=ppc).
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>