linux-stable.git/arch, branch linux-2.6.16.y

[IA64] Fix unaligned handler for floating point instructions with base update

2008-07-19T20:45:33+00:00

The compiler team did the hard work for this distilling a problem in
large fortran application which showed up when applied to a 290MB input
data set down to this instruction:

        ldfd f34=[r17],-8

Which they noticed incremented r17 by 0x10 rather than decrementing it
by 8 when the value in r17 caused an unaligned data fault.  I tracked
it down to some bad instruction decoding in unaligned.c. The code
assumes that the 'x' bit can determine whether the instruction is
an "ldf" or "ldfp" ... which it is for opcode=6 (see table 4-29 on
page 3:302 of the SDM).  But for opcode=7 the 'x' bit is irrelevent,
all variants are "ldf" instructions (see table 4-36 on page 3:306).

Note also that interpreting the instruction as "ldfp" means that the
"paired" floating point register (f35 in the example here) will also
be corrupted.

Signed-off-by: Tony Luck 
Signed-off-by: Adrian Bunk

alpha: build fixes - force architecture

2008-07-19T13:15:43+00:00

Override compiler .arch directive for generic kernel build.

Signed-off-by: Ivan Kokshaysky 
Signed-off-by: Richard Henderson 
Signed-off-by: Adrian Bunk

x86: Replace NSC/Cyrix specific chipset access macros by inlined functions.

2008-07-14T18:08:34+00:00

Due to index register access ordering problems, when using macros a line
like this fails (and does nothing):

    setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88);

With inlined functions this line will work as expected.

Note about a side effect: Seems on Geode GX1 based systems the
"suspend on halt power saving feature" was never enabled due to this
wrong macro expansion. With inlined functions it will be enabled, but
this will stop the TSC when the CPU runs into a HLT instruction.
Kernel output something like this:
    Clocksource tsc unstable (delta = -472746897 ns)

This is the 3rd version of this patch.

 - Adding missed arch/i386/kernel/cpu/mtrr/state.c
    Thanks to Andres Salomon
 - Adding some big fat comments into the new header file
    Suggested by Andi Kleen

AK: fixed x86-64 compilation

Adrian Bunk:
Added workaround for x86_64 compilation.

Signed-off-by: Juergen Beisert 
Signed-off-by: Andi Kleen 
Signed-off-by: Adrian Bunk

[SPARC64]: Fix endless loop in cheetah_xcall_deliver().

2008-01-15T23:11:38+00:00

[ Upsteam commit: 0de56d1ab83323d604d95ca193dcbd28388dbabb ]

We need to mask out the proper bits when testing the dispatch status
register else we can see unrelated NACK bits from previous cross call
sends.

Signed-off-by: David S. Miller 
Signed-off-by: Adrian Bunk

i386: fixup TRACE_IRQ breakage

2007-11-13T07:46:02+00:00

The TRACE_IRQS_ON function in iret_exc: calls a C function without
ensuring that the segments are set properly. Move the trace function and
the enabling of interrupt into the C stub.

Signed-off-by: Peter Zijlstra 
Signed-off-by: Adrian Bunk

Handle bogus %cs selector in single-step instruction decoding (CVE-2007-3731)

2007-11-13T07:43:25+00:00

The code for LDT segment selectors was not robust in the face of a bogus
selector set in %cs via ptrace before the single-step was done.

Signed-off-by: Roland McGrath 
Signed-off-by: Adrian Bunk

[SPARC64]: Fix show_stack() when stack argument is NULL.

2007-11-02T21:56:18+00:00

It didn't handle that case at all, and now dump_stack()
can be implemented directly as show_stack(current, NULL)

Signed-off-by: David S. Miller 
Signed-off-by: Adrian Bunk

[SPARC64] pass correct addr in get_fb_unmapped_area(MAP_FIXED)

2007-11-01T02:57:51+00:00

Looks like the MAP_FIXED case is using the wrong address hint.  I'd
expect the comment "don't mess with it" means pass the request
straight on through, not change the address requested to -ENOMEM.

Signed-off-by: Chris Wright 
Signed-off-by: David S. Miller 
Signed-off-by: Adrian Bunk

[IA64] fix ia64 is_hugepage_only_range

2007-10-28T20:40:41+00:00

fix is_hugepage_only_range() definition to be "overlaps"
instead of "within architectural restricted hugetlb address
range".  Simplify the ia64 specific code that used to use
is_hugepage_only_range() to just check which region the
address is in.

Signed-off-by: Ken Chen 
Signed-off-by: Tony Luck 
Signed-off-by: Adrian Bunk

x86: HUGETLBFS and DEBUG_PAGEALLOC are incompatible

2007-10-18T23:59:17+00:00

DEBUG_PAGEALLOC is not compatible with hugetlb page support.  That debug
option turns off PSE.  Once it is turned off in CR4, the cpu will ignore
pse bit in the pmd and causing infinite page-not- present faults.

So disable DEBUG_PAGEALLOC if the user selected hugetlbfs.

Signed-off-by: Ken Chen 
Signed-off-by: Adrian Bunk