26 Sep, 2006
40 commits
-
kexec: Avoid overwriting the current pgd (V4, i386)
This patch upgrades the i386-specific kexec code to avoid overwriting the
current pgd. Overwriting the current pgd is bad when CONFIG_CRASH_DUMP is used
to start a secondary kernel that dumps the memory of the previous kernel.The code introduces a new set of page tables. These tables are used to provide
an executable identity mapping without overwriting the current pgd.Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
The EDD code would scan the command line as a fixed array, without
taking account of either whitespace, null-termination, the old
command-line protocol, late overrides early, or the fact that the
command line may not be reachable from INITSEG.This should fix those problems, and enable us to use a longer command
line.Signed-off-by: H. Peter Anvin
Signed-off-by: Andi Kleen -
In i386's entry.S, FIX_STACK() needs annotation because it
replaces the stack pointer. And the rest of nmi() needs
annotation in order to compile with these new annotations.Signed-off-by: Chuck Ebbert
Signed-off-by: Andi Kleen -
A kprobe executes IRET early and that could cause NMI recursion and stack
corruption.Note: This problem was originally spotted and solved by Andi Kleen in the
x86_64 architecture. This patch is an adaption of his patch for i386.AK: Merged with current code which was a bit different.
AK: Removed printk in nmi handler that shouldn't be there in the first time
AK: Added missing include.
AK: added KPROBES_ENDSigned-off-by: Fernando Vazquez
Signed-off-by: Andi Kleen -
A kprobe executes IRET early and that could cause NMI recursion and stack
corruption.Note: This problem was originally spotted by Andi Kleen. This patch
adds fixes not included in his original patch.
[AK: Jan Beulich originally discovered these classes of bugs]Signed-off-by: Fernando Vazquez
Signed-off-by: Andi Kleen -
Mark i386-specific cpu cache functions as __cpuinit. They are all
only called from arch/i386/common.c:display_cache_info() that already is
marked as __cpuinit.Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
Mark i386-specific cpu identification functions as __cpuinit. They are all
only called from arch/i386/common.c:identify_cpu() that already is marked as
__cpuinit.Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
Mark i386-specific cpu init functions as __cpuinit. They are all
only called from arch/i386/common.c:identify_cpu() that already is marked as
__cpuinit. This patch also removes the empty function init_umc().Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
The different cpu_dev structures are all used from __cpuinit callers what
I can tell. So mark them as __cpuinitdata instead of __initdata. I am a
little bit unsure about arch/i386/common.c:default_cpu, especially when it
comes to the purpose of this_cpu.Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
cpu_dev->c_identify is only called from arch/i386/common.c:identify_cpu(), and
this after generic_identify() already has been called. There is no need to call
this function twice and hook it in c_identify - but I may be wrong, please
double check before applying.This patch also removes generic_identify() from cpu.h to avoid unnecessary
future nesting.Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
This patch enables ACPI based physical CPU hotplug support for x86_64.
Implements acpi_map_lsapic() and acpi_unmap_lsapic() to support physical cpu
hotplug.Signed-off-by: Ashok Raj
Signed-off-by: Andi Kleen
Cc: Andi Kleen
Cc: "Brown, Len"
Signed-off-by: Andrew Morton -
cyrix_identify() should be __init because transmeta_identify() is.
tsc_init() is only called from setup_arch() which is marked as __init.These two section mismatches have been detected using running modpost on
a vmlinux image compiled with CONFIG_RELOCATABLE=y.Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
There is no need to duplicate the topology_init() function.
Signed-off-by: Magnus Damm
Signed-off-by: Andi Kleen -
The implementation comes from Zach's [RFC, PATCH 10/24] i386 Vmi
descriptor changes:Descriptor and trap table cleanups. Add cleanly written accessors for
IDT and GDT gates so the subarch may override them. Note that this
allows the hypervisor to transparently tweak the DPL of the descriptors
as well as the RPL of segments in those descriptors, with no unnecessary
kernel code modification. It also allows the hypervisor implementation
of the VMI to tweak the gates, allowing for custom exception frames or
extra layers of indirection above the guest fault / IRQ handlers.Signed-off-by: Zachary Amsden
Signed-off-by: Rusty Russell
Signed-off-by: Andi Kleen -
And add proper CFI annotation to it which was previously
impossible. This prevents "stuck" messages by the dwarf2 unwinder
when reaching the top of a kernel stack.Includes feedback from Jan Beulich
Cc: jbeulich@novell.com
Signed-off-by: Andi Kleen -
enable_local_apic can now become static.
Cc: len.brown@intel.com
Signed-off-by: Adrian Bunk
Signed-off-by: Andi Kleen -
acpi_force can become static.
Cc: len.brown@intel.com
Signed-off-by: Adrian Bunk
Signed-off-by: Andi Kleen -
It's needed for external debuggers and overhead is very small.
Also make the actual notifier chain they use static
Cc: jbeulich@novell.com
Signed-off-by: Andi Kleen
-
Fix
linux/arch/i386/kernel/mpparse.c: In function #MP_bus_info#:
linux/arch/i386/kernel/mpparse.c:232: warning: comparison is always false due to limited range of data typeSigned-off-by: Andi Kleen
-
Improve Kconfig description of CONFIG_CRASH_DUMP. Previously
it was too brief to be useful.Cc: vgoyal@in.ibm.com
Cc: ebiederm@xmission.comSigned-off-by: Andi Kleen
-
This patch moves the entry.S:error_entry to .kprobes.text section,
since code marked unsafe for kprobes jumps directly to entry.S::error_entry,
that must be marked unsafe as well.
This patch also moves all the ".previous.text" asm directives to ".previous"
for kprobes section.AK: Following a similar i386 patch from Chuck Ebbert
AK: Also merged Jeremy's fix in.+From: Jeremy Fitzhardinge
KPROBE_ENTRY does a .section .kprobes.text, and expects its users to
do a .previous at the end of the function.Unfortunately, if any code within the function switches sections, for
example .fixup, then the .previous ends up putting all subsequent code
into .fixup. Worse, any subsequent .fixup code gets intermingled with
the code its supposed to be fixing (which is also in .fixup). It's
surprising this didn't cause more havok.The fix is to use .pushsection/.popsection, so this stuff nests
properly. A further cleanup would be to get rid of all
.section/.previous pairs, since they're inherently fragile.+From: Chuck Ebbert
Because code marked unsafe for kprobes jumps directly to
entry.S::error_code, that must be marked unsafe as well.
The easiest way to do that is to move the page fault entry
point to just before error_code and let it inherit the same
section.Also moved all the ".previous" asm directives for kprobes
sections to column 1 and removed ".text" from them.Signed-off-by: Chuck Ebbert
Signed-off-by: Andi Kleen -
We have a test that looks for invalid pairings of certain athlon/durons
that weren't designed for SMP, and taint accordingly (with 'S') if we find
such a configuration. However, this test shouldn't fire if there's only
a single CPU present. It's perfectly valid for an SMP kernel to boot on UP
hardware for example.AK: changed to num_possible_cpus()
Signed-off-by: Dave Jones
Signed-off-by: Andi Kleen -
Fix a very dubious piece of code in
arch/i386/kernel/cpu/common.c:cpu_init(). This clears out %fs and
%gs, but clobbers %eax in the process without telling gcc. It turns
out that gcc happens to be not using %eax at that point anyway so it
doesn't matter much, but it looks like a bomb waiting to go off.This does end up saving an instruction, because gcc wants %eax==0 for
the set_debugreg()s below.Signed-off-by: Jeremy Fitzhardinge
Signed-off-by: Andi Kleen -
Cc: jbeulich@novell.com
Signed-off-by: Andi Kleen -
Cc: jbeulich@novell.com
Signed-off-by: Andi Kleen -
Following x86-64 patches. Reuses code from them in fact.
Convert the standard backtracer to do all output using
callbacks. Use the x86-64 stack tracer implementation
that uses these callbacks to implement the stacktrace interface.This allows to use the new dwarf2 unwinder for stacktrace
and get better backtraces.Cc: mingo@elte.hu
Signed-off-by: Andi Kleen
-
- Remove unused all_contexts parameter
No caller used it
- Move skip argument into the structure (needed for
followon patches)Cc: mingo@elte.hu
Signed-off-by: Andi Kleen
-
is_at_popf() needs to test for the iret instruction as well as
popf. So add that test and rename it to is_setting_trap_flag().Also change max insn length from 16 to 15 to match reality.
LAHF / SAHF can't affect TF, so the comment in x86_64 is removed.
Signed-off-by: Chuck Ebbert
Signed-off-by: Andi Kleen -
And move one into proto.h
Cc: len.brown@intel.com
Signed-off-by: Andi Kleen -
Removes code duplication between i386/x86-64.
Not needed anymore in setup.c since early_param cleanup
Cc: len.brown@intel.com
Signed-off-by: Andi Kleen -
This patch replaces the open-coded early commandline parsing
throughout the i386 boot code with the generic mechanism (already used
by ppc, powerpc, ia64 and s390). The code was inconsistent with
whether it deletes the option from the cmdline or not, meaning some of
these will get passed through the environment into init.This transformation is mainly mechanical, but there are some notable
parts:1) Grammar: s/linux never set's it up/linux never sets it up/
2) Remove hacked-in earlyprintk= option scanning. When someone
actually implements CONFIG_EARLY_PRINTK, then they can use
early_param().
[AK: actually it is implemented, but I'm adding the early_param it in the next
x86-64 patch]3) Move declaration of generic_apic_probe() from setup.c into asm/apic.h
4) Various parameters now moved into their appropriate files (thanks Andi).
5) All parse functions which examine arg need to check for NULL,
except one where it has subtle humor value.AK: readded acpi_sci handling which was completely dropped
AK: moved some more variables into acpi/boot.cCc: len.brown@intel.com
Signed-off-by: Rusty Russell
Signed-off-by: Andi Kleen -
Lock sections don't work the new dwarf2 unwinder
This generates slightly smaller code. It adds one more taken
jump to the fast path.Also move the trampolines into semaphore.S and add proper CFI
annotations.Cc: jbeulich@novell.com
Signed-off-by: Andi Kleen
-
Move initialization of all memory end variables to as early as
possible, so that dependent code doesn't need to check whether these
variables have already been set.Change the range check in kunmap_atomic to actually make use of this
so that the no-mapping-estabished path (under CONFIG_DEBUG_HIGHMEM)
gets used only when the address is inside the lowmem area (and BUG()
otherwise).Signed-off-by: Jan Beulich
Signed-off-by: Andi Kleen -
... instead of using a CONFIG option. The config option still controls
if the resulting executable actually has unwind information.This is useful to prevent compilation errors when users select
CONFIG_STACK_UNWIND on old binutils and also allows to use
CFI in the future for non kernel debugging applications.Cc: jbeulich@novell.com
Cc: sam@ravnborg.orgSigned-off-by: Andi Kleen
-
Remove some unlinuxy ways to write function parameter definitions.
Remove some stray "return;"sNo functional change.
Cc: len.brown@intel.com
Signed-off-by: Andi Kleen -
Bugzilla #6552 says:
"In arch/i386/boot/setup.S, movw is used instead of movb for PS/2 mouse
information, although it is unsigned char. This does not harm, because
the jmp instruction overwritten by movw is used before executing movw,
and never be used again"I've no idea if this is a real bug or how it gets fixed, so I'm submitting
it for review instead of letting it die of boredom in bugzilla. Aditionally
to i386, I've changed x86-64, which mirrors the same code.Credits to Yoshinori K. Okuji, who found the problem and suggested a fix.
Signed-off-by: Diego Calleja
Signed-off-by: Andi Kleen -
The IO APIC code had lots of duplicated code to read/write 64bit
routing entries into the IO-APIC. Factor this out int common read/write
functionsIn a few cases the IO APIC lock is taken more often now, but this
isn't a problem because it's all initialization/shutdown only
slow path code.Similar to earlier x86-64 patch.
Includes a fix by Jiri Slaby for a mistake that broke resume
Signed-off-by: Andi Kleen
-
It apparently has never triggered in many years.
Signed-off-by: Andi Kleen
-
- Move them to a pure assembly file. Previously they were in
a C file that only consisted of inline assembly. Doing it in pure
assembler is much nicer.
- Add a frame.i include with FRAME/ENDFRAME macros to easily
add frame pointers to assembly functions
- Add dwarf2 annotation to them so that the new dwarf2 unwinder
doesn't get stuck on them
- Random cleanupsIncludes feedback from Jan Beulich and a UML build fix from Andrew
Morton.Cc: jbeulich@novell.com
Cc: jdike@addtoit.com
Signed-off-by: Andi Kleen -
This ports the algorithm from x86-64 (with improvements) to i386.
Previously this only worked for frame pointer enabled kernels.
But spinlocks have a very simple stack frame that can be manually
analyzed. Do this.Signed-off-by: Andi Kleen
