Default CONFIG_PHYSICAL_START and CONFIG_PHYSICAL_ALIGN each to 16 MB,
so that both non-relocatable and relocatable kernels are loaded at
16 MB by a non-relocating bootloader. This is somewhat hacky, but it
appears to be the only way to do this that does not break some some
set of existing bootloaders.
We want to avoid the bottom 16 MB because of large page breakup,
memory holes, and ZONE_DMA. Embedded systems may need to reduce this,
or update their bootloaders to be aware of the new min_alignment field.
[ Impact: performance improvement, avoids problems on some systems ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Make the kernel_alignment field adjustable; this allows us to set it
to a large value (intended to be 16 MB to avoid ZONE_DMA contention,
memory holes and other weirdness) while a smart bootloader can still
force a loading at a lesser alignment if absolutely necessary.
Also export pref_address (preferred loading address, corresponding to
the link-time address) and init_size, the total amount of linear
memory the kernel will require during initialization.
[ Impact: allows better kernel placement, gives bootloader more info ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Remove a couple of lines of dead code from
arch/x86/boot/compressed/head_*.S; all of these update registers that
are dead in the current code.
[ Impact: cleanup ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Use LOAD_PHYSICAL_ADDR instead of CONFIG_PHYSICAL_START in the 64-bit
decompression code, for equivalence with the 32-bit code.
[ Impact: cleanup, increases code similarity ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Make symbols from the main vmlinux, as opposed to just
compressed/vmlinux, available to header.S. Also, export a few
additional symbols.
This will be used in a subsequent patch to export the total memory
footprint of the kernel.
[ Impact: enable future enhancement ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Determine the compressed code offset (from the kernel runtime address)
at compile time. This allows some minor optimizations in
arch/x86/boot/compressed/head_*.S, but more importantly it makes this
value available to the build process, which will enable a future patch
to export the necessary linear memory footprint into the bzImage
header.
[ Impact: cleanup, future patch enabling ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
In the pre-decompression code, use the appropriate largest possible
rep movs and rep stos to move code and clear bss, respectively. For
reverse copy, do note that the initial values are supposed to be the
address of the first (highest) copy datum, not one byte beyond the end
of the buffer.
rep strings are not necessarily the fastest way to perform these
operations on all current processors, but are likely to be in the
future, and perhaps more importantly, we want to encourage the
architecturally right thing to do here.
This also fixes a couple of trivial inefficiencies on 64 bits.
[ Impact: trivial performance enhancement, increase code similarity ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
The 64-bit code already clears EFLAGS as soon as it has a stack. This
seems like a reasonable precaution, so do it on 32 bits as well.
[ Impact: extra paranoia ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Set up the decompression stack as soon as we know where it needs to
go. That way we have a full-service stack as soon as possible, rather
than relying on the BP_scratch field.
Note that the stack does need to be empty during bss zeroing (or
else the stack needs to be moved out of the bss segment, which is also
an option.)
[ Impact: cleanup, minor paranoia ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Both on 32 and 64 bits, we copy all the way up to the end of bss,
except that on 64 bits there is a hack to avoid copying on top of the
page tables. There is no point in copying bss at all, especially
since we are just about to zero it all anyway.
To clean up and unify the handling, we now do:
- copy from startup_32 to _bss.
- zero from _bss to _ebss.
- the _ebss symbol is aligned to an 8-byte boundary.
- the page tables are moved to a separate section.
Use _bss as the copy endpoint since _edata may be misaligned.
[ Impact: cleanup, trivial performance improvement ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Clean up style issues in arch/x86/boot/compressed/head_64.S. This
file had a lot fewer style issues than its 32-bit cousin, but the ones
it has are worth fixing, especially since it makes the two files more
similar.
[ Impact: cleanup, no object code change ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Reformat arch/x86/boot/compressed/head_32.S to be closer to currently
preferred kernel assembly style, that is:
- opcode and operand separated by tab
- operands separated by ", "
- C-style comments
This also makes it more similar to head_64.S.
[ Impact: cleanup, no object code change ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
When generating the compression suffix in
arch/x86/boot/compressed/Makefile, follow standard Kbuild
conventions, that is:
- Use a dash not underscore before y/m/n endings
- Use := whenever possible.
Requested-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Simplify the arch/x86/boot/compressed/Makefile, by using the new
capability of specifying multiple inputs to a compressor, and the
CONFIG_X86_NEED_RELOCS Kconfig symbol.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
We only need to build relocations when we are building a 32-bit
relocatable kernel. Rather than unnecessarily complicating the
Makefiles, make an explicit Kbuild symbol for this.
[ Impact: permits future cleanup ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Aligning the .bss section makes it trivial to use large operation
sizes for moving the initialized sections and clearing the .bss.
The alignment chosen (L1 cache) is somewhat arbitrary, but should be
large enough to avoid all known performance traps and small enough to
not cause troubles.
[ Impact: trivial performance enhancement, future patch prep ]
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
32 bit:
- increase alignment from 4 to 8 for .parainstructions
- increase alignment from 4 to 8 for .altinstructions
64 bit:
- move ALIGN() outside output section for .altinstructions
None of the above should result in any functional change.
[ Impact: refactor and unify linker script ]
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tim Abbott <tabbott@MIT.EDU>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <1240991249-27117-10-git-send-email-sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
32-bit:
- Move definition of __init_begin outside output_section
because it covers more than one section
- Move ALIGN() for end-of-section inside .smp_locks output section.
Same effect but the intent is better documented that
we need both start and end aligned.
64-bit:
- Move ALIGN() outside output section in .init.setup
- Deleted unused __smp_alt_* symbols
None of the above should result in any functional change.
[ Impact: refactor and unify linker script ]
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tim Abbott <tabbott@MIT.EDU>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <1240991249-27117-9-git-send-email-sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
For 64 bit the following functional changes are introduced:
- .data.page_aligned has moved
- .data.cacheline_aligned has moved
- .data.read_mostly has moved
- ALIGN() moved out of output section for .data.cacheline_aligned
- ALIGN() moved out of output section for .data.page_aligned
Notice that 32 bit and 64 bit has different location of _edata.
.data_nosave is 32 bit only as 64 bit is special due to PERCPU.
[ Impact: 32-bit: cleanup, 64-bit: use 32-bit linker script ]
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tim Abbott <tabbott@MIT.EDU>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <1240991249-27117-7-git-send-email-sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
32 bit x86 had a dedicated .text.head output section,
whereas 64 bit had it all in a single output section.
In the unified version the dedicated .text.head output section
was kept to have full control over the head code.
32 bit:
- Moved definition of _stext to the linker script.
The definition is located _after_ .text.page_aligned as this
is what 32 bit did before.
The ALIGN(8) was introduced so we hit the exact same address
(on the tested config) before and after the move.
I assume that it is a bug that _stext did not cover the
.text.page_aligned section - if this is true it can be fixed
in a follow-up patch (and the ugly ALIGN() can be dropped).
[ Impact: 64-bit: cleanup, 32-bit: use the 64-bit linker script ]
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tim Abbott <tabbott@MIT.EDU>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <1240991249-27117-5-git-send-email-sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
PHDRS are not equal for the two - so
use ifdefs to cover up for that.
On the assumption that they may become equal the ifdef
is inside the PHDRS definiton.
[ Impact: cleanup ]
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tim Abbott <tabbott@MIT.EDU>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <1240991249-27117-3-git-send-email-sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Beautify vmlinux_32.lds.S:
- Use tabs for indent
- Located curly braces like in C code
- Rearranged a few comments
To see actual differences use "git diff -b" which
ignore 'whitespace' changes.
The beautification is done to prepare a unification
of the _32 and _64 variants of the linker scripts.
[ Impact: cleanup ]
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tim Abbott <tabbott@MIT.EDU>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <1240991249-27117-1-git-send-email-sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Look at the:
diff -u arch/x86/boot/compressed/vmlinux_*.lds
output and realize that they're basially exactly the same except for
trivial naming differences, and the fact that the 64-bit version has a
"pgtable" thing.
So unify them.
There's some trivial cleanup there (make the output format a Kconfig thing
rather than doing #ifdef's for it, and unify both 32-bit and 64-bit BSS
end to "_ebss", where 32-bit used to use the traditional "_end"), but
other than that it's really very mindless and straigt conversion.
For example, I think we should aim to remove "startup_32" vs "startup_64",
and just call it "startup", and get rid of one more difference. I didn't
do that.
Also, notice the comment in the unified vmlinux.lds.S talks about
"head_64" and "startup_32" which is an odd and incorrect mix, but that was
actually what the old 64-bit only lds file had, so the confusion isn't
new, and now that mixing is arguably more accurate thanks to the
vmlinux.lds.S file being shared between the two cases ;)
[ Impact: cleanup, unification ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Beautify vmlinux_64.lds.S:
- Use tabs for indent
- Located curly braces like in C code
- Rearranged a few comments
There is no functional changes in this patch
The beautification is done to prepare a unification
of the _32 and the _64 variants of the linker scripts.
[ Impact: cleanup ]
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tim Abbott <tabbott@MIT.EDU>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <20090426210742.GA3464@uranus.ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Introducing this Kbuild file allow us to:
make arch/x86/
And thus building all the core part of x86.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Jaswinder Singh Rajput <jaswinder@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
It turns out that 'smp_call_function_many()' doesn't work at all like
'smp_call_function_single()', and my change to Andrew's patch to use it
rather than a loop over all CPU's acpi-cpufreq doesn't work.
My bad.
'smp_call_function_many()' has two "features" (aka "documented bugs"):
(a) it needs to be called with preemption disabled, because it uses
smp_processor_id() without guarding the CPU lookup with 'get_cpu()'
and 'put_cpu()' like the 'single' variant does.
(b) even if the current CPU is part of the CPU mask, it won't do the
call on that CPU.
Still, we're better off trying to use 'smp_call_function_many()' than
looping over CPU's, since it at least in theory allows us to use a
broadcast IPI and do it all in parallel. So let's just work around the
silly semantic bugs in that function.
Reported-and-tested-by: Ali Gholami Rudi <ali@rudi.ir>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Dave Jones <davej@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Revert part of af5c820a31 ("x86: cpumask:
use work_on_cpu in arch/x86/kernel/microcode_core.c")
That change is causing only one Intel CPU's microcode to be updated e.g.
microcode: CPU3 updated from revision 0x9 to 0x17, date = 2005-04-22
where before it announced that also for CPU0 and CPU1 and CPU2.
We cannot use work_on_cpu() in the CONFIG_MICROCODE_OLD_INTERFACE code,
because Intel's request_microcode_user() involves a copy_from_user() from
/sbin/microcode_ctl, which therefore needs to be on that CPU at the time.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We ended up incorrectly using '&cur' instead of '&readin' in the
work_on_cpu() -> smp_call_function_single() transformation in commit
01599fca67 ("cpufreq: use
smp_call_function_[single|many]() in acpi-cpufreq.c").
Andrew explains:
"OK, the acpi tree went and had conflicting changes merged into it after
I'd written the patch and it appears that I incorrectly reverted part
of 18b2646fe3 while fixing the resulting
rejects.
Switching it to `readin' looks correct."
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'for-rc1/xen/core' of git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen:
xen: add FIX_TEXT_POKE to fixmap
xen: honour VCPU availability on boot
xen: clean up gate trap/interrupt constants
xen: set _PAGE_NX in __supported_pte_mask before pagetable construction
xen: resume interrupts before system devices.
xen/mmu: weaken flush_tlb_other test
xen/mmu: some early pagetable cleanups
Xen: Add virt_to_pfn helper function
x86-64: remove PGE from must-have feature list
xen: mask XSAVE from cpuid
NULL noise: arch/x86/xen/smp.c
xen: remove xen_load_gdt debug
xen: make xen_load_gdt simpler
xen: clean up xen_load_gdt
xen: split construction of p2m mfn tables from registration
xen: separate p2m allocation from setting
xen: disable preempt for leave_lazy_mmu
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: add linux kernel support for YMM state
x86: fix wrong section of pat_disable & make it static
x86: Fix section mismatches in mpparse
x86: fix set_fixmap to use phys_addr_t
x86: Document get_user_pages_fast()
x86, intr-remap: fix eoi for interrupt remapping without x2apic
Atttempting to rid us of the problematic work_on_cpu(). Just use
smp_call_fuction_single() here.
This repairs a 10% sysbench(oltp)+mysql regression which Mike reported,
due to
commit 6b44003e5c
Author: Andrew Morton <akpm@linux-foundation.org>
Date: Thu Apr 9 09:50:37 2009 -0600
work_on_cpu(): rewrite it to create a kernel thread on demand
It seems that the kernel calls these acpi-cpufreq functions at a quite
high frequency.
Valdis Kletnieks also reports that this causes 70-90 forks per second on
his hardware.
Cc: Valdis.Kletnieks@vt.edu
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Dave Jones <davej@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Mike Galbraith <efault@gmx.de>
Cc: "Zhang, Yanmin" <yanmin_zhang@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Ingo Molnar <mingo@elte.hu>
[ Made it use smp_call_function_many() instead of looping over cpu's
with smp_call_function_single() - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Impact: save/restore Intel-AVX state properly between tasks
Intel Advanced Vector Extensions (AVX) introduce 256-bit vector processing
capability. More about AVX at http://software.intel.com/sites/avx
Add OS support for YMM state management using xsave/xrstor infrastructure
to support AVX.
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
LKML-Reference: <1239402084.27006.8057.camel@localhost.localdomain>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
pat_disable cannot be __cpuinit anymore because it's called from pat_init
and the callchain looks like this:
pat_disable [cpuinit] <- pat_init <- generic_set_all <-
ipi_handler <- set_mtrr <- (other non init/cpuinit functions)
WARNING: arch/x86/mm/built-in.o(.text+0x449e): Section mismatch in reference
from the function pat_init() to the function .cpuinit.text:pat_disable()
The function pat_init() references
the function __cpuinit pat_disable().
This is often because pat_init lacks a __cpuinit
annotation or the annotation of pat_disable is wrong.
Non CONFIG_X86_PAT version of pat_disable is static inline, so this version
can be static too (and there are no callers outside of this file).
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
LKML-Reference: <49DFB055.6070405@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: fix section mismatch
In arch/x86/kernel/mpparse.c, smp_reserve_bootmem() has been called
and also refers to a function which is in .init section. Thus causes
the first warning. And check_irq_src() also requires an __init,
because it refers to an .init section.
Signed-off-by: Rakib Mullick <rakib.mullick@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
LKML-Reference: <b9df5fa10904102004g51265d9axc8d07278bfdb6ba0@mail.gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: fix kprobes crash on 32-bit with RAM above 4G
Use phys_addr_t for receiving a physical address argument
instead of unsigned long. This allows fixmap to handle
pages higher than 4GB on x86-32.
Signed-off-by: Masami Hiramatsu <mhiramat@redhat.com>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: systemtap-ml <systemtap@sources.redhat.com>
Cc: Gary Hade <garyhade@us.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <49DE3695.6040800@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
While better than get_user_pages(), the usage of gupf(),
especially the return values and the fact that it can
potentially only partially pin the range, warranted some
documentation.
Signed-off-by: Andy Grover <andy.grover@oracle.com>
Cc: npiggin@suse.de
Cc: akpm@linux-foundation.org
LKML-Reference: <1239320729-3262-1-git-send-email-andy.grover@oracle.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
To simplify level irq migration in the presence of interrupt-remapping,
Suresh used a virtual vector (io-apic pin number) to eliminate io-apic
RTE modification. Level triggered interrupt will appear as an edge to
the local apic cpu but still as level to the IO-APIC. So in addition to
do the local apic EOI, it still needs to do IO-APIC directed EOI to clear
the remote IRR bit in the IO-APIC RTE. Pls refer to Suresh's patch for
more details (commit 0280f7c416).
Now interrupt remapping is decoupled from x2apic, it also needs to do the
directed EOI for apic. Otherwise, apic interrupts won't work correctly.
Signed-off-by: Weidong Han <weidong.han@intel.com>
Cc: iommu@lists.linux-foundation.org
Cc: Weidong Han <weidong.han@intel.com>
Cc: suresh.b.siddha@intel.com
Cc: dwmw2@infradead.org
Cc: allen.m.kay@intel.com
LKML-Reference: <1239355037-22856-1-git-send-email-weidong.han@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Use phys_addr_t for receiving a physical address argument instead of
unsigned long. This allows fixmap to handle pages higher than 4GB on
x86-32.
Signed-off-by: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Acked-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: cpu_debug remove execute permission
x86: smarten /proc/interrupts output for new counters
x86: DMI match for the Dell DXP061 as it needs BIOS reboot
x86: make 64 bit to use default_inquire_remote_apic
x86, setup: un-resequence mode setting for VGA 80x34 and 80x60 modes
x86, intel-iommu: fix X2APIC && !ACPI build failure
H. Peter Anvin