Upcoming 64 bit processors from Centaur can use sysenter.
Signed-off-by: Dave Jones <davej@codemonkey.org.uk>
Signed-off-by: Jesse Ahrens <jahrens@centtech.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
On AMD SMM protected memory is part of the address map, but handled
internally like an MTRR. That leads to large pages getting split
internally which has some performance implications. Check for the
AMD TSEG MSR and split the large page mapping on that area
explicitely if it is part of the direct mapping.
There is also SMM ASEG, but it is in the first 1MB and already covered by
the earlier split first page patch.
Idea for this came from an earlier patch by Andreas Herrmann
On a RevF dual Socket Opteron system kernbench shows a clear
improvement from this:
(together with the earlier patches in this series, especially the
split first 2MB patch)
[lower is better]
no split stddev split stddev delta
Elapsed Time 87.146 (0.727516) 84.296 (1.09098) -3.2%
User Time 274.537 (4.05226) 273.692 (3.34344) -0.3%
System Time 34.907 (0.42492) 34.508 (0.26832) -1.1%
Percent CPU 322.5 (38.3007) 326.5 (44.5128) +1.2%
=> About 3.2% improvement in elapsed time for kernbench.
With GB pages on AMD Fam1h the impact of splitting is much higher of course,
since it would split two full GB pages (together with the first
1MB split patch) instead of two 2MB pages. I could not benchmark
a clear difference in kernbench on gbpages, so I kept it disabled
for that case
That was only limited benchmarking of course, so if someone
was interested in running more tests for the gbpages case
that could be revisited (contributions welcome)
I didn't bother implementing this for 32bit because it is very
unlikely the 32bit lowmem mapping overlaps into the TSEG near 4GB
and the 2MB low split is already handled for both.
[ mingo@elte.hu: do it on gbpages kernels too, there's no clear reason
why it shouldnt help there. ]
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: andreas.herrmann3@amd.com
Cc: mingo@elte.hu
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
When end_pfn is not aligned to 2MB (or 1GB) then the kernel might
map more memory than end_pfn. Account this in max_pfn_mapped.
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: andreas.herrmann3@amd.com
Cc: mingo@elte.hu
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
- Fix the the build breakage when PARAVIRT is defined
but PCI is not
This fixes problem reported at:
http://marc.info/?l=linux-kernel&m=120525966600698&w=2
- Make is_vsmp_box() available even when PARAVIRT is not defined.
This is needed to determine if tsc's are reliable as a time source
even when PARAVIRT is not defined.
- split vsmp_init to use is_vsmp_box() and set_vsmp_pv_ops()
set_vsmp_pv_ops will do nothing if PCI is not enabled in the config.
Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
provide two specialized identify_secondary_cpu() and identify_boot_cpu()
routines for x86_64. Although not strictly needed, they are functionally
correct, and will ease integration with i386
Signed-off-by: Glauber Costa <gcosta@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
store initial_apicid from early identify. it is could be different from
phys_proc_id later.
also print it out in /proc/cpuinfo.
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
for system with apicid lifting, boot cpu apicid will be 4
got:
CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
CPU: L2 Cache: 512K (64 bytes/line)
CPU 0/4 -> Node 0
CPU: Physical Processor ID: 1
CPU: Processor Core ID: 0
so try to offset apicid back before get phys_proc_id with bits shift.
then we can get correct socket ID
also remove remove cpu_data(0) reference.
because cpu_data(0) only be ready after smp_prepare_cpus with the assignment
from boot_cpu_data to current_cpu_data aka cpu_data(0).
and check_bugs()==>identify_cpu(&boot_cpu_data) is quite before than
smp_prepare_cpus. So just use boot_cpu_id instead.
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
early_init_intel() on 64-bit is introduced by
commit 2b16a23538
Author: Andi Kleen <ak@suse.de>
Date: Wed Jan 30 13:32:40 2008 +0100
x86: move X86_FEATURE_CONSTANT_TSC into early cpu feature detection
sets CONSTANT_TSC for intel cpus - but it is already set in init_intel().
don't need to set that two times in early_init_intel() and init_intel().
this patch removes the init_intel() one.
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
e820_resource_resources could use insert_resource instead of request_resource
also move code_resource, data_resource, bss_resource, and crashk_res
out of e820_reserve_resources.
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
x86 /proc/cpuinfo code can be unified.
This is the first step of unification.
Signed-off-by: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
It becomes to early for ioremap, so we use early_ioremap
Signed-off-by: Glauber Costa <gcosta@redhat.com>
Signed-off-by: Ravikiran Thirumalai <kiran@scalemp.com>
Acked-by: Shai Fultheim <shai@scalemp.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Revert
commit f62f1fc9ef
Author: Yinghai Lu <yhlu.kernel@gmail.com>
Date: Fri Mar 7 15:02:50 2008 -0800
x86: reserve dma32 early for gart
The patch has a dependency on bootmem modifications which are not .25
material that late in the -rc cycle. The problem which is addressed by
the patch is limited to machines with 256G and more memory booted with
NUMA disabled. This is not a .25 regression and the audience which is
affected by this problem is very limited, so it's safer to do the
revert than pulling in intrusive bootmem changes right now.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
a system with 256 GB of RAM, when NUMA is disabled crashes the
following way:
Your BIOS doesn't leave a aperture memory hole
Please enable the IOMMU option in the BIOS setup
This costs you 64 MB of RAM
Cannot allocate aperture memory hole (ffff8101c0000000,65536K)
Kernel panic - not syncing: Not enough memory for aperture
Pid: 0, comm: swapper Not tainted 2.6.25-rc4-x86-latest.git #33
Call Trace:
[<ffffffff84037c62>] panic+0xb2/0x190
[<ffffffff840381fc>] ? release_console_sem+0x7c/0x250
[<ffffffff847b1628>] ? __alloc_bootmem_nopanic+0x48/0x90
[<ffffffff847b0ac9>] ? free_bootmem+0x29/0x50
[<ffffffff847ac1f7>] gart_iommu_hole_init+0x5e7/0x680
[<ffffffff847b255b>] ? alloc_large_system_hash+0x16b/0x310
[<ffffffff84506a2f>] ? _etext+0x0/0x1
[<ffffffff847a2e8c>] pci_iommu_alloc+0x1c/0x40
[<ffffffff847ac795>] mem_init+0x45/0x1a0
[<ffffffff8479ff35>] start_kernel+0x295/0x380
[<ffffffff8479f1c2>] _sinittext+0x1c2/0x230
the root cause is : memmap PMD is too big,
[ffffe200e0600000-ffffe200e07fffff] PMD ->ffff81383c000000 on node 0
almost near 4G..., and vmemmap_alloc_block will use up the ram under 4G.
solution will be:
1. make memmap allocation get memory above 4G...
2. reserve some dma32 range early before we try to set up memmap for all.
and release that before pci_iommu_alloc, so gart or swiotlb could get some
range under 4g limit for sure.
the patch is using method 2.
because method1 may need more code to handle SPARSEMEM and SPASEMEM_VMEMMAP
will get
Your BIOS doesn't leave a aperture memory hole
Please enable the IOMMU option in the BIOS setup
This costs you 64 MB of RAM
Mapping aperture over 65536 KB of RAM @ 4000000
Memory: 264245736k/268959744k available (8484k kernel code, 4187464k reserved, 4004k data, 724k init)
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> Diffing dmesg between git7 and git8 doesn't sched any light since
> git8 also removed the printouts of the x86 caps as they were being
> initialised and updated. I'm currently adding those printouts back
> in the hope of seeing where and when the caps get broken.
That turned out to be very illuminating:
--- dmesg-2.6.24-git7 2008-02-24 18:01:25.295851000 +0100
+++ dmesg-2.6.24-git8 2008-02-24 18:01:25.530358000 +0100
...
CPU: After generic identify, caps: 00000003 00000000 00000000 00000000 00000000 00000000 00000000 00000000
CPU: After all inits, caps: 00000003 00000000 00000000 00000000 00000000 00000000 00000000 00000000
+CPU: After applying cleared_cpu_caps, caps: 00000013 00000000 00000000 00000000 00000000 00000000 00000000 00000000
Notice how the TSC cap bit goes from Off to On.
(The first two lines are printout loops from -git7 forward-ported
to -git8, the third line is the same printout loop added just after
the xor-with-cleared_cpu_caps[] loop.)
Here's how the breakage occurs:
1. arch/x86/kernel/tsc_32.c:tsc_init() sees !cpu_has_tsc,
so bails and calls setup_clear_cpu_cap(X86_FEATURE_TSC).
2. include/asm-x86/cpufeature.h:setup_clear_cpu_cap(bit) clears
the bit in boot_cpu_data and sets it in cleared_cpu_caps
3. arch/x86/kernel/cpu/common.c:identify_cpu() XORs all caps
in with cleared_cpu_caps
HOWEVER, at this point c->x86_capability correctly has TSC
Off, cleared_cpu_caps has TSC On, so the XOR incorrectly
sets TSC to On in c->x86_capability, with disastrous results.
The real bug is that clearing bits with XOR only works if the
bits are known to be 1 prior to the XOR, and that's not true here.
A simple fix is to convert the XOR to AND-NOT instead. The following
patch does that, and allows my 486 to boot 2.6.25-rc kernels again.
[ mingo@elte.hu: fixed a similar bug in setup_64.c as well. ]
The breakage was introduced via commit 7d851c8d3d.
Signed-off-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
srat_detect_node() is only used by __cpuinit init_intel().
So the trivial fix is to annotate srat_detect_node() with __cpuinit.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
nearby_node() were only used by __cpuinit amd_detect_cmp()
So annotating nearby_node() __cpuinit was the trivial fix.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Use the BOOTMEM_EXCLUSIVE, introduced in the previous patch, to avoid
conflicts while reserving the memory for the kdump capture kernel
(crashkernel=).
Signed-off-by: Bernhard Walle <bwalle@suse.de>
Cc: <linux-arch@vger.kernel.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patchset adds a flags variable to reserve_bootmem() and uses the
BOOTMEM_EXCLUSIVE flag in crashkernel reservation code to detect collisions
between crashkernel area and already used memory.
This patch:
Change the reserve_bootmem() function to accept a new flag BOOTMEM_EXCLUSIVE.
If that flag is set, the function returns with -EBUSY if the memory already
has been reserved in the past. This is to avoid conflicts.
Because that code runs before SMP initialisation, there's no race condition
inside reserve_bootmem_core().
[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: fix powerpc build]
Signed-off-by: Bernhard Walle <bwalle@suse.de>
Cc: <linux-arch@vger.kernel.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Move the CPU feature string names to a separate file (common to 32
and 64 bits); additionally, make <asm/cpufeature.h> includable by host
code in preparation for including the CPU feature strings in the boot
code.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This patch adds a new configuration option, which adds support for a new
early_param which gets checked in arch/x86/kernel/setup_{32,64}.c:setup_arch()
to decide wether OHCI-1394 FireWire controllers should be initialized and
enabled for physical DMA access to allow remote debugging of early problems
like issues ACPI or other subsystems which are executed very early.
If the config option is not enabled, no code is changed, and if the boot
paramenter is not given, no new code is executed, and independent of that,
all new code is freed after boot, so the config option can be even enabled
in standard, non-debug kernels.
With specialized tools, it is then possible to get debugging information
from machines which have no serial ports (notebooks) such as the printk
buffer contents, or any data which can be referenced from global pointers,
if it is stored below the 4GB limit and even memory dumps of of the physical
RAM region below the 4GB limit can be taken without any cooperation from the
CPU of the host, so the machine can be crashed early, it does not matter.
In the extreme, even kernel debuggers can be accessed in this way. I wrote
a small kgdb module and an accompanying gdb stub for FireWire which allows
to gdb to talk to kgdb using remote remory reads and writes over FireWire.
An version of the gdb stub fore FireWire is able to read all global data
from a system which is running a a normal kernel without any kernel debugger,
without any interruption or support of the system's CPU. That way, e.g. the
task struct and so on can be read and even manipulated when the physical DMA
access is granted.
A HOWTO is included in this patch, in Documentation/debugging-via-ohci1394.txt
and I've put a copy online at
ftp://ftp.suse.de/private/bk/firewire/docs/debugging-via-ohci1394.txt
It also has links to all the tools which are available to make use of it
another copy of it is online at:
ftp://ftp.suse.de/private/bk/firewire/kernel/ohci1394_dma_early-v2.diff
Signed-Off-By: Bernhard Kaindl <bk@suse.de>
Tested-By: Thomas Renninger <trenn@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This patch fixes some bugs of EFI memory handing code.
- On x86_64, it is possible that EFI memory map can not be mapped via
identity map, so efi_map_memmap is removed, just use early_ioremap.
- On i386, the EFI memory map mapping take effect cross paging_init,
so it is not necessary to use efi_map_memmap.
- EFI memory map is unmapped in efi_enter_virtual_mode to avoid
early_ioremap leak.
Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
in init/main.c boot_cpu_init() already does that before setup_arch
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Fix the following warning:
WARNING: arch/x86/kernel/built-in.o(.cpuinit.text+0x7a3): Section mismatch: reference to .init.text:amd_detect_cmp in 'init_amd'
The function amd_detect_cmp were annotated __init and
was only used from init_amd() which are annotated __cpuinit.
Annotate amd_detect_cmp() with _cpuinit to fix it.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
No one uses struct cpu_model_info on x86_64 now.
Signed-off-by: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Change static bios_cpu_apicid array to a per_cpu data variable.
This includes using a static array used during initialization
similar to the way x86_cpu_to_apicid[] is handled.
There is one early use of bios_cpu_apicid in apic_is_clustered_box().
The other reference in cpu_present_to_apicid() is called after
smp_set_apicids() has setup the percpu version of bios_cpu_apicid.
Signed-off-by: Mike Travis <travis@sgi.com>
Reviewed-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Add a generic option to clear any cpuid bit. I added it because it was
very easy to add with the new generic cpuid disable bitmap and perhaps
it will be useful in the future.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
To disable CLFLUSH usage, especially in change_page_attr().
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
There are already various options to disable specific cpuid bits
on the command line. They all use their own variable. Add a generic
mask to make this easier in the future.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Move the select_idle_routine() call to after the detect_ht() call at
identify_cpu() on 64-bit.
This change is for printing the polling idle and HT enabled warning
message properly.
Signed-off-by: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Andi's patch
"
x86: move X86_FEATURE_CONSTANT_TSC into early cpu feature detection
Need this in the next patch in time_init and that happens early.
This includes a minor fix on i386 where early_intel_workarounds()
[which is now called early_init_intel] really executes early as
the comments say.
"
calling early_init_amd in early_identify_cpu and identify_cpu two times.
this patch remove the one in identify_cpu
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
On some machines, buggy BIOSes don't properly setup WB MTRRs to cover all
available RAM, meaning the last few megs (or even gigs) of memory will be
marked uncached. Since Linux tends to allocate from high memory addresses
first, this causes the machine to be unusably slow as soon as the kernel
starts really using memory (i.e. right around init time).
This patch works around the problem by scanning the MTRRs at boot and
figuring out whether the current end_pfn value (setup by early e820 code)
goes beyond the highest WB MTRR range, and if so, trimming it to match. A
fairly obnoxious KERN_WARNING is printed too, letting the user know that
not all of their memory is available due to a likely BIOS bug.
Something similar could be done on i386 if needed, but the boot ordering
would be slightly different, since the MTRR code on i386 depends on the
boot_cpu_data structure being setup.
This patch fixes a bug in the last patch that caused the code to run on
non-Intel machines (AMD machines apparently don't need it and it's untested
on other non-Intel machines, so best keep it off).
Further enhancements and fixes from:
Yinghai Lu <Yinghai.Lu@Sun.COM>
Andi Kleen <ak@suse.de>
Signed-off-by: Jesse Barnes <jesse.barnes@intel.com>
Tested-by: Justin Piszcz <jpiszcz@lucidpixels.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
On x86-64 there are several memory allocations before bootmem. To avoid
them stomping on each other they used to be all hard coded in bad_area().
Replace this with an array that is filled as needed.
This cleans up the code considerably and allows to expand its use.
Cc: peterz@infradead.org
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Previously there was a AMD specific quirk to handle the case of
AMD Fam10h MWAIT not supporting any C states. But it turns out
that CPUID already has ways to detectly detect that without
using special quirks.
The new code simply checks if MWAIT supports at least C1 and doesn't
use it if it doesn't. No more vendor specific code.
Note this is does not simply clear MWAIT because MWAIT can be still
useful even without C states.
Credit goes to Ben Serebrin for pointing out the (nearly) obvious.
Cc: "Andreas Herrmann" <andreas.herrmann3@amd.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Change static bios_cpu_apicid array to a per_cpu data variable.
This includes using a static array used during initialization
similar to the way x86_cpu_to_apicid[] is handled.
There is one early use of bios_cpu_apicid in apic_is_clustered_box().
The other reference in cpu_present_to_apicid() is called after
smp_set_apicids() has setup the percpu version of bios_cpu_apicid.
[ mingo@elte.hu: build fix ]
Signed-off-by: Mike Travis <travis@sgi.com>
Reviewed-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Change the following static arrays sized by NR_CPUS to
per_cpu data variables:
char cpu_to_node_map[NR_CPUS];
Signed-off-by: Mike Travis <travis@sgi.com>
Reviewed-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Clean up references to x86_cpu_to_apicid. Removes extraneous
comments and standardizes on "x86_*_early_ptr" for the early
kernel init references.
Signed-off-by: Mike Travis <travis@sgi.com>
Reviewed-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
For K8 system: 4G RAM with memory hole remapping enabled, or more than
4G RAM installed.
when try to use kexec second kernel, and the first doesn't include
gart_shutdown. the second kernel could have different aper position than
the first kernel. and second kernel could use that hole as RAM that is
still used by GART set by the first kernel. esp. when try to kexec
2.6.24 with sparse mem enable from previous kernel (from RHEL 5 or SLES
10). the new kernel will use aper by GART (set by first kernel) for
vmemmap. and after new kernel setting one new GART. the position will be
real RAM. the _mapcount set is lost.
Bad page state in process 'swapper'
page:ffffe2000e600020 flags:0x0000000000000000 mapping:0000000000000000 mapcount:1 count:0
Trying to fix it up, but a reboot is needed
Backtrace:
Pid: 0, comm: swapper Not tainted 2.6.24-rc7-smp-gcdf71a10-dirty #13
Call Trace:
[<ffffffff8026401f>] bad_page+0x63/0x8d
[<ffffffff80264169>] __free_pages_ok+0x7c/0x2a5
[<ffffffff80ba75d1>] free_all_bootmem_core+0xd0/0x198
[<ffffffff80ba3a42>] numa_free_all_bootmem+0x3b/0x76
[<ffffffff80ba3461>] mem_init+0x3b/0x152
[<ffffffff80b959d3>] start_kernel+0x236/0x2c2
[<ffffffff80b9511a>] _sinittext+0x11a/0x121
and
[ffffe2000e600000-ffffe2000e7fffff] PMD ->ffff81001c200000 on node 0
phys addr is : 0x1c200000
RHEL 5.1 kernel -53 said:
PCI-DMA: aperture base @ 1c000000 size 65536 KB
new kernel said:
Mapping aperture over 65536 KB of RAM @ 3c000000
So could try to disable that GART if possible.
According to Ingo
> hm, i'm wondering, instead of modifying the GART, why dont we simply
> _detect_ whatever GART settings we have inherited, and propagate that
> into our e820 maps? I.e. if there's inconsistency, then punch that out
> from the memory maps and just dont use that memory.
>
> that way it would not matter whether the GART settings came from a [old
> or crashing] Linux kernel that has not called gart_iommu_shutdown(), or
> whether it's a BIOS that has set up an aperture hole inconsistent with
> the memory map it passed. (or the memory map we _think_ i tried to pass
> us)
>
> it would also be more robust to only read and do a memory map quirk
> based on that, than actively trying to change the GART so early in the
> bootup. Later on we have to re-enable the GART _anyway_ and have to
> punch a hole for it.
>
> and as a bonus, we would have shored up our defenses against crappy
> BIOSes as well.
add e820 modification for gart inconsistent setting.
gart_fix_e820=off could be used to disable e820 fix.
Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
