Previously the numa hash code would be confused by holes in the node space
and stop early. This is the first part of the fix for the non boot issue
with empty nodes on Opterons.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Code was refusing good SRATs because about 12K got lost somewhere.
Allow less than 1MB of difference before rejecting it.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The early initialization of cpu_to_node code as it is now only updates the
cpu_to_node array, and does not update cpu_pda()->nodemember. This will
cause numa_node_id() to return 0 on systems where CPU 0 is not on Node 0.
This leads to a kernel panic in slab.c.
I've tested the patch below on a 16 processor x86_64 ES7000-600 server, and
no longer see the panic I saw with the original 2.6.16-rc3.
Signed-off-by: Dan Yeisley <dan.yeisley@unisys.com>
Acked-by: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Checking of the validity of pointers should be consistently done before
dereferencing the pointer.
Signed-Off-By: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
- Check if the processor/memory affinity entries are long enough
according to the ACPI 3.0 spec.
- Ignore memory affinity entries that define a zero length region.
All based on BIOS issues found in the field @)
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add x86-64 specific memory hot-add functions, Kconfig options,
and runtime kernel page table update functions to make
hot-add usable on x86-64 machines. Also, fixup the nefarious
conditional locking and exports pointed out by Andi.
Tested on Intel and IBM x86-64 memory hot-add capable systems.
Signed-off-by: Matt Tolentino <matthew.e.tolentino@intel.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Another try at this.
For 32bit follow the 32bit implementation from Ingo -
mappings are growing down from the end of stack now
and vary randomly by 1GB.
Randomized mappings for 64bit just vary the normal mmap break
by 1TB. I didn't bother implementing full flex mmap for 64bit
because it shouldn't be needed there.
Cc: mingo@elte.hu
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The comparison of the initrd start address against "&_end" is
unnecessary and incorrect. Make it match the x86 code that just
compares the passed-in arguments.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
For not fully explained reasons it broke mem=... on several setups.
Also minor cleanup.
Cc: axboe@suse.de
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Previously they would be only allocated before the kernel text at
1MB. This limited the maximum supported memory to 128GB.
Now allow the e820 allocator to put them everywhere. Try
to put them beyond any DMA zones to avoid filling them up.
This should free some GFP_DMA memory compared to earlier kernels.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Saves about ~18K .text in defconfig
There would be more optimization potential, but that's for later.
Suggestion originally from Bill Irwin.
Fix from Andy Whitcroft.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Patch enables early intialization of cpu_to_node.
apicid_to_node is built by reading the SRAT table, from acpi_numa_init with
ACPI_NUMA and k8_scan_nodes with K8_NUMA.
x86_cpu_to_apicid is built by parsing the ACPI MADT table, from acpi_boot_init.
We combine these two tables and setup cpu_to_node.
Early intialization helps the static per_cpu_areas in getting pages from
correct node.
Change since last release:
Do not initialize early init_cpu_to_node for faking node cases.
Patch tested on TYAN dual core 4P board with K8 only, ACPI_NUMA.
Tested on EM64T NUMA. Also tested with numa=off, numa=fake, and running
a kernel compiled with NUMA on a regular EM64 2 way SMP.
Signed-off-by: Alok N Kataria <alokk@calsoftinc.com>
Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Shai Fultheim <shai@scalex86.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
AK: I hacked Muli's original patch a lot and there were a lot
of changes - all bugs are probably to blame on me now.
There were also some changes in the fall back behaviour
for swiotlb - in particular it doesn't try to use GFP_DMA
now anymore. Also all DMA mapping operations use the
same core dma_alloc_coherent code with proper fallbacks now.
And various other changes and cleanups.
Known problems: iommu=force swiotlb=force together breaks
needs more testing.
This patch cleans up x86_64's DMA mapping dispatching code. Right now
we have three possible IOMMU types: AGP GART, swiotlb and nommu, and
in the future we will also have Xen's x86_64 swiotlb and other HW
IOMMUs for x86_64. In order to support all of them cleanly, this
patch:
- introduces a struct dma_mapping_ops with function pointers for each
of the DMA mapping operations of gart (AMD HW IOMMU), swiotlb
(software IOMMU) and nommu (no IOMMU).
- gets rid of:
if (swiotlb)
return swiotlb_xxx();
- PCI_DMA_BUS_IS_PHYS is now checked against the dma_ops being set
This makes swiotlb faster by avoiding double copying in some cases.
Signed-Off-By: Muli Ben-Yehuda <mulix@mulix.org>
Signed-Off-By: Jon D. Mason <jdmason@us.ibm.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Broken BIOS on Iwill 8way systems reports these and it causes the bootmem
allocator to crash. Add a sanity check if all the PXMs in the
SRAT table cover all memory as reported by e820. If the sanity
check fails the SRAT is rejected and the code will fall back
to discover the NUMA topology using the K8 northbridge registers
when applicable.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix off by one when checking if the machine has enougn memory to need IOMMU
This caused the IOMMUs to be needlessly enabled for mem=4G
Based on a patch from Jon Mason
Signed-off-by: jdmason@us.ibm.com
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Much better to deal with these than with the magic numbers.
And remove the comment describing the bits - kernel source
is no replacement for an architecture manual.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Don't need to do the vmalloc check for the module range because its
PML4 is shared with the kernel text.
Also removed an unnecessary TLB flush.
Pointed out by Jan Beulich
Cc: jbeulich@novell.com
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
When we don't know the node a PCI bus is connected to return -1.
This matches the generic code.
Noticed by Ravikiran G Thirumalai <kiran@scalex86.org>
Cc: Ravikiran G Thirumalai <kiran@scalex86.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
A lot of Opteron BIOS just pass 10 in all SLIT entries (10 is the
normalized unit). This is actually worse than the default heuristic
because it leads to pci_distance not knowing the difference between
local and remote nodes anymore. This messes up some NUMA
heuristics in generic code.
In this case it's better to fall back to the default heuristic
which just does nodea == nodeb ? 10 : 20.
This patch does some basic sanity checking on the SLIT and only accepts
the SLIT when it passes.
Invariants enforced are:
- Node to itself shall be 10
- Any other distance shouldn't be 10
- Distances smaller than 10 are illegal
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Adjust page fault protection error check before considering it to be
a vmalloc synchronization candidate.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This adjusts things so that handlers of the die() notifier will have
sufficient information about the trap currently being handled. It also
adjusts the notify_die() prototype to (again) match that of i386.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
x86-64 specific parts to make the .rodata section read only
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Bug fix required for the .rodata work on x86-64:
when change_page_attr() and friends need to break up a 2Mb page into 4Kb
pages, it always set the NX bit on the PMD, which causes the cpu to consider
the entire 2Mb region to be NX regardless of the actual PTE perms. This is
fine in general, with one big exception: the 2Mb page that covers the last
part of the kernel .text! The fix is to not invent a new permission for the
new PMD entry, but to just inherit the existing one minus the PSE bit.
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Currently, we do not pass the correct start_pfn to e820_hole_size, to
calculate holes. Following patch fixes that.
The bug results in incorrect number of node_present_pages for each pgdat
and causes ugly output in /sys and probably VM inbalances.
Signed-off-by: Alok N Kataria <alokk@calsoftinc.com>
Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Andi Kleen <ak@suse.de>
Sighed-off-by: Shair Fultheim <shai@scalex86.org>
Sighed-off-by: Linus Torvalds <torvalds@osdl.org>
As reported by Keith Mannthey, there are problems in populate_memnodemap()
The bug was that the compute_hash_shift() was returning 31, with incorrect
initialization of memnodemap[]
To correct the bug, we must use (1UL << shift) instead of (1 << shift) to
avoid an integer overflow, and we must check that shift < 64 to avoid an
infinite loop.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
It's illegal because it can sleep.
Use a two step lookup scheme instead. First look up the vm_struct, then
change the direct mapping, then finally unmap it. That's ok because nobody
can change the particular virtual address range as long as the vm_struct is
still in the global list.
Also added some LinuxDoc documentation to iounmap.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix up booting with sparse mem enabled. Otherwise it would just
cause an early PANIC at boot.
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
CONFIG_CHECKING covered some debugging code used in the early times
of the port. But it wasn't even SMP safe for quite some time
and the bugs it checked for seem to be gone.
This patch removes all the code to verify GS at kernel entry. There
haven't been any new bugs in this area for a long time.
Previously it also covered the sysctl for the page fault tracing.
That didn't make much sense because that code was unconditionally
compiled in. I made that a boot option now because it is typically
only useful at boot.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The current x86_64 NUMA memory code is inconsequent when it comes to node
memory ranges. The exact behaviour varies depending on which config option
that is used.
setup_node_bootmem() has start and end as arguments and these are used to
calculate the size of the node like this: (end - start). This is all fine
if end is pointing to the first non-available byte. The problem is that the
current x86_64 code sometimes treats it as the last present byte and sometimes
as the first non-available byte. The result is that some configurations might
lose a page at the end of the range.
This patch tries to fix CONFIG_ACPI_NUMA, CONFIG_K8_NUMA and CONFIG_NUMA_EMU
so they all treat the end variable as the first non-available byte. This is
the same way as the single node code.
The patch is boot tested on dual x86_64 hardware with the above configurations,
but maybe the removed code is needed as some workaround?
Signed-off-by: Magnus Damm <magnus@valinux.co.jp>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Compute the highest possible value for memnode_shift, in order to reduce
footprint of memnodemap[] to the minimum, thus making all users
(phys_to_nid(), kfree()), more cache friendly.
Before the patch :
Node 0 MemBase 0000000000000000 Limit 00000001ffffffff
Node 1 MemBase 0000000200000000 Limit 00000003ffffffff
Using 23 for the hash shift. Max adder is 3ffffffff
After the patch :
Node 0 MemBase 0000000000000000 Limit 00000001ffffffff
Node 1 MemBase 0000000200000000 Limit 00000003ffffffff
Using 33 for the hash shift.
In this case, only 2 bytes of memnodemap[] are used, instead of 2048
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Minor victory on the continuous quest against all stray extern.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Adding __initdata_* to asm-generic/sections.h
Replaces a lot of open coded externs in arch/x86_64/*
I had to change __bss_end to __bss_stop to match the other architectures.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
We should zap the low mappings, as soon as possible, so that we can catch
kernel bugs more effectively. Previously early boot had NULL mapped
and didn't trap on NULL references.
This patch introduces boot_level4_pgt, which will always have low identity
addresses mapped. Druing boot, all the processors will use this as their
level4 pgt. On BP, we will switch to init_level4_pgt as soon as we enter C
code and zap the low mappings as soon as we are done with the usage of
identity low mapped addresses. On AP's we will zap the low mappings as
soon as we jump to C code.
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Not go from the CPU number to an mapping array.
Mode number is often used now in fast paths.
This also adds a generic numa_node_id to all the topology includes
Suggested by Eric Dumazet
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The VM needs to know about lost memory in zones to accurately
balance dirty pages. This patch accounts mem_map in there too,
which fixes a constant errror of a few percent. Also some
other misc mappings and the kernel text itself are accounted
too.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add a new 4GB GFP_DMA32 zone between the GFP_DMA and GFP_NORMAL zones.
As a bit of historical background: when the x86-64 port
was originally designed we had some discussion if we should
use a 16MB DMA zone like i386 or a 4GB DMA zone like IA64 or
both. Both was ruled out at this point because it was in early
2.4 when VM is still quite shakey and had bad troubles even
dealing with one DMA zone. We settled on the 16MB DMA zone mainly
because we worried about older soundcards and the floppy.
But this has always caused problems since then because
device drivers had trouble getting enough DMA able memory. These days
the VM works much better and the wide use of NUMA has proven
it can deal with many zones successfully.
So this patch adds both zones.
This helps drivers who need a lot of memory below 4GB because
their hardware is not accessing more (graphic drivers - proprietary
and free ones, video frame buffer drivers, sound drivers etc.).
Previously they could only use IOMMU+16MB GFP_DMA, which
was not enough memory.
Another common problem is that hardware who has full memory
addressing for >4GB misses it for some control structures in memory
(like transmit rings or other metadata). They tended to allocate memory
in the 16MB GFP_DMA or the IOMMU/swiotlb then using pci_alloc_consistent,
but that can tie up a lot of precious 16MB GFPDMA/IOMMU/swiotlb memory
(even on AMD systems the IOMMU tends to be quite small) especially if you have
many devices. With the new zone pci_alloc_consistent can just put
this stuff into memory below 4GB which works better.
One argument was still if the zone should be 4GB or 2GB. The main
motivation for 2GB would be an unnamed not so unpopular hardware
raid controller (mostly found in older machines from a particular four letter
company) who has a strange 2GB restriction in firmware. But
that one works ok with swiotlb/IOMMU anyways, so it doesn't really
need GFP_DMA32. I chose 4GB to be compatible with IA64 and because
it seems to be the most common restriction.
The new zone is so far added only for x86-64.
For other architectures who don't set up this
new zone nothing changes. Architectures can set a compatibility
define in Kconfig CONFIG_DMA_IS_DMA32 that will define GFP_DMA32
as GFP_DMA. Otherwise it's a nop because on 32bit architectures
it's normally not needed because GFP_NORMAL (=0) is DMA able
enough.
One problem is still that GFP_DMA means different things on different
architectures. e.g. some drivers used to have #ifdef ia64 use GFP_DMA
(trusting it to be 4GB) #elif __x86_64__ (use other hacks like
the swiotlb because 16MB is not enough) ... . This was quite
ugly and is now obsolete.
These should be now converted to use GFP_DMA32 unconditionally. I haven't done
this yet. Or best only use pci_alloc_consistent/dma_alloc_coherent
which will use GFP_DMA32 transparently.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
First step in pushing down the page_table_lock. init_mm.page_table_lock has
been used throughout the architectures (usually for ioremap): not to serialize
kernel address space allocation (that's usually vmlist_lock), but because
pud_alloc,pmd_alloc,pte_alloc_kernel expect caller holds it.
Reverse that: don't lock or unlock init_mm.page_table_lock in any of the
architectures; instead rely on pud_alloc,pmd_alloc,pte_alloc_kernel to take
and drop it when allocating a new one, to check lest a racing task already
did. Similarly no page_table_lock in vmalloc's map_vm_area.
Some temporary ugliness in __pud_alloc and __pmd_alloc: since they also handle
user mms, which are converted only by a later patch, for now they have to lock
differently according to whether or not it's init_mm.
If sources get muddled, there's a danger that an arch source taking
init_mm.page_table_lock will be mixed with common source also taking it (or
neither take it). So break the rules and make another change, which should
break the build for such a mismatch: remove the redundant mm arg from
pte_alloc_kernel (ppc64 scrapped its distinct ioremap_mm in 2.6.13).
Exceptions: arm26 used pte_alloc_kernel on user mm, now pte_alloc_map; ia64
used pte_alloc_map on init_mm, now pte_alloc_kernel; parisc had bad args to
pmd_alloc and pte_alloc_kernel in unused USE_HPPA_IOREMAP code; ppc64
map_io_page forgot to unlock on failure; ppc mmu_mapin_ram and ppc64 im_free
took page_table_lock for no good reason.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Noticed by Terence Ripperda
Undo wrong change in global_flush_tlb. We need to flush the caches in all
cases, not just when pages were reverted. This was a bogus optimization
added earlier, but it was wrong.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The tests Alok carried out on Petr's box confirmed that cpu_to_node[BP] is
not setup early enough by numa_init_array due to the x86_64 changes in
2.6.14-rc*, and unfortunately set wrongly by the work around code in
numa_init_array(). cpu_to_node[0] gets set with 1 early and later gets set
properly to 0 during identify_cpu() when all cpus are brought up, but
confusing the numa slab in the process.
Here is a quick fix for this. The right fix obviously is to have
cpu_to_node[bsp] setup early for numa_init_array(). The following patch
will fix the problem now, and the code can stay on even when
cpu_to_node{BP] gets fixed early correctly.
Thanks to Petr for access to his box.
Signed off by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Alok N Kataria <alokk@calsoftinc.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix the BP node_to_cpumask. 2.6.14-rc* broke the boot cpu bit as the
cpu_to_node(0) is now not setup early enough for numa_init_array.
cpu_to_node[] is setup much later at srat_detect_node on acpi srat based
em64t machines. This seems like a problem on amd machines too, Tested on
em64t though. /sys/devices/system/node/node0/cpumap shows up sanely after
this patch.
Signed off by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Shai Fultheim <shai@scalex86.org>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Rather than blindly re-enabling interrupts in oops_end(), save their state
in oope_begin() and then restore that state.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Andi Kleen