Optimize the critical zonelist scanning for free pages in the kernel memory
allocator by caching the zones that were found to be full recently, and
skipping them.
Remembers the zones in a zonelist that were short of free memory in the
last second. And it stashes a zone-to-node table in the zonelist struct,
to optimize that conversion (minimize its cache footprint.)
Recent changes:
This differs in a significant way from a similar patch that I
posted a week ago. Now, instead of having a nodemask_t of
recently full nodes, I have a bitmask of recently full zones.
This solves a problem that last weeks patch had, which on
systems with multiple zones per node (such as DMA zone) would
take seeing any of these zones full as meaning that all zones
on that node were full.
Also I changed names - from "zonelist faster" to "zonelist cache",
as that seemed to better convey what we're doing here - caching
some of the key zonelist state (for faster access.)
See below for some performance benchmark results. After all that
discussion with David on why I didn't need them, I went and got
some ;). I wanted to verify that I had not hurt the normal case
of memory allocation noticeably. At least for my one little
microbenchmark, I found (1) the normal case wasn't affected, and
(2) workloads that forced scanning across multiple nodes for
memory improved up to 10% fewer System CPU cycles and lower
elapsed clock time ('sys' and 'real'). Good. See details, below.
I didn't have the logic in get_page_from_freelist() for various
full nodes and zone reclaim failures correct. That should be
fixed up now - notice the new goto labels zonelist_scan,
this_zone_full, and try_next_zone, in get_page_from_freelist().
There are two reasons I persued this alternative, over some earlier
proposals that would have focused on optimizing the fake numa
emulation case by caching the last useful zone:
1) Contrary to what I said before, we (SGI, on large ia64 sn2 systems)
have seen real customer loads where the cost to scan the zonelist
was a problem, due to many nodes being full of memory before
we got to a node we could use. Or at least, I think we have.
This was related to me by another engineer, based on experiences
from some time past. So this is not guaranteed. Most likely, though.
The following approach should help such real numa systems just as
much as it helps fake numa systems, or any combination thereof.
2) The effort to distinguish fake from real numa, using node_distance,
so that we could cache a fake numa node and optimize choosing
it over equivalent distance fake nodes, while continuing to
properly scan all real nodes in distance order, was going to
require a nasty blob of zonelist and node distance munging.
The following approach has no new dependency on node distances or
zone sorting.
See comment in the patch below for a description of what it actually does.
Technical details of note (or controversy):
- See the use of "zlc_active" and "did_zlc_setup" below, to delay
adding any work for this new mechanism until we've looked at the
first zone in zonelist. I figured the odds of the first zone
having the memory we needed were high enough that we should just
look there, first, then get fancy only if we need to keep looking.
- Some odd hackery was needed to add items to struct zonelist, while
not tripping up the custom zonelists built by the mm/mempolicy.c
code for MPOL_BIND. My usual wordy comments below explain this.
Search for "MPOL_BIND".
- Some per-node data in the struct zonelist is now modified frequently,
with no locking. Multiple CPU cores on a node could hit and mangle
this data. The theory is that this is just performance hint data,
and the memory allocator will work just fine despite any such mangling.
The fields at risk are the struct 'zonelist_cache' fields 'fullzones'
(a bitmask) and 'last_full_zap' (unsigned long jiffies). It should
all be self correcting after at most a one second delay.
- This still does a linear scan of the same lengths as before. All
I've optimized is making the scan faster, not algorithmically
shorter. It is now able to scan a compact array of 'unsigned
short' in the case of many full nodes, so one cache line should
cover quite a few nodes, rather than each node hitting another
one or two new and distinct cache lines.
- If both Andi and Nick don't find this too complicated, I will be
(pleasantly) flabbergasted.
- I removed the comment claiming we only use one cachline's worth of
zonelist. We seem, at least in the fake numa case, to have put the
lie to that claim.
- I pay no attention to the various watermarks and such in this performance
hint. A node could be marked full for one watermark, and then skipped
over when searching for a page using a different watermark. I think
that's actually quite ok, as it will tend to slightly increase the
spreading of memory over other nodes, away from a memory stressed node.
===============
Performance - some benchmark results and analysis:
This benchmark runs a memory hog program that uses multiple
threads to touch alot of memory as quickly as it can.
Multiple runs were made, touching 12, 38, 64 or 90 GBytes out of
the total 96 GBytes on the system, and using 1, 19, 37, or 55
threads (on a 56 CPU system.) System, user and real (elapsed)
timings were recorded for each run, shown in units of seconds,
in the table below.
Two kernels were tested - 2.6.18-mm3 and the same kernel with
this zonelist caching patch added. The table also shows the
percentage improvement the zonelist caching sys time is over
(lower than) the stock *-mm kernel.
number 2.6.18-mm3 zonelist-cache delta (< 0 good) percent
GBs N ------------ -------------- ---------------- systime
mem threads sys user real sys user real sys user real better
12 1 153 24 177 151 24 176 -2 0 -1 1%
12 19 99 22 8 99 22 8 0 0 0 0%
12 37 111 25 6 112 25 6 1 0 0 -0%
12 55 115 25 5 110 23 5 -5 -2 0 4%
38 1 502 74 576 497 73 570 -5 -1 -6 0%
38 19 426 78 48 373 76 39 -53 -2 -9 12%
38 37 544 83 36 547 82 36 3 -1 0 -0%
38 55 501 77 23 511 80 24 10 3 1 -1%
64 1 917 125 1042 890 124 1014 -27 -1 -28 2%
64 19 1118 138 119 965 141 103 -153 3 -16 13%
64 37 1202 151 94 1136 150 81 -66 -1 -13 5%
64 55 1118 141 61 1072 140 58 -46 -1 -3 4%
90 1 1342 177 1519 1275 174 1450 -67 -3 -69 4%
90 19 2392 199 192 2116 189 176 -276 -10 -16 11%
90 37 3313 238 175 2972 225 145 -341 -13 -30 10%
90 55 1948 210 104 1843 213 100 -105 3 -4 5%
Notes:
1) This test ran a memory hog program that started a specified number N of
threads, and had each thread allocate and touch 1/N'th of
the total memory to be used in the test run in a single loop,
writing a constant word to memory, one store every 4096 bytes.
Watching this test during some earlier trial runs, I would see
each of these threads sit down on one CPU and stay there, for
the remainder of the pass, a different CPU for each thread.
2) The 'real' column is not comparable to the 'sys' or 'user' columns.
The 'real' column is seconds wall clock time elapsed, from beginning
to end of that test pass. The 'sys' and 'user' columns are total
CPU seconds spent on that test pass. For a 19 thread test run,
for example, the sum of 'sys' and 'user' could be up to 19 times the
number of 'real' elapsed wall clock seconds.
3) Tests were run on a fresh, single-user boot, to minimize the amount
of memory already in use at the start of the test, and to minimize
the amount of background activity that might interfere.
4) Tests were done on a 56 CPU, 28 Node system with 96 GBytes of RAM.
5) Notice that the 'real' time gets large for the single thread runs, even
though the measured 'sys' and 'user' times are modest. I'm not sure what
that means - probably something to do with it being slow for one thread to
be accessing memory along ways away. Perhaps the fake numa system, running
ostensibly the same workload, would not show this substantial degradation
of 'real' time for one thread on many nodes -- lets hope not.
6) The high thread count passes (one thread per CPU - on 55 of 56 CPUs)
ran quite efficiently, as one might expect. Each pair of threads needed
to allocate and touch the memory on the node the two threads shared, a
pleasantly parallizable workload.
7) The intermediate thread count passes, when asking for alot of memory forcing
them to go to a few neighboring nodes, improved the most with this zonelist
caching patch.
Conclusions:
* This zonelist cache patch probably makes little difference one way or the
other for most workloads on real numa hardware, if those workloads avoid
heavy off node allocations.
* For memory intensive workloads requiring substantial off-node allocations
on real numa hardware, this patch improves both kernel and elapsed timings
up to ten per-cent.
* For fake numa systems, I'm optimistic, but will have to leave that up to
Rohit Seth to actually test (once I get him a 2.6.18 backport.)
Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: Rohit Seth <rohitseth@google.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: David Rientjes <rientjes@cs.washington.edu>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The zone table is mostly not needed. If we have a node in the page flags
then we can get to the zone via NODE_DATA() which is much more likely to be
already in the cpu cache.
In case of SMP and UP NODE_DATA() is a constant pointer which allows us to
access an exact replica of zonetable in the node_zones field. In all of
the above cases there will be no need at all for the zone table.
The only remaining case is if in a NUMA system the node numbers do not fit
into the page flags. In that case we make sparse generate a table that
maps sections to nodes and use that table to to figure out the node number.
This table is sized to fit in a single cache line for the known 32 bit
NUMA platform which makes it very likely that the information can be
obtained without a cache miss.
For sparsemem the zone table seems to be have been fairly large based on
the maximum possible number of sections and the number of zones per node.
There is some memory saving by removing zone_table. The main benefit is to
reduce the cache foootprint of the VM from the frequent lookups of zones.
Plus it simplifies the page allocator.
[akpm@osdl.org: build fix]
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
With CONFIG_SMP=n:
drivers/input/ff-memless.c:384: warning: implicit declaration of function 'local_bh_disable'
drivers/input/ff-memless.c:393: warning: implicit declaration of function 'local_bh_enable'
Really linux/spinlock.h should include linux/interrupt.h. But interrupt.h
includes sched.h which will need spinlock.h.
So the patch breaks the _bh declarations out into a separate header and
includes it in both interrupt.h and spinlock.h.
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Cc: Andi Kleen <ak@suse.de>
Cc: <stable@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
[MIPS] Import updates from i386's i8259.c
[MIPS] *-berr: Header inclusions for DEC bus error handlers
[MIPS] Compile __do_IRQ() when really needed
[MIPS] genirq: use name instead of typename
[MIPS] Do not use handle_level_irq for ioasic_dma_irq_type.
[MIPS] pte_offset(dir,addr): parenthesis fix
Any code that relies on the volatile would be a bug waiting to happen
anyway.
Don't encourage people to think that putting 'volatile' on data
structures somehow fixes problems. We should always use proper locking
(and other serialization) techniques.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is a resubmission of patches originally created by Ingo Molnar.
The link below is the initial (?) posting of the patch.
http://marc.theaimsgroup.com/?l=linux-kernel&m=115217423929806&w=2
Remove 'volatile' from spinlock_types as it causes GCC to generate bad
code (see link) and locking should be used on kernel data.
Signed-off-by: Art Haas <ahaas@airmail.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch adds missing parenthesis around 'dir' argument in pte_offset()
macro definition.
It also removes an extra space in the definition of pte_offset_kernel()
macro.
Signed-off-by: Franck Bui-Huu <fbuihuu@gmail.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
* master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh-2.6: (43 commits)
sh: sh775x/titan fixes for irq header changes.
sh: update r7780rp defconfig.
sh: compile fixes for header cleanup.
sh: Fixup pte_mkhuge() build failure.
sh: set KBUILD_IMAGE to something sensible.
sh: show held locks in stack trace with lockdep.
sh: platform_pata support for R7780RP
sh: stacktrace/lockdep/irqflags tracing support.
sh: Fixup movli.l/movco.l atomic ops for gcc4.
sh: dyntick infrastructure.
sh: Clock framework tidying.
sh: Turn off IRQs around get_timer_offset() calls.
sh: Get the PGD right in oops case with 64-bit PTEs.
sh: Fix store queue bitmap end.
sh: More flexible + SH7780 earlyprintk SCIF support.
sh: Fixup various PAGE_SIZE == 4096 assumptions.
sh: Fixup 4K irq stacks.
sh: dma-api channel capability extensions.
sh: Drop name overload in dma-sh.
sh: Make dma-isa depend on ISA_DMA_API.
...
* git://git.infradead.org/users/dhowells/workq-2.6:
Actually update the fixed up compile failures.
WorkQueue: Fix up arch-specific work items where possible
WorkStruct: make allyesconfig
WorkStruct: Pass the work_struct pointer instead of context data
WorkStruct: Merge the pending bit into the wq_data pointer
WorkStruct: Typedef the work function prototype
WorkStruct: Separate delayable and non-delayable events.
The first patch is to the 2.6 kernel include file (for m68knommu), to get
rid of the conditional definitions, otherwise the structures have different
sizes depending on whether there's an FPU or not.
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add a null definition for irq_canonicalize(). It is used in the gerneric
serial subsystem code, can't compile without it.
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This adds support for RTCs (through genrtc) for M68KNOMMU.
Board-specific code will have to link the appropriate RTC driver to the
mach_hwclk callback, at minimum.
Signed-off-by: Gavin Lambert <gavinl@compacsort.com>
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Conflicts:
drivers/pcmcia/ds.c
Fix up merge failures with Linus's head and fix new compile failures.
Signed-Off-By: David Howells <dhowells@redhat.com>
The following moves the creation of IPR interupts into setup-7750.c
and updates a few other things to make it all work after the "Drop
CPU subtype IRQ headers" commit. It boots and runs fine on my titan
board.
- adds an ipr_idx to the ipr_data and uses a function in the subtype
code to calculate the address of the IPR registers
- adds a function to enable individual interrupt mode for externals
in the subtype code and calls that from the titan board code
instead of doing it directly.
- I changed the shift in the ipr_data to be the actual # of bits to
shift, instead of the numnber / 4 - made it easier to match with
the manual.
Signed-off-by: Jamie Lenehan <lenehan@twibble.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
When hugetlbpage support isn't enabled, this can be bogus.
Wrap it back in _PAGE_FLAGS_HARD to avoid changes to the
base PTE when not aiming for larger sizes.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
gcc4 gets a bit pissy about the outputs:
include/asm/atomic.h: In function 'atomic_add':
include/asm/atomic.h:37: error: invalid lvalue in asm statement
include/asm/atomic.h:30: error: invalid lvalue in asm output 1
...
this ended up being a thinko anyways, so just fix it up.
Verified for proper behaviour with the older toolchains, too.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
This adds basic NO_IDLE_HZ support to the SH timer API so timers
are able to wire it up. Taken from the ARM version, as it fit in
to our API with very few changes needed.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
This syncs up the SH clock framework with the linux/clk.h API,
for which there were only some minor changes required, namely
the clk_get() dev_id and subsequent callsites.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
There were a number of places that made evil PAGE_SIZE == 4k
assumptions that ended up breaking when trying to play with
8k and 64k page sizes, this fixes those up.
The most significant change is the way we load THREAD_SIZE,
previously this was done via:
mov #(THREAD_SIZE >> 8), reg
shll8 reg
to avoid a memory access and allow the immediate load. With
a 64k PAGE_SIZE, we're out of range for the immediate load
size without resorting to special instructions available in
later ISAs (movi20s and so on). The "workaround" for this is
to bump up the shift to 10 and insert a shll2, which gives a
bit more flexibility while still being much cheaper than a
memory access.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
This extends the SH DMA API for allowing handling of DMA
channels based off of their respective capabilities.
A couple of functions are added to the existing API,
the core bits are register_chan_caps() for registering
channel capabilities, and request_dma_bycap() for fetching
a channel dynamically based off of a capability set.
Signed-off-by: Mark Glaisher <mark.glaisher@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Two of the fields in /proc/[number]/stat are documented in
proc(5) as:
kstkesp %lu
The current value of esp (stack pointer), as
found in the kernel stack page for the process.
kstkeip %lu
The current EIP (instruction pointer).
The SH currently prints the the last SP and PC of the process
inside the kernel, while most other archs use the last user
space values.
This patch modifes the SH to display the user space values.
Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Handle simple TLB miss faults which can be resolved completely
from the page table in assembler.
Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
This adds support for a generic push switch framework. Adaptable for
various switches, including GPIO switches and the push switches commonly
found on Renesas debug boards.
This allows switch states to be trivially reported through sysfs.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Remove extra bits from the pmd structure and store a kernel logical
address rather than a physical address. This allows it to be directly
dereferenced. Another piece of wierdness inherited from x86.
Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Add TTB accessor functions and give it a sensible default
value. We will use this later for optimizing the fault
path.
Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Remove the previous saving of fault codes into the thread_struct
as they are never used, and appeared to be inherited from x86.
Signed-off-by: Stuart Menefy <stuart.menefy@st.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
This adds some preliminary support for the SH-X2 MMU, used by
newer SH-4A parts (particularly SH7785).
This MMU implements a 'compat' mode with SH-X MMUs and an
'extended' mode for SH-X2 extended features. Extended features
include additional page sizes (8kB, 4MB, 64MB), as well as the
addition of page execute permissions.
The extended mode attributes are placed in a second data array,
which requires us to switch to 64-bit PTEs when in X2 mode.
With the addition of the exec perms, we also overhaul the mmap
prots somewhat, now that it's possible to handle them more
intelligently.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
This drops the various IRQ headers that were floating around
and primarily providing hardcoded IRQ definitions for the
various CPU subtypes. This quickly got to be an unmaintainable
mess, made even more evident by the subtle breakage introduced
by the SH-2 and SH-2A changes.
Now that subtypes are able to register IRQ maps directly, just
rip all of the headers out.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Fix up arch-specific work items where possible to use the new work_struct and
delayed_work structs.
Three places that enqueue bits of their stack and then return have been marked
with #error as this is not permitted.
Signed-Off-By: David Howells <dhowells@redhat.com>
Conflicts:
drivers/ata/libata-scsi.c
include/linux/libata.h
Futher merge of Linus's head and compilation fixups.
Signed-Off-By: David Howells <dhowells@redhat.com>
Conflicts:
drivers/infiniband/core/iwcm.c
drivers/net/chelsio/cxgb2.c
drivers/net/wireless/bcm43xx/bcm43xx_main.c
drivers/net/wireless/prism54/islpci_eth.c
drivers/usb/core/hub.h
drivers/usb/input/hid-core.c
net/core/netpoll.c
Fix up merge failures with Linus's head and fix new compilation failures.
Signed-Off-By: David Howells <dhowells@redhat.com>
CONFIG_LBD and CONFIG_LSF are spread into asm/types.h for no particularly
good reason.
Centralising the definition in linux/types.h means that arch maintainers
don't need to bother adding it, as well as fixing the problem with
x86-64 users being asked to make a decision that has absolutely no
effect.
The H8/300 porters seem particularly confused since I'm not aware of any
microcontrollers that need to support 2TB filesystems.
Signed-off-by: Matthew Wilcox <matthew@wil.cx>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* master.kernel.org:/pub/scm/linux/kernel/git/paulus/powerpc: (194 commits)
[POWERPC] Add missing EXPORTS for mpc52xx support
[POWERPC] Remove obsolete PPC_52xx and update CLASSIC32 comment
[POWERPC] ps3: add a default zImage target
[POWERPC] Add of_platform_bus support to mpc52xx psc uart driver
[POWERPC] typo fix and whitespace cleanup on mpc52xx-uart driver
[POWERPC] Fix debug printks for 32-bit resources in the PCI code
[POWERPC] Replace kmalloc+memset with kzalloc
[POWERPC] Linkstation / kurobox support
[POWERPC] Add the e300c3 core to the CPU table.
[POWERPC] ppc: m48t35 add missing bracket
[POWERPC] iSeries: don't build head_64.o unnecessarily
[POWERPC] iSeries: stop dt_mod.o being rebuilt unnecessarily
[POWERPC] Fix cputable.h for combined build
[POWERPC] Allow CONFIG_BOOTX_TEXT on iSeries
[POWERPC] Allow xmon to build on legacy iSeries
[POWERPC] Change ppc64_defconfig to use AUTOFS_V4 not V3
[POWERPC] Tell firmware we can handle POWER6 compatible mode
[POWERPC] Clean images in arch/powerpc/boot
[POWERPC] Fix OF pci flags parsing
[POWERPC] defconfig for lite5200 board
...
s->functions needs to be initialized earlier, for the "let's see
how high it increases" approach means that pcmcia_request_irq()
(which makes use of this value) is confused, and might request
an exclusive IRQ first even though it is not supposed to.
Also, a CIS override autoloaded using the firmware loader may
allow for the use of more or less functions in a multifunction
card. Therefore, we may need to schedule a call to add this
second function later on, or simply remove the other function
(it's always the first -valid- function which reaches this
codepath).
Many thanks to Fabrice Bellet for debugging and testing patches.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
For the definition of atomic64_t atomic.h was relying on <asm/types.h>
having been included previously. Before changeset
d89d8e0637a5e4e0a12e90c4bc934d0d4c335239 this was happening as a
side effect of including <linux/spinlock.h>.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev: (82 commits)
[PATCH] pata_ali: small fixes
[PATCH] pata_via: VIA 8251 bridged systems are now out and about
[PATCH] trivial piix: swap bogus dot for comma space
[PATCH] sata_promise: PHYMODE4 fixup
[PATCH] libata: always use polling IDENTIFY
[libata] pata_cs5535: fix build
[PATCH] ahci: do not powerdown during initialization
[PATCH] libata: prepare ata_sg_clean() for invocation from EH
[PATCH] libata: separate out rw ATA taskfile building into ata_build_rw_tf()
[PATCH] libata: implement ata_exec_internal_sg()
[PATCH] libata: make sure IRQ is cleared after ata_bmdma_freeze()
[PATCH] libata: move BMDMA host status recording from EH to interrupt handler
[PATCH] libata: make sure sdev doesn't go away while rescanning
[PATCH] libata: don't request sense if the port is frozen
[PATCH] libata: fix READ CAPACITY simulation
[PATCH] libata: implement ATA_FLAG_SETXFER_POLLING and use it in pata_via, take #2
[PATCH] libata: set IRQF_SHARED for legacy PCI IDE IRQs
[PATCH] libata: remove unused HSM_ST_UNKNOWN
[PATCH] libata: kill unnecessary sht->max_sectors initializations
[PATCH] libata: add missing sht->slave_destroy
...
This was exposed by Al's recent header file dependency reduction
patches..
Cc: Al Viro <viro@ftp.linux.org.uk>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6: (34 commits)
[S390] Don't use small stacks when lockdep is used.
[S390] cio: Use device_reprobe() instead of bus_rescan_devices().
[S390] cio: Retry internal operations after vary off.
[S390] cio: Use path verification for last path gone after vary off.
[S390] non-unique constant/macro identifiers.
[S390] Memory detection fixes.
[S390] cio: Make ccw_dev_id_is_equal() more robust.
[S390] Convert extmem spin_lock into a mutex.
[S390] set KBUILD_IMAGE.
[S390] lockdep: show held locks when showing a stackdump
[S390] Add dynamic size check for usercopy functions.
[S390] Use diag260 for memory size detection.
[S390] pfault code cleanup.
[S390] Cleanup memory_chunk array usage.
[S390] Misaligned wait PSW at memory detection.
[S390] cpu shutdown rework
[S390] cpcmd <-> __cpcmd calling issues
[S390] Bad kexec control page allocation.
[S390] Reset infrastructure for re-IPL.
[S390] Some documentation typos.
...
Make the m68knommu DMA handling consistent with other architectures.
Compile problems pointed out by Adrian Bunk <bunk@stusta.de>
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Here is a small patch to automatically detect the DRAM size on m520x.
It was generated against 2.6.17-uc0, and tested on an Intec 5208 dev board.
(This part of the patch if the memory register defines for the 520x
ColdFire CPU family - Greg).
Signed-off-by: Michael Broughton <mbobowik@telusplanet.net>
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
VMALLOC_END on 31bit should be 0x8000000UL instead of 0x7fffffffL.
The page mask which is used to make sure memory_end is on 4MB/2MB
boundary is wrong and not needed. Therefore remove it.
Make sure a vmalloc area does also exist and work on (future)
machines with 4TB and more memory.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Using memcmp to compare ccw_dev_id implies that the whole structure (incl.
padding) has always been completely initialized to sane values. Comparing
the structures field by field doesn't make such assumptions.
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Need this at yet another file and don't want to add yet another
extern...
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Let one master cpu kill all other cpus instead of sending an external
interrupt to all other cpus so they can kill themselves.
Simplifies reipl/shutdown functions a lot.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
In case of reipl cpcmd gets called when all other cpus are not running
anymore. To prevent deadlocks change __cpcmd so that it doesn't take
any locks and call cpcmd or __cpcmd, whatever is correct in the current
context.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
KEXEC_CONTROL_MEMORY_LIMIT is an unsigned long value and therefore
should be defined as one. Otherwise the kexec control page can be
allocated above 2GB which will cause a specification exception on the
sam31 instruction in the s390 kexec relocation code.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
In case of re-IPL and diag308 doesn't work we have to reset all devices
manually and wait synchronously that each reset finished.
This patch adds the necessary infrastucture and the first exploiter of it.
Subsystems that need to add a function that needs to be called at re-IPL
may register/unregister this function via
struct reset_call {
struct reset_call *next;
void (*fn)(void);
};
void register_reset_call(struct reset_call *reset);
void unregister_reset_call(struct reset_call *reset);
When the registered function get called the context is:
- all cpus beside the current one are stopped
- all machine checks and interrupts are disabled
- prefixing is disabled
- a default machine check handler is available for use
The registered functions may not take any locks are sleep.
For the common I/O layer part of this patch:
Introduce a reset_call css_reset that does the following:
- clear all subchannels
- perform a rchp on all channel paths and wait for the resulting
machine checks
This replaces the calls to clear_all_subchannels() and
cio_reset_channel_paths() for kexec and ccw reipl. reipl_ccw_dev() now
uses reipl_find_schid() to determine the subchannel id for a given
device id.
Also remove cio_reset_channel_paths() and friends since they are not
needed anymore.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Get rid of our own user_termio_to_kernel_termios() and
kernel_termios_to_user_termio() macros which didn't check for errors
on user space accesses. Instead use the generic functions which
handle this properly.
In addition the generic version of user_termio_to_kernel_termios()
also copies the c_line member which was missing in our variant.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Follow other architectures and add __must_check to uaccess functions.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Remove CPU_FTR_16M_PAGE from the cupfeatures mask at runtime on iSeries.
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Adds utility routines used by 52xx device drivers and board support
code. Main functionality is to add device nodes to the of_platform_bus,
retrieve the IPB bus frequency, and find+ioremap device registers.
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
There is no need to expose these settings outside the scope
of the interrupt controller code itself.
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>
To allow arch/powerpc/kernel/crash.c to build on 32-bit we need a
definition of hard_irq_disable(). 32-bit doesn't support the lazy
interrupt disabling mechanism, so on 32-bit hard_irq_disable() is
simply local_irq_disable(). Add a definition for hard_irq_enable()
just for completeness.
This allows (KEXEC=y && PPC32=y) to build again. Broken since
d04c56f73c.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>
"extern inline" generates a warning with -Wmissing-prototypes and I'm
currently working on getting the kernel cleaned up for adding this to
the CFLAGS since it will help us to avoid a nasty class of runtime
errors.
If there are places that really need a forced inline, __always_inline
would be the correct solution.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Paul Mackerras <paulus@samba.org>
At least the ide driver calls pcibus_to_node, which is not
defined when CONFIG_PCI is disabled. This adds a nop function
for the !PCI case.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
A few code paths need to check whether or not they are running
on the PS3's LV1 hypervisor before making hcalls. This introduces
a new firmware feature bit for this, FW_FEATURE_PS3_LV1.
Now when both PS3 and IBM_CELL_BLADE are enabled, but not PSERIES,
FW_FEATURE_PS3_LV1 and FW_FEATURE_LPAR get enabled at compile time,
which is a bug. The same problem can also happen for (PPC_ISERIES &&
!PPC_PSERIES && PPC_SOMETHING_ELSE). In order to solve this, I
introduce a new CONFIG_PPC_NATIVE option that is set when at least
one platform is selected that can run without a hypervisor and then
turns the firmware feature check into a run-time option.
The new cell oprofile support that was recently merged does not
work on hypervisor based platforms like the PS3, therefore make
it depend on PPC_CELL_NATIVE instead of PPC_CELL. This may change
if we get oprofile support for PS3.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
When renaming CONFIG_PS3 to CONFIG_PPC_PS3, a few occurrences have been
missed.
I also fixed up the alignment in arch/powerpc/platforms/Makefile.
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
It may be desireable to build a kernel for cell without
spufs, e.g. as the initial kboot kernel. This requires
that the SPU specific parts of the core dump and the xmon
code depend on CONFIG_SPU_BASE instead of CONFIG_PPC_CELL.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Adds a PS3 system bus driver. This system bus is a virtual bus used to present
the PS3 system devices in the LDM.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Adds some needed bits for a config option PS3_USE_LPAR_ADDR that disables
the PS3 lpar address translation mechanism. This is a currently needed
workaround for limitations in the design of the generic cell spu support.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Adds the core platform support for the PS3 game console and other devices
using the PS3 hypervisor.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This adds a platform specific spu management abstraction and the coresponding
routines to support the IBM Cell Blade. It also removes the hypervisor only
resources that were included in struct spu.
Three new platform specific routines are introduced, spu_enumerate_spus(),
spu_create_spu() and spu_destroy_spu(). The underlying design uses a new
type, struct spu_management_ops, to hold function pointers that the platform
setup code is expected to initialize to instances appropriate to that platform.
For the IBM Cell Blade support, I put the hypervisor only resources that were
in struct spu into a platform specific data structure struct spu_pdata.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This adds an accessor routine virq_to_hw() to the
virq routines which hides the implementation details
of the virq to hwirq map.
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
It saves #ifdef'ing in callers if we at least define the 64-bit cpu
features for 32-bit also.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This patch adds SPU elf notes to the coredump. It creates a separate note
for each of /regs, /fpcr, /lslr, /decr, /decr_status, /mem, /signal1,
/signal1_type, /signal2, /signal2_type, /event_mask, /event_status,
/mbox_info, /ibox_info, /wbox_info, /dma_info, /proxydma_info, /object-id.
A new macro, ARCH_HAVE_EXTRA_NOTES, was created for architectures to
specify they have extra elf core notes.
A new macro, ELF_CORE_EXTRA_NOTES_SIZE, was created so the size of the
additional notes could be calculated and added to the notes phdr entry.
A new macro, ELF_CORE_WRITE_EXTRA_NOTES, was created so the new notes
would be written after the existing notes.
The SPU coredump code resides in spufs. Stub functions are provided in the
kernel which are hooked into the spufs code which does the actual work via
register_arch_coredump_calls().
A new set of __spufs_<file>_read/get() functions was provided to allow the
coredump code to read from the spufs files without having to lock the
SPU context for each file read from.
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This adds code to look at the properties firmware puts in the device
tree to determine what compatibility mode the partition is in on
POWER6 machines, and set the ELF aux vector AT_HWCAP and AT_PLATFORM
entries appropriately.
Specifically, we look at the cpu-version property in the cpu node(s).
If that contains a "logical" PVR value (of the form 0x0f00000x), we
call identify_cpu again with this PVR value. A value of 0x0f000001
indicates the partition is in POWER5+ compatibility mode, and a value
of 0x0f000002 indicates "POWER6 architected" mode, with various
extensions disabled. We also look for various other properties:
ibm,dfp, ibm,purr and ibm,spurr.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Add PPU event-based and cycle-based profiling support to Oprofile for Cell.
Oprofile is expected to collect data on all CPUs simultaneously.
However, there is one set of performance counters per node. There are
two hardware threads or virtual CPUs on each node. Hence, OProfile must
multiplex in time the performance counter collection on the two virtual
CPUs.
The multiplexing of the performance counters is done by a virtual
counter routine. Initially, the counters are configured to collect data
on the even CPUs in the system, one CPU per node. In order to capture
the PC for the virtual CPU when the performance counter interrupt occurs
(the specified number of events between samples has occurred), the even
processors are configured to handle the performance counter interrupts
for their node. The virtual counter routine is called via a kernel
timer after the virtual sample time. The routine stops the counters,
saves the current counts, loads the last counts for the other virtual
CPU on the node, sets interrupts to be handled by the other virtual CPU
and restarts the counters, the virtual timer routine is scheduled to run
again. The virtual sample time is kept relatively small to make sure
sampling occurs on both CPUs on the node with a relatively small
granularity. Whenever the counters overflow, the performance counter
interrupt is called to collect the PC for the CPU where data is being
collected.
The oprofile driver relies on a firmware RTAS call to setup the debug bus
to route the desired signals to the performance counter hardware to be
counted. The RTAS call must set the routing registers appropriately in
each of the islands to pass the signals down the debug bus as well as
routing the signals from a particular island onto the bus. There is a
second firmware RTAS call to reset the debug bus to the non pass thru
state when the counters are not in use.
Signed-off-by: Carl Love <carll@us.ibm.com>
Signed-off-by: Maynard Johnson <mpjohn@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The following routines are added to arch/powerpc/platforms/cell/pmu.c:
cbe_clear_pm_interrupts()
cbe_enable_pm_interrupts()
cbe_disable_pm_interrupts()
cbe_query_pm_interrupts()
cbe_pm_irq()
cbe_init_pm_irq()
This also adds a routine in arch/powerpc/platforms/cell/interrupt.c and
some macros in cbe_regs.h to manipulate the IIC_IR register:
iic_set_interrupt_routing()
Signed-off-by: Kevin Corry <kevcorry@us.ibm.com>
Signed-off-by: Carl Love <carll@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Move some PMU-related macros and function prototypes from cbe_regs.h
and pmu.h in arch/powerpc/platforms/cell/ to a new header at
include/asm-powerpc/cell-pmu.h
This is cleaner to use from the oprofile code, since that sits in
arch/powerpc/oprofile, not in the cell platform directory.
Signed-off-by: Kevin Corry <kevcorry@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The /lslr file gives read access to the SPU_LSLR register in hex; 0x3fff
for example The /dma_info file provides read access to the SPU Command
Queue in a binary format. The /proxydma_info files provides read access
access to the Proxy Command Queue in a binary format. The spu_info.h
file provides data structures for interpreting the binary format of
/dma_info and /proxydma_info.
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Change the definition of powerpc's cond_syscall() to use the standard gcc
weak attribute specifier which provides proper support for C linkage as
needed by spu_syscall_table[].
Fixes this powerpc build error with CONFIG_SPU_FS=y, CONFIG_PPC_RTAS=n:
arch/powerpc/platforms/built-in.o: undefined reference to `ppc_rtas'
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This is a small driver for the Xserve G5 CPU-meter blue LEDs on the
front-panel. It might work on the Xserve G4 as well though that was
not tested. It's pretty basic and could use some improvements if
somebody cares doing them. :)
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Per email discussion, it appears that rtas_stop_self()
and pSeries_mach_cpu_die() should not be compiled if
CONFIG_HOTPLUG_CPU is not defined. This patch adds
#ifdefs around these bits of code.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Rewrite local_get_flags and local_irq_disable to use r13 explicitly,
to avoid the risk that gcc will split get_paca()->soft_enabled into a
sequence unsafe against preemption. Similar care in local_irq_restore.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The recent IO accessor changes broke IDE on arch/ppc due to the IDE
stream IO macros using the new reads/writes{b,w,l} accessors that
are only defined for arch/powerpc. This adds them to arch/ppc.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The new IO accessor code allows to stick a token in the top bit of MMIO
addresses which gets masked out during actual accesses. However, the
__raw_* accessors forgot to mask it out. This fixes it.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
powerpc: Merge 32 and 64 bits asm-powerpc/io.h
The rework on io.h done for the new hookable accessors made it easier,
so I just finished the work and merged 32 and 64 bits io.h for arch/powerpc.
arch/ppc still uses the old version in asm-ppc, there is just too much gunk
in there that I really can't be bothered trying to cleanup.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch adds full cell iommu support (and iommu disabled mode).
It implements mapping/unmapping of iommu pages on demand using the
standard powerpc iommu framework. It also supports running with
iommu disabled for machines with less than 2GB of memory. (The
default is off in that case, though it can be forced on with the
kernel command line option iommu=force).
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch adds an optional global offset that can be added to DMA addresses
when using the direct DMA operations.
That brings it a step closer to the 32 bits direct DMA operations, and makes
it useable on Cell when the MMU is disabled and we are using a spider
southbridge.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>