The Marvell Kirkwood (88F6000) is a family of ARM SoCs based on a
Shiva CPU core, and features a DDR2 controller, a x1 PCIe interface,
a USB 2.0 interface, a SPI controller, a crypto accelerator, a TS
interface, and IDMA/XOR engines, and depending on the model, also
features one or two Gigabit Ethernet interfaces, two SATA II
interfaces, one or two TWSI interfaces, one or two UARTs, a
TDM/SLIC interface, a NAND controller, an I2S/SPDIF interface, and
an SDIO interface.
This patch adds supports for the Marvell DB-88F6281-BP Development
Board and the RD-88F6192-NAS and the RD-88F6281 Reference Designs,
enabling support for the PCIe interface, the USB interface, the
ethernet interfaces, the SATA interfaces, the TWSI interfaces, the
UARTs, and the NAND controller.
This patch adds support for the unified Feroceon L2 cache controller
as found in e.g. the Marvell Kirkwood and Marvell Discovery Duo
families of ARM SoCs.
Note that:
- Page table walks are outer uncacheable on Kirkwood and Discovery
Duo, since the ARMv5 spec provides no way to indicate outer
cacheability of page table walks (specifying it in TTBR[4:3] is
an ARMv6+ feature).
This requires adding L2 cache clean instructions to
proc-feroceon.S (dcache_clean_area(), set_pte()) as well as to
tlbflush.h ({flush,clean}_pmd_entry()). The latter case is handled
by defining a new TLB type (TLB_FEROCEON) which is almost identical
to the v4wbi one but provides a TLB_L2CLEAN_FR flag.
- The Feroceon L2 cache controller supports L2 range (i.e. 'clean L2
range by MVA' and 'invalidate L2 range by MVA') operations, and this
patch uses those range operations for all Linux outer cache
operations, as they are faster than the regular per-line operations.
L2 range operations are not interruptible on this hardware, which
avoids potential livelock issues, but can be bad for interrupt
latency, so there is a compile-time tunable (MAX_RANGE_SIZE) which
allows you to select the maximum range size to operate on at once.
(Valid range is between one cache line and one 4KiB page, and must
be a multiple of the line size.)
The Marvell Loki (88RC8480) is an ARM SoC based on a Feroceon CPU
core running at between 400 MHz and 1.0 GHz, and features a 64 bit
DDR controller, 512K of internal SRAM, two x4 PCI-Express ports,
two Gigabit Ethernet ports, two 4x SAS/SATA controllers, two UARTs,
two TWSI controllers, and IDMA/XOR engines.
This patch adds support for the Marvell LB88RC8480 Development
Board, enabling the use of the PCIe interfaces, the ethernet
interfaces, the TWSI interfaces and the UARTs.
Ke Wei [Fri, 23 May 2008 08:23:22 +0000 (10:23 +0200)]
[ARM] Orion: add a separate BRIDGE_INT_TIMER1_CLR define
Some Feroceon-based SoCs have an MBUS bridge interrupt controller
that requires writing a one instead of a zero to clear edge
interrupt sources such as timer expiry.
This patch adds a new BRIDGE_INT_TIMER1_CLR define, which platform
code can set to either ~BRIDGE_INT_TIMER1 (write-zero-to-clear) or
BRIDGE_INT_TIMER1 (write-one-to-clear) depending on the platform.
Ke Wei [Sun, 22 Jun 2008 20:45:00 +0000 (22:45 +0200)]
[ARM] Feroceon: allow more old Feroceon IDs
There are a couple more Feroceon-based SoCs out in the field that use
different Variant and Architecture fields in their Main ID registers
-- this patch tweaks the processor match/mask in proc-feroceon.S to
catch those SoCs as well.
Signed-off-by: Ke Wei <kewei@marvell.com> Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Nicolas Pitre [Tue, 3 Jun 2008 21:06:21 +0000 (23:06 +0200)]
[ARM] Feroceon: catch other Feroceon CPU IDs in head.S
Tweak the Feroceon match/mask in arch/arm/boot/compressed/head.S to
match a couple of newer Feroceon cores (such as the 88fr571vd with
CPU ID 0x56155710, and the 88fr131 with CPU ID 0x56251310) as well.
Signed-off-by: Nicolas Pitre <nico@marvell.com> Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Nicolas Pitre [Thu, 24 Apr 2008 00:04:54 +0000 (02:04 +0200)]
[ARM] Feroceon: speed up flushing of the entire cache
Flushing the L1 D cache with a test/clean/invalidate loop is very
easy in software, but it is not the quickest way of doing it, as
there is a lot of overhead involved in re-scanning the cache from
the beginning every time we hit a dirty line.
This patch makes proc-feroceon.S use "clean+invalidate by set/way"
loops according to possible cache configuration of Feroceon CPUs
(either direct-mapped or 4-way set associative).
Signed-off-by: Nicolas Pitre <nico@marvell.com> Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
[ARM] Orion: implement power-off method for Kurobox Pro
This patch implements the communication with the microcontroller on the
Kurobox Pro and Linkstation Pro/Live boards. This is allowing to send
the commands needed to power-off the board correctly.
Signed-off-by: Sylver Bruneau <sylver.bruneau@googlemail.com> Acked-by: Russell King <linux@arm.linux.org.uk> Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
The mv643xx_eth platform data field ->force_phy_addr only needs
to be set if the passed-in ->phy_addr field is zero (to distinguish
the case of not having specified a phy address
(force_phy_addr = 0) from the case where a phy address of zero needs
to be used (force_phy_addr = 1.))
Also, the ->force_phy_addr field will hopefully disappear in a
future mv643xx_eth reorganisation.
Therefore, this patch deletes the ->force_phy_addr field initialiser
from all Orion board code.
[ARM] Orion: remove error printks in ->map_irq() implementations
If all PCI devices are working as expected, the error printks in the
various implementations of ->map_irq() doesn't really provide any
useful info. And if something is not working as expected, turning
on pci=debug gives you more useful information than the printk calls
in ->map_irq(), since the former also tells you which devices _did_
get IRQs successfully assigned. Therefore, delete these printks
entirely.
Spotted by Russell King.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Russell King <linux@arm.linux.org.uk>
Instead of having board code poke directly into the MPP configuration
registers, and separately calling orion5x_gpio_set_valid_pins() to
indicate which MPP pins can be used as GPIO pins, introduce a helper
function for configuring the roles of each of the MPP pins, and have
that helper function handle gpio validity internally.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Sylver Bruneau <sylver.bruneau@googlemail.com> Acked-by: Russell King <linux@arm.linux.org.uk>
[ARM] Orion: move setting up PCIe WA window into PCIe setup path
It makes no sense to do PCIe WA window setup in the individual
board support files while the decision whether or not to use the
PCIe WA access method is made in a different place, in the PCIe
support code.
This patch moves the configuration of a PCIe WA window from the
individual Orion board support files to the central Orion PCIe
support code.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Russell King <linux@arm.linux.org.uk>
[ARM] Orion: move EHCI/I2C/UART peripheral init into board code
This patch moves initialisation of EHCI/I2C/UART platform devices
from the common orion5x_init() into the board support code.
The rationale behind this is that only the board support code knows
whether certain peripherals have been brought out on the board, and
not initialising peripherals that haven't been brought out is
desirable for example:
- to reduce user confusion (e.g. seeing both 'eth0' and 'eth1'
appear while there is only one ethernet port on the board); and
- to allow for future power savings (peripherals that have not
been brought out can be clock gated off entirely).
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Russell King <linux@arm.linux.org.uk>
Make it clear that Orion top-level IRQs are level-triggered. This
means that we don't need an ->ack() handler, or at least, we don't
need the ->ack() handler (or the acking part of the ->mask_ack()
handler) to actually do anything.
Given that, we might as well point our ->mask_ack() handler at the
->mask() handler instead of providing a dummy ->ack() handler, since
providing a ->mask_ack() handler on level IRQ sources will prevent
->ack() from ever being called.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Russell King <linux@arm.linux.org.uk>
[ARM] Orion: DRAM mapping granularity is 64KiB, not 16MiB
The DRAM base address and size fields in the CPU's MBUS bridge have
64KiB granularity, instead of the currently used 16MiB. Since all
of the currently supported MBUS peripherals support 64KiB granularity
as well, this patch changes the Orion address map code to stop
rounding base addresses down and sizes up to multiples of 16MiB.
Found by Ke Wei <kewei@marvell.com>.
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Russell King <linux@arm.linux.org.uk>
Currently, Orion window setup uses hardcoded window indexes for each
of the boot/cs0/cs1/cs2/PCIe WA windows. The static window allocation
used can clash if board support code will ever attempt to configure
both a dev2 and a PCIe WA window, as both of those use CPU mbus window
#7 at present.
This patch keeps track of the last used window, and opens subsequently
requested windows sequentially, starting from 4. (Windows 0-3 are used
as MEM/IO windows for the PCI/PCIe buses.)
[ARM] Orion: fix various whitespace and coding style issues
More cosmetic cleanup:
- Replace 8-space indents by proper tab indents.
- In structure initialisers, use a trailing comma for every member.
- Collapse "},\n{" in structure initialiers to "}, {".
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com> Acked-by: Russell King <linux@arm.linux.org.uk>
Nicolas Pitre [Mon, 31 Mar 2008 16:38:31 +0000 (12:38 -0400)]
[ARM] cache align destination pointer when copying memory for some processors
The implementation for memory copy functions on ARM had a (disabled)
provision for aligning the source pointer before loading registers with
data. Turns out that aligning the _destination_ pointer is much more
useful, as the read side is already sufficiently helped with the use of
preload.
So this changes the definition of the CALGN() macro to target the
destination pointer instead, and turns it on for Feroceon processors
where the gain is very noticeable.
Signed-off-by: Nicolas Pitre <nico@marvell.com> Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Nicolas Pitre [Thu, 24 Apr 2008 05:31:46 +0000 (01:31 -0400)]
[ARM] latencytop support
Available for !SMP only at the moment.
From Russell:
|Basically, if a thread is running on a CPU, thread_saved_fp() is invalid.
|So, the question is: what guarantees do we have here that 'tsk' is not
|running on another CPU?
* git://git.kernel.org/pub/scm/linux/kernel/git/bart/ide-2.6:
BAST: Remove old IDE driver
pcmcia ide kingston compactflash's have a new manufacturer id
pcmcia: add another pata/ide ID
pcmcia: add an pata/ide ID
ide: increase timeout in wait_drive_not_busy()
palm_bk3710: fix resource management
Linus Torvalds [Fri, 20 Jun 2008 19:41:10 +0000 (12:41 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
ieee1394: Kconfig menu touch-up
firewire: Kconfig menu touch-up
firewire: deadline for PHY config transmission
firewire: fw-ohci: unify printk prefixes
firewire: fill_bus_reset_event needs lock protection
firewire: fw-ohci: write selfIDBufferPtr before LinkControl.rcvSelfID
firewire: fw-ohci: disable PHY packet reception into AR context
firewire: fw-ohci: use of uninitialized data in AR handler
firewire: don't panic on invalid AR request buffer
Linus Torvalds [Fri, 20 Jun 2008 19:39:12 +0000 (12:39 -0700)]
Merge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6
* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6:
ACPI: no AC status notification
ACPI Exception (video-1721): UNKNOWN_STATUS_CODE, Cant attach device
Linus Torvalds [Fri, 20 Jun 2008 19:38:18 +0000 (12:38 -0700)]
Merge branch 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6
* 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (21 commits)
drm: only trust core drm ioctls - driver ioctls are a mess.
drm/i915: add support for Intel series 4 chipsets.
drm/radeon: add hier-z registers for r300 and r500 chipsets
drm/radeon: use DSTCACHE_CTLSTAT rather than RB2D_DSTCACHE_CTLSTAT
drm/radeon: switch IGP gart to use radeon_write_agp_base()
drm/radeon: Restore sw interrupt on resume
drm/r500: add support for AGP based cards.
drm/radeon: fix texture uploads with large 3d textures (bug 13980)
drm/radeon: add initial r500 support.
drm/radeon: init pipe setup in kernel code.
drm/radeon: fixup radeon_do_engine_reset
drm/radeon: fix pixcache and purge/cache flushing registers
drm/radeon: write AGP_BASE_2 on chips that support it.
drm/radeon: merge IGP chip setup and fixup RS400 vs RS480 support
drm/radeon: IGP clean up register and magic numbers.
drm/rs690: set base 2 to 0.
drm/rs690: set all of gart base address.
radeon: add production microcode from AMD
drm: pcigart use proper pci map interfaces.
drm: the sg alloc ioctl should write back the handle to userspace
...
Linus Torvalds [Fri, 20 Jun 2008 19:37:55 +0000 (12:37 -0700)]
Merge branch 'agp-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/agp-2.6
* 'agp-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/agp-2.6:
[agp]: fixup chipset flush for new Intel G4x.
agp: brown paper bag patch - put back the two lines it took out.
Linus Torvalds [Fri, 20 Jun 2008 19:37:13 +0000 (12:37 -0700)]
Merge branch 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
softlockup: fix NMI hangs due to lock race - 2.6.26-rc regression
rcupreempt: remove export of rcu_batches_completed_bh
cpuset: limit the input of cpuset.sched_relax_domain_level
Linus Torvalds [Fri, 20 Jun 2008 19:36:55 +0000 (12:36 -0700)]
Merge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
sched, delay accounting: fix incorrect delay time when constantly waiting on runqueue
sched: CPU hotplug events must not destroy scheduler domains created by the cpusets
sched: rt-group: fix RR buglet
sched: rt-group: heirarchy aware throttle
sched: rt-group: fix hierarchy
sched: NULL pointer dereference while setting sched_rt_period_us
sched: fix defined-but-unused warning
Linus Torvalds [Fri, 20 Jun 2008 19:36:38 +0000 (12:36 -0700)]
Merge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86, geode: add a VSA2 ID for General Software
x86: use BOOTMEM_EXCLUSIVE on 32-bit
x86, 32-bit: fix boot failure on TSC-less processors
x86: fix NULL pointer deref in __switch_to
x86: set PAE PHYSICAL_MASK_SHIFT to 44 bits.
Linus Torvalds [Fri, 20 Jun 2008 19:34:43 +0000 (12:34 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/blackfin-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/blackfin-2.6:
Blackfin Serial Driver: Use timer to poll CTS PIN instead of workqueue.
Blackfin arch: fix typo error in bf548 serial header file
Linus Torvalds [Fri, 20 Jun 2008 19:31:03 +0000 (12:31 -0700)]
Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
ahci: sis can't do PMP
ata_piix: add TECRA M4 to broken suspend list
LIBATA: Add HAVE_PATA_PLATFORM to select PATA_PLATFORM driver
sata_mv: warn on PIO with multiple DRQs
sata_mv: enable async_notify for 60x1 Rev.C0 and higher
libata: don't check whether to use DMA or not for no data commands
ahci: jmb361 has only one port
Linus Torvalds [Fri, 20 Jun 2008 19:19:28 +0000 (12:19 -0700)]
[watchdog] hpwdt: fix use of inline assembly
The inline assembly in drivers/watchdog/hpwdt.c was incredibly broken,
and included all the function prologue and epilogue stuff, even though
it was itself then inside a C function where the compiler would add its
own prologue and epilogue on top of it all.
This then just _happened_ to work if you had exactly the right compiler
version and exactly the right compiler flags, so that gcc just happened
to not create any prologue at all (the gcc-generated epilogue wouldn't
matter, since it would never be reached).
But the more proper way to fix it is to simply not do this. Move the
inline asm to the top level, with no surrounding function at all (the
better alternative would be to remove the prologue and make it actually
use proper description of the arguments to the inline asm, but that's a
bigger change than the one I'm willing to make right now).
pcmcia ide kingston compactflash's have a new manufacturer id
Up to now, Kingston compactflash cards (ab)used the Toshiba Manufacturer's ID,
In their new CF cards, they use a new one. Let's the ide subsystem
recognize CF cards with the new id.
Signed-off-by: Christophe Niclaes <cniclaes@develtech.com> Acked-by: Philippe De Muyter <phdm@macqel.be> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Dominik Brodowski <linux@dominikbrodowski.net> Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Some ATAPI devices take longer than the current max timeout value to
become ready (i.e. TEAC DV-W28ECW takes 6 ms) so increase the timeout
value to 10 ms.
This fixes kernel.org bugzilla bug #10887:
http://bugzilla.kernel.org/show_bug.cgi?id=10887
Sergei Shtylyov [Fri, 20 Jun 2008 18:53:32 +0000 (20:53 +0200)]
palm_bk3710: fix resource management
The driver expected a *virtual* address in the IDE platform device's memory
resource and didn't request the memory region for the register block. Fix this
taking into account the fact that DaVinci SoC devices are fixed-mapped to the
virtual memory early and we can get their virtual addresses using IO_ADDRESS()
macro, not having to call ioremap()...
While at it, also do some cosmetic changes...
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com> Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Linus Torvalds [Fri, 20 Jun 2008 18:18:25 +0000 (11:18 -0700)]
Reinstate ZERO_PAGE optimization in 'get_user_pages()' and fix XIP
KAMEZAWA Hiroyuki and Oleg Nesterov point out that since the commit 557ed1fa2620dc119adb86b34c614e152a629a80 ("remove ZERO_PAGE") removed
the ZERO_PAGE from the VM mappings, any users of get_user_pages() will
generally now populate the VM with real empty pages needlessly.
We used to get the ZERO_PAGE when we did the "handle_mm_fault()", but
since fault handling no longer uses ZERO_PAGE for new anonymous pages,
we now need to handle that special case in follow_page() instead.
In particular, the removal of ZERO_PAGE effectively removed the core
file writing optimization where we would skip writing pages that had not
been populated at all, and increased memory pressure a lot by allocating
all those useless newly zeroed pages.
This reinstates the optimization by making the unmapped PTE case the
same as for a non-existent page table, which already did this correctly.
While at it, this also fixes the XIP case for follow_page(), where the
caller could not differentiate between the case of a page that simply
could not be used (because it had no "struct page" associated with it)
and a page that just wasn't mapped.
We do that by simply returning an error pointer for pages that could not
be turned into a "struct page *". The error is arbitrarily picked to be
EFAULT, since that was what get_user_pages() already used for the
equivalent IO-mapped page case.
[ Also removed an impossible test for pte_offset_map_lock() failing:
that's not how that function works ]
Jordan Crouse [Wed, 18 Jun 2008 17:34:38 +0000 (11:34 -0600)]
x86, geode: add a VSA2 ID for General Software
General Software writes their own VSA2 module for their version
of the Geode BIOS, which returns a different ID then the standard
VSA2. This was causing the framebuffer driver to break for most
GSW boards.
Bharath Ravi [Mon, 16 Jun 2008 09:41:01 +0000 (15:11 +0530)]
sched, delay accounting: fix incorrect delay time when constantly waiting on runqueue
This patch corrects the incorrect value of per process run-queue wait
time reported by delay statistics. The anomaly was due to the following
reason. When a process leaves the CPU and immediately starts waiting for
CPU on the runqueue (which means it remains in the TASK_RUNNABLE state),
the time of re-entry into the run-queue is never recorded. Due to this,
the waiting time on the runqueue from this point of re-entry upto the
next time it hits the CPU is not accounted for. This is solved by
recording the time of re-entry of a process leaving the CPU in the
sched_info_depart() function IF the process will go back to waiting on
the run-queue. This IF condition is verified by checking whether the
process is still in the TASK_RUNNABLE state.
The patch was tested on 2.6.26-rc6 using two simple CPU hog programs.
The values noted prior to the fix did not account for the time spent on
the runqueue waiting. After the fix, the correct values were reported
back to user space.
Signed-off-by: Bharath Ravi <bharathravi1@gmail.com> Signed-off-by: Madhava K R <madhavakr@gmail.com> Cc: dhaval@linux.vnet.ibm.com Cc: vatsa@in.ibm.com Cc: balbir@in.ibm.com Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Ingo Molnar <mingo@elte.hu>
x86, 32-bit: fix boot failure on TSC-less processors
Booting 2.6.26-rc6 on my 486 DX/4 fails with a "BUG: Int 6"
(invalid opcode) and a kernel halt immediately after the
kernel has been uncompressed. The BUG shows EIP pointing
to an rdtsc instruction in native_read_tsc(), invoked from
native_sched_clock().
(This error occurs so early that not even the serial console
can capture it.)
>x86: distangle user disabled TSC from unstable
>
>tsc_enabled is set to 0 from the command line switch "notsc" and from
>the mark_tsc_unstable code. Seperate those functionalities and replace
>tsc_enable with tsc_disable. This makes also the native_sched_clock()
>decision when to use TSC understandable.
>
>Preparatory patch to solve the sched_clock() issue on 32 bit.
>
>Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The core reason for this bug is that native_sched_clock() gets
called before tsc_init().
Before the commit above, tsc_32.c used a "tsc_enabled" variable
which defaulted to 0 == disabled, and which only got enabled late
in tsc_init(). Thus early calls to native_sched_clock() would skip
the TSC and use jiffies instead.
After the commit above, tsc_32.c uses a "tsc_disabled" variable
which defaults to 0, meaning that the TSC is Ok to use. Early calls
to native_sched_clock() now erroneously try to use the TSC on
!cpu_has_tsc processors, leading to invalid opcode exceptions.
My proposed fix is to initialise tsc_disabled to a "soft disabled"
state distinct from the hard disabled state set up by the "notsc"
kernel option. This fixes the native_sched_clock() problem. It also
allows tsc_init() to be simplified: instead of setting tsc_disabled = 1
on every error return, we just set tsc_disabled = 0 once when all
checks have succeeded.
I've verified that this lets my 486 boot again. I've also verified
that a Core2 machine still uses the TSC as clocksource after the patch.
Signed-off-by: Mikael Pettersson <mikpe@it.uu.se> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Suresh Siddha [Fri, 13 Jun 2008 22:47:12 +0000 (15:47 -0700)]
x86: fix NULL pointer deref in __switch_to
Patrick McHardy reported a crash:
> > I get this oops once a day, its apparently triggered by something
> > run by cron, but the process is a different one each time.
> >
> > Kernel is -git from yesterday shortly before the -rc6 release
> > (last commit is the usb-2.6 merge, the x86 patches are missing),
> > .config is attached.
> >
> > I'll retry with current -git, but the patches that have gone in
> > since I last updated don't look related.
> >
> > [62060.043009] BUG: unable to handle kernel NULL pointer dereference at
> > 000001ff
> > [62060.043009] IP: [<c0102a9b>] __switch_to+0x2f/0x118
> > [62060.043009] *pde = 00000000
> > [62060.043009] Oops: 0002 [#1] PREEMPT
Vegard Nossum analyzed it:
> This decodes to
>
> 0: 0f ae 00 fxsave (%eax)
>
> so it's related to the floating-point context. This is the exact
> location of the crash:
>
> $ addr2line -e arch/x86/kernel/process_32.o -i ab0
> include/asm/i387.h:232
> include/asm/i387.h:262
> arch/x86/kernel/process_32.c:595
>
> ...so it looks like prev_task->thread.xstate->fxsave has become NULL.
> Or maybe it never had any other value.
Somehow (as described below) TS_USEDFPU is set but the fpu is not
allocated or freed.
Another possible FPU pre-emption issue with the sleazy FPU optimization
which was benign before but not so anymore, with the dynamic FPU allocation
patch.
New task is getting exec'd and it is prempted at the below point.
Now when it context switches in again, as the used_math() is still set
and fpu_counter can be > 5, we will do a math_state_restore() which sets
the task's TS_USEDFPU. After it continues from the above preemption point
it does clear_used_math() and much later free_thread_xstate().
Now, at the next context switch, it is quite possible that xstate is
null, used_math() is not set and TS_USEDFPU is still set. This will
trigger unlazy_fpu() causing kernel oops.
Fix this by clearing tsk's fpu_counter before clearing task's fpu.
When a 64-bit x86 processor runs in 32-bit PAE mode, a pte can
potentially have the same number of physical address bits as the
64-bit host ("Enhanced Legacy PAE Paging"). This means, in theory,
we could have up to 52 bits of physical address in a pte.
The 32-bit kernel uses a 32-bit unsigned long to represent a pfn.
This means that it can only represent physical addresses up to 32+12=44
bits wide. Rather than widening pfns everywhere, just set 2^44 as the
Linux x86_32-PAE architectural limit for physical address size.
This is a bugfix for two cases:
1. running a 32-bit PAE kernel on a machine with
more than 64GB RAM.
2. running a 32-bit PAE Xen guest on a host machine with
more than 64GB RAM
In both cases, a pte could need to have more than 36 bits of physical,
and masking it to 36-bits will cause fairly severe havoc.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: Jan Beulich <jbeulich@novell.com> Cc: <stable@kernel.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Jason Wessel [Tue, 27 May 2008 17:23:29 +0000 (12:23 -0500)]
softlockup: fix NMI hangs due to lock race - 2.6.26-rc regression
The touch_nmi_watchdog() routine on x86 ultimately calls
touch_softlockup_watchdog(). The problem is that to touch the
softlockup watchdog, the cpu_clock code has to be called which could
involve multiple cpu locks and can lead to a hard hang if one of the
locks is held by a processor that is not going to return anytime soon
(such as could be the case with kgdb or perhaps even with some other
kind of exception).
This patch causes the public version of the
touch_softlockup_watchdog() to defer the cpu clock access to a later
point.
The test case for this problem is to use the following kernel config
options:
It should be noted that kgdb test suite and these options were not
available until 2.6.26-rc2, so it was necessary to patch the kgdb
test suite during the bisection.
I would consider this patch a regression fix because the problem first
appeared in commit 27ec4407790d075c325e1f4da0a19c56953cce23 when some
logic was added to try to periodically sync the clocks. It was
possible to work around this particular problem by simply not
performing the sync anytime the system was in a critical context.
This was ok until commit 3e51f33fcc7f55e6df25d15b55ed10c8b4da84cd,
which added config option CONFIG_HAVE_UNSTABLE_SCHED_CLOCK and some
multi-cpu locks to sync the clocks. It became clear that accessing
this code from an nmi was the source of the lockups. Avoiding the
access to the low level clock code from an code inside the NMI
processing also fixed the problem with the 27ec44... commit.
Signed-off-by: Jason Wessel <jason.wessel@windriver.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Steven Rostedt [Thu, 22 May 2008 18:18:17 +0000 (14:18 -0400)]
rcupreempt: remove export of rcu_batches_completed_bh
In rcupreempt, rcu_batches_completed_bh is defined as a static inline in
the header file. This does not need to be exported, and not only that,
this breaks my PPC build.
Signed-off-by: Steven Rostedt <srostedt@redhat.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: paulus@samba.org Cc: linuxppc-dev@ozlabs.org Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Li Zefan [Tue, 13 May 2008 02:27:17 +0000 (10:27 +0800)]
cpuset: limit the input of cpuset.sched_relax_domain_level
We allow the inputs to be [-1 ... SD_LV_MAX), and return -EINVAL
for inputs outside this range.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com> Acked-by: Paul Menage <menage@google.com> Acked-by: Paul Jackson <pj@sgi.com> Acked-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Max Krasnyansky [Thu, 29 May 2008 18:17:01 +0000 (11:17 -0700)]
sched: CPU hotplug events must not destroy scheduler domains created by the cpusets
First issue is not related to the cpusets. We're simply leaking doms_cur.
It's allocated in arch_init_sched_domains() which is called for every
hotplug event. So we just keep reallocation doms_cur without freeing it.
I introduced free_sched_domains() function that cleans things up.
Second issue is that sched domains created by the cpusets are
completely destroyed by the CPU hotplug events. For all CPU hotplug
events scheduler attaches all CPUs to the NULL domain and then puts
them all into the single domain thereby destroying domains created
by the cpusets (partition_sched_domains).
The solution is simple, when cpusets are enabled scheduler should not
create default domain and instead let cpusets do that. Which is
exactly what the patch does.
Signed-off-by: Max Krasnyansky <maxk@qualcomm.com> Cc: pj@sgi.com Cc: menage@google.com Cc: rostedt@goodmis.org Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Peter Zijlstra [Thu, 19 Jun 2008 07:06:59 +0000 (09:06 +0200)]
sched: rt-group: fix RR buglet
In tick_task_rt() we first call update_curr_rt() which can dequeue a runqueue
due to it running out of runtime, and then we try to requeue it, of it also
having exhausted its RR quota. Obviously requeueing something that is no longer
on the runqueue will not have the expected result.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Tested-by: Daniel K. <dk@uw.no> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Peter Zijlstra [Thu, 19 Jun 2008 07:06:57 +0000 (09:06 +0200)]
sched: rt-group: heirarchy aware throttle
The bandwidth throttle code dequeues a group when it runs out of quota, and
re-queues it once the period rolls over and the quota gets refreshed.
Sadly it failed to take the hierarchy into consideration. Share more of the
enqueue/dequeue code with regular task opterations.
Also, some operations like sched_setscheduler() can dequeue/enqueue tasks that
are in throttled runqueues, we should not inadvertly re-enqueue empty runqueues
so check for that.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Tested-by: Daniel K. <dk@uw.no> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Linus Torvalds [Thu, 19 Jun 2008 04:52:35 +0000 (21:52 -0700)]
Merge branch 'agp-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/agp-2.6
* 'agp-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/agp-2.6:
agp/intel: cleanup some serious whitespace badness
[AGP] intel_agp: Add support for Intel 4 series chipsets
[AGP] intel_agp: extra stolen mem size available for IGD_GM chipset
agp: more boolean conversions.
drivers/char/agp - use bool
agp: two-stage page destruction issue
agp/via: fixup pci ids
drm/radeon: fix texture uploads with large 3d textures (bug 13980)
Texture uploads could hit the blitter coordinate limit, adjust the texture
offset when uploading the pieces. Make sure to check the end address of the
upload too.
projects / linux-2.6 / log