Commit Graph

45 Commits

Author SHA1 Message Date
Stephen Boyd 7fa22bd546 ARM: 6993/1: platsmp: Allow secondary cpu hotplug with maxcpus=1
If an ARM system has multiple cpus in the same socket and the
kernel is booted with maxcpus=1, secondary cpus are possible but
not present due to how platform_smp_prepare_cpus() is called.
Since most typical ARM processors don't actually support physical
hotplug, initialize the present map to be equal to the possible
map in generic ARM SMP code. Also, always call
platform_smp_prepare_cpus() as long as max_cpus is non-zero (0
means no SMP) to allow platform code to do any SMP setup.

After applying this patch it's possible to boot an ARM system
with maxcpus=1 on the command line and then hotplug in secondary
cpus via sysfs. This is more in line with how x86 does things.

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Cc: David Brown <davidb@codeaurora.org>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Srinidhi Kasagar <srinidhi.kasagar@stericsson.com>
Cc: Linus Walleij <linus.walleij@stericsson.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-07 14:45:07 +01:00
Russell King 0f7b332f97 ARM: consolidate SMP cross call implementation
Rather than having each platform class provide a mach/smp.h header for
smp_cross_call(), arrange for them to register the function with the
core ARM SMP code instead.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-05-23 16:53:17 +01:00
Russell King 0462b4477e ARM: realview/vexpress: consolidate SMP bringup code
Realview and Versatile Express share the same SMP bringup code, so
consolidate the two implementations.

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-02-19 11:11:41 +00:00
Russell King ec15038f2b ARM: realview,vexpress: fix section mismatch warning for pen_release
Fix two section mismatch warnings in the platform SMP bringup code for
Realview and Versatile Express:

WARNING: arch/arm/mach-realview/built-in.o(.text+0x8ac): Section mismatch in reference from the function write_pen_release() to the variable .cpuinit.data:pen_release
The function write_pen_release() references
the variable __cpuinitdata pen_release.
This is often because write_pen_release lacks a __cpuinitdata
annotation or the annotation of pen_release is wrong.

WARNING: arch/arm/mach-vexpress/built-in.o(.text+0x7b4): Section mismatch in reference from the function write_pen_release() to the variable .cpuinit.data:pen_release
The function write_pen_release() references
the variable __cpuinitdata pen_release.
This is often because write_pen_release lacks a __cpuinitdata
annotation or the annotation of pen_release is wrong.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-01-25 15:08:01 +00:00
Russell King 4073723acb Merge branch 'misc' into devel
Conflicts:
	arch/arm/Kconfig
	arch/arm/common/Makefile
	arch/arm/kernel/Makefile
	arch/arm/kernel/smp.c
2011-01-06 22:32:52 +00:00
Russell King 3705ff6da5 ARM: Fix subtle race in CPU pen_release hotplug code
There is a subtle race in the CPU hotplug code, where a CPU which has
been offlined can online itself before being requested, which results
in things going astray on the next online/offline cycle.

What happens in the normal online/offline/online cycle is:

	CPU0			CPU3
	requests boot of CPU3
	pen_release = 3
	flush cache line
				checks pen_release, reads 3
				starts boot
				pen_release = -1
	... requests CPU3 offline ...
				... dies ...
				checks pen_release, reads -1
	requests boot of CPU3
	pen_release = 3
	flush cache line
				checks pen_release, reads 3
				starts boot
				pen_release = -1

However, as the write of -1 of pen_release is not fully flushed back to
memory, and the checking of pen_release is done with caches disabled,
this allows CPU3 the opportunity to read the old value of pen_release:

	CPU0			CPU3
	requests boot of CPU3
	pen_release = 3
	flush cache line
				checks pen_release, reads 3
				starts boot
				pen_release = -1
	... requests CPU3 offline ...
				... dies ...
				checks pen_release, reads 3
				starts boot
				pen_release = -1
	requests boot of CPU3
	pen_release = 3
	flush cache line

Fix this by grouping the write of pen_release along with its cache line
flushing code to ensure that any update to pen_release is always pushed
out to physical memory.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-20 15:09:13 +00:00
Russell King 2c0136dba4 ARM: SMP: consolidate trace_hardirqs_off() into common SMP code
All platforms call trace_hardirqs_off() in their secondary startup code,
so move this into the core SMP code - it doesn't need to be in the
per-platform code.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-20 15:09:09 +00:00
Russell King 05c74a6cbc ARM: SMP: consolidate the common parts of smp_prepare_cpus()
There is a certain amount of smp_prepare_cpus() which doesn't belong
in the platform support code - that is, code which is invariant to the
SMP implementation.  Move this code into arch/arm/kernel/smp.c, and
add a platform_ prefix to the original function.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-20 15:09:08 +00:00
Russell King aec66ba1f7 ARM: SMP: get rid of references to poke_milo and Milo
Milo is an old boot loader, which is no longer relevant for these
platforms.  References to it are misleading.  Move the code out
of poke_milo(), and remove references to milo in comments.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-20 15:08:28 +00:00
Russell King fd778f0ad7 ARM: SMP: get rid of get_core_count()
We don't need this small function as well as scu_get_core_count()

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-20 15:08:27 +00:00
Russell King 8975b6c0fd ARM: SMP: Clean up ncores sanity checks
scu_get_core_count() never returns zero cores, so we don't need to
check and correct if ncores is zero.

Tegra was missing the check against NR_CPUS, leading to a potential
bitfield overflow if this becomes the case.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-20 15:08:27 +00:00
Russell King bbc3d14e9a ARM: SMP: move CPU number sanity checks to smp_init_cpus()
Ensure that the number of CPUs is sanity checked before setting
the number of possible CPUs.  This avoids any chance of overflowing
the cpu_possible bitmap.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-20 15:08:26 +00:00
Russell King 384895330e ARM: GIC: Remove MMIO address from gic_cpu_init, rename to gic_secondary_init
We don't need to re-pass the base address for the CPU interfaces to the
GIC for secondary CPUs, as it will never be different from the boot CPU
- and even if it was, we'd overwrite the boot CPU's base address.

Get rid of this argument, and rename to gic_secondary_init().

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-14 19:21:40 +00:00
Russell King ad3b6993b9 ARM: SMP: pass an ipi number to smp_cross_call()
This allows us to use smp_cross_call() to trigger a number of different
software generated interrupts, rather than combining them all on one
SGI.  Recover the SGI number via do_IPI.

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-12-03 08:26:30 +00:00
Colin Tuckley 157aed7473 RealView: Remove duplicated #define REALVIEW_SYS_FLAGS* statements
The platsmp.c file defines the REALVIEW_SYS_FLAGS* macros which are
already present in platform.h. Just use the latter.

Signed-off-by: Colin Tuckley <colin.tuckley@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-11-05 10:10:36 +00:00
Catalin Marinas dff2ab16ac Thumb-2: Pass a Thumb-2 address to the secondary CPUs to jump to
This patch sets bit 0 in the startup address passed to the secondary
CPUs so that they branch into Thumb-2 mode.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 12:33:00 +01:00
Russell King 42578c82e0 Merge branch 'for-rmk' of git://linux-arm.org/linux-2.6 into devel
Conflicts:
	arch/arm/Kconfig
	arch/arm/kernel/smp.c
	arch/arm/mach-realview/Makefile
	arch/arm/mach-realview/platsmp.c
2009-06-11 15:35:00 +01:00
Colin Tuckley 1b504bbe7a RealView: Add support for the RealView/PBX platform
This is a RealView platform supporting core tiles with ARM11MPCore,
Cortex-A8 or Cortex-A9 (multicore) processors. It has support for MMC,
CompactFlash, PCI-E.

Signed-off-by: Colin Tuckley <colin.tuckley@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-05-30 13:56:12 +01:00
Russell King e03cdade0c [ARM] smp: use new cpumask functions
Convert cpu_*_mask bit twiddling to the new set_cpu_*() API.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-05-28 14:16:52 +01:00
Russell King a8cbcd92bd [ARM] smp: separate SCU support code from realview
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-05-17 19:00:37 +01:00
Russell King 49613d4d9a [ARM] smp: SCU is used on non-realview platforms
The SCU can be used by non-realview platforms, so make it visible
for other people to use rather than having them copy the header file.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-05-17 18:59:57 +01:00
Russell King bc28248ee2 [ARM] smp: move core localtimer support out of platform specific files
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-05-17 18:58:34 +01:00
Russell King ee348d5a1d [ARM] realview: fix broadcast tick support
Having discussed broadcast tick support with Thomas Glexiner, the
broadcast tick devices should be registered with a higher rating
than the global tick device, and it should have the ONESHOT and
PERIODIC feature flags set.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Acked-by: Thomas Glexiner <tglx@linutronix.de>
2009-05-17 17:11:35 +01:00
Russell King 78d236c2b3 [ARM] realview: remove useless smp_cross_call_done()
smp_cross_call_done() is a no-op for MPCore, and since it's only
used by platform code, there's no point in having it unless it's
doing something.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-05-17 16:23:45 +01:00
Russell King 826681043d [ARM] smp: fix cpumask usage in ARM SMP code
The ARM SMP code wasn't properly updated for the cpumask changes, which
results in smp_timer_broadcast() broadcasting ticks to non-online CPUs.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-05-17 16:22:46 +01:00
Russell King 934848daa3 [ARM] Fix realview build
arch/arm/mach-realview/platsmp.c:140: error: 'jiffies' undeclared (first use in this function)
drivers/amba/bus.c:246: error: 'NO_IRQ' undeclared (first use in this function)

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-01-08 16:29:41 +00:00
Catalin Marinas 1bbdf637ba RealView: Clean up the machine_is_*() calls in platsmp.c
Some of the calls weren't necessary and some others were duplicated.
This patch tidies up the platsmp.c file.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-12-01 14:54:58 +00:00
Catalin Marinas ebac6546df RealView: Use only the shadow mapping of ARM11MPCore local timers
All the cases where the local timer for a CPU is accessed happen on the
corresponding current CPU, hence no need to access the per-CPU local
timer mappings.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-12-01 14:54:57 +00:00
Jon Callan 4c3ea37171 RealView: Add Cortex-A9 support to the EB board
This patch adds the necessary definitions and Kconfig entries to enable
Cortex-A9 (ARMv7 SMP) tiles on the RealView/EB board.

Signed-off-by: Jon Callan <Jon.Callan@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-12-01 14:54:56 +00:00
Russell King fced80c735 [ARM] Convert asm/io.h to linux/io.h
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-09-06 12:10:45 +01:00
Russell King a09e64fbc0 [ARM] Move include/asm-arm/arch-* to arch/arm/*/include/mach
This just leaves include/asm-arm/plat-* to deal with.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-07 09:55:48 +01:00
Russell King be50972935 [ARM] Remove asm/hardware.h, use asm/arch/hardware.h instead
Remove includes of asm/hardware.h in addition to asm/arch/hardware.h.
Then, since asm/hardware.h only exists to include asm/arch/hardware.h,
update everything to directly include asm/arch/hardware.h and remove
asm/hardware.h.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-07 09:40:08 +01:00
Catalin Marinas 08383ef29f [ARM] 5131/1: Annotate platform_secondary_init with trace_hardirqs_off
This patch annotates the platform_secondary_init function in
arch/arm/mach-realview/platsmp.c with trace_hardirqs_off to avoid a
warning when LOCKDEP and TRACE_IRQFLAGS are enabled.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-06-30 19:08:53 +01:00
Catalin Marinas e67172f579 RealView: Add the SMP initialisation support for PB11MPCore
This patch adds the initialisation calls for the SMP support on the
PB11MPCore platform.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-04-18 22:43:13 +01:00
Catalin Marinas b7b0ba942f RealView: Move the SCU initialisation out of __v6_setup
This patch moves the SCU initialisation from __v6_setup to the
smp_prepare_cpus() function as it relies on platform-specific
settings. Changes to get_core_count() are mainly for allowing cleaner
code with the upcoming PB11MPCore patches.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-04-18 22:43:08 +01:00
Catalin Marinas 7dd19e755d [ARM] 4818/1: RealView: Add core-tile detection
This patch adds the core-tile detection and only enables devices if the
corresponding tile is present. It currently detects the ARM11MPCore via
the core_tile_eb11mp() macro.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-02-04 17:52:24 +00:00
Catalin Marinas 356cb470b8 [ARM] 4816/1: RealView: Move the platform-specific definitions into board-eb.h
This patch moves the platform specific definitions from platform.h into
the board-eb.h file. It drops the INT_* definitions as they are no
longer used in irqs.h (moved to board-eb.h). It renames REALVIEW_*
macros to REALVIEW_EB_* or REALVIEW_EB11MP_* to distinguish between
standard EB and EB + the ARM11MPCore tile. The platform.h file contains
common definitions to the RealView platforms and it is only directly
included in board-*.h files.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-02-04 17:52:21 +00:00
Catalin Marinas a8655e83fc [ARM] 4814/1: RealView: Add broadcasting clockevents support for ARM11MPCore
This patch adds dummy local timers for each CPU so that the board clock
device is used to broadcast events to the other CPUs. The patch also
adds the declaration for the dummy_timer_setup function (the equivalent
of local_timer_setup when CONFIG_LOCAL_TIMERS is not set).

Due to the way clockevents work, the dummy timer on the first CPU has to
be registered before the board timer.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-02-04 17:52:19 +00:00
Catalin Marinas 0e0ba76926 [ARM] 4201/1: SMP barriers pair needed for the secondary boot process
In some situations, the pen_release store in platform_secondary_init()
may stay forever in the write buffer while the CPU is waiting on the
boot_lock to be released in boot_secondary(). The primary CPU could
never see the pen_release update without the barriers.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-02-16 13:06:56 +00:00
Catalin Marinas b3a1bde4db [ARM] 4108/2: Allow multiple GIC interrupt controllers in a system
The current implementation only assumes one GIC to be present in the
system. However, there are platforms with more than one cascaded interrupt
controllers (RealView/EB MPCore for example).

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-02-15 14:44:10 +00:00
Russell King 7bbb794031 [ARM] Fix SMP initialisation oops
A change to the SMP initialisation caused the following oops:

 CPU1: Booted secondary processor
 CPU1: D VIPT write-back cache
 CPU1: I cache: 32768 bytes, associativity 4, 32 byte lines, 256 sets
 CPU1: D cache: 32768 bytes, associativity 4, 32 byte lines, 256 sets
 <7>Calibrating delay loop... 83.14 BogoMIPS (lpj=415744)
 <1>Unable to handle kernel NULL pointer dereference at virtual address 0000001c
 ...
 PC is at enqueue_task+0x1c/0x64
 LR is at activate_task+0xcc/0xe4

SMP initialisation now requires cpu_possible_map to be initialised in
setup_arch().  Move this from smp_prepare_cpus() to smp_init_cpus()
and call it from our setup_arch() if CONFIG_SMP is enabled.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-02-16 11:08:09 +00:00
Russell King ce07d90aa8 [ARM] Fix arch-realview/system.h to use __io_address()
Move __io_address to arch-realview/hardware.h, drop core.h from platsmp.c
and localtimer.c, and include asm/io.h where required.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-11-16 14:38:19 +00:00
Russell King 2a98beb639 [ARM SMP] Add local timer support for Realview MPcore
Add platform specific parts for local timer support for the
Realview board.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-11-09 10:50:29 +00:00
Russell King 5d43045bcd [ARM SMP] Fix some sparse warnings in SMP code
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-11-08 10:44:46 +00:00
Russell King 862184fe01 [ARM SMP] Add Realview MPcore SMP support
Add SMP support for the MPcore tile fitted to the Realview ARM
platform.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-11-07 21:05:42 +00:00