I have had four seperate system lockups attributable to this exact problem
in two days of testing. Instead of trying to handle all the weird end
cases and wrap, how about changing it to look for exactly what we appear
to want.
The following patch removes a couple races in setup_APIC_timer. One occurs
when the HPET advances the COUNTER past the T0_CMP value between the time
the T0_CMP was originally read and when COUNTER is read. This results in
a delay waiting for the counter to wrap. The other results from the counter
wrapping.
This change takes a snapshot of T0_CMP at the beginning of the loop and
simply loops until T0_CMP has changed (a tick has happened).
<later>
I have one small concern about the patch. I am not sure it meets the intent
as well as it should. I think we are trying to match APIC timer interrupts up
with the hpet counter increment. The event which appears to be disturbing
this loop in our test environment is the NMI watchdog. What we believe has
been happening with the existing code is the setup_APIC_timer loop has read
the CMP value, and the NMI watchdog code fires for the first time. This
results in a series of icache miss slowdowns and by the time we get back to
things it has wrapped.
I think this code is trying to get the CMP as close to the counter value as
possible. If that is the intent, maybe we should really be testing against a
"window" around the CMP. Something like COUNTER = CMP+/2. It appears COUNTER
should get advanced every 89nSec (IIRC). The above seems like an unreasonably
small window, but may be necessary. Without documentation, I am not sure of
the original intent with this code.
In summary, this code fixes my boot hangs, but since I am not certain of the
intent of the existing code, I am not certain this has not introduced new bugs
or unexpected behaviors.
Signed-off-by: Robin Holt <holt@sgi.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Vojtech Pavlik <vojtech@suse.cz>
Cc: "Aaron Durbin" <adurbin@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
My "slices" address space management code that was added in the 2.6.22
implementation of get_unmapped_area() doesn't properly check that the
size is a multiple of the requested page size. This allows userland to
create VMAs that aren't a multiple of the huge page size with hugetlbfs
(since hugetlbfs entirely relies on get_unmapped_area() to do that
checking) which leads to a kernel BUG() when such areas are torn down.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
powerpc has a couple of bugs in the usage of dma_masks that tend to
break when drivers explicitly try to set a 32-bit mask for example.
First, the code that generates the pci devices from the OF device-tree
doesn't initialize the mask properly, then our implementation of
set_dma_mask() was trying to validate the -previous- mask value, not the
one passed in as an argument.
This fixes these problems.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This reverts commit 3baee95595.
That commit was a mistake from the start; I added mdio type to the
bus scan list early on in my ucc_geth migrate to phylib development,
which is just pure wrong (the ucc_geth_mii driver creates the mii
bus and the PHY layer handles PHY enumeration without translation).
This follows on from commit 77926826f3:
Revert "[POWERPC] Don't complain if size-cells == 0 in prom_parse()"
which was basically trying to hide a symptom of the original mistake
this revert fixes.
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Remove the Kconfig message that indicates the PS3 platform support is
incomplete.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Fix the PS3 storage probe logic to properly find device regions on cold
startup.
o Change the storage probe event mask from notify_device_ready
to notify_region_update.
o Improve the storage probe error handling.
o Change ps3_storage_wait_for_device() to use a temporary variable to hold
the buffer address.
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch moves affinity initialization code from spu_base.c to a
new spu_management_of_ops function (init_affinity), which is empty
in the case of PS3. This fixes a linking problem that was happening
when compiling for PS3.
Also, some small code style changes were made.
Signed-off-by: Andre Detsch <adetsch@br.ibm.com>
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Acked-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
We were getting a duplicate entry in the SLB shadow buffer in
slb_flush_and_rebolt() if the kernel stack was in the same segment
as PAGE_OFFSET, which on POWER6 causes the hypervisor to terminate
the partition with an error. This fixes it.
Also we were not creating an SLB entry (or an SLB shadow buffer
entry) for the kernel stack on secondary CPUs when starting the
CPU. This isn't a major problem, since an appropriate entry will
be created on demand, but this fixes that also for consistency.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Every time a cpu is added via hotplug, we allocate the per-cpu MONDO
queues but we never free them up. Freeing isn't easy since the first
cpu gets this memory from bootmem.
Therefore, the simplest thing to do to fix this bug is to allocate the
queues for all possible cpus at boot time.
Signed-off-by: David S. Miller <davem@davemloft.net>
Check the cpu type in the OBP device tree before committing to
using the optimized Niagara memcpy and memset implementation.
If we don't recognize the cpu type, use a completely generic
version.
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch addresses some issues in x86/x86-64 acpi-cpufreq driver:
1. Current memory allocation for acpi_perf_data is actually open-coded
alloc_percpu(). The patch defines and handles acpi_perf_data as percpu
data. The code will be cleaner and easier to be maintained with this
change.
2. Won't load driver in acpi_cpufreq_early_init() failure case.
3. Add __init for acpi_cpufreq_early_init().
Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>
Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Dave Jones <davej@codemonkey.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Len Brown <len.brown@intel.com>
Now that __initmv references the machvec section unconditionally
there have been cases where folks have been mistakenly flagging
non-machvec structures with the machvec section attribute (presumably
to shut up modpost). This leads to obscure breakage in earlyprintk
amongst other places and is rather non-obvious.
Add a simple sanity check to try and catch __initmv misuse and
panic early.
Reported-by: Markus Brunner <super.firetwister@gmail.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Remove cruft - CONFIG_DEBUG_SERIAL_EARLY_INIT didn't work that well, and
DEBUG_KERNEL_START was just implmented poorly. Will replace with a new
checkin.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Fix a problem reported in the forums - libstdc++ can call writev with an
iovec containing { NULL, 0 }, which works fine on i686-linux, but fails on
Blackfin. Fixed by allowing size 0 transfers to/from userspace regardless
of the address.
Signed-off-by: Bernd Schmidt <bernd.schmidt@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Now that we have moved head.S into the init section, the load
address is not safe to point to as a workaround for ANOMALY 05000281
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
show_mem can not be marked as init, since it is called during
OOM condition from /mm/oom_kill.c:out_of_memory() and
/mm/page_alloc.c:__alloc_pages()
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
- flush/inv the correct range
- dmacopy test failed when policy is write_back - invalidate before dma
http://blackfin.uclinux.org/gf/project/uclinux-dist/tracker/?action=TrackerItemEdit&tracker_item_id=3367
It's the cache invalidate what is causing the issue.
There is no invalidate only instruction it's always: FLUSHINV
So when we "invalidate" after the DMA we might (do) overwrite freshly
dma'ed data by dirty Cache WB content.
Fixed by moving the "invalidate" at the beginning of dma_memcpy.
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Comply with revised Anomaly Workarounds for BF533 05000311 and BF561 05000323
accoring to BF533 anomaly sheet Rev. A 09/04/07
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Print out debug info, as early as possible - even before the
kernel initializes the interrupt vectors. Now we can print out debug
messages almost anytime during the boot process.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Also ensure that the traps_c code doesn't cause a double fault, by
sending a signal to a faulting kernel before the memory subsystem
is fully initialized, by printing out the error message before sending
the signal.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Today when a double fault happens (exception during an exception
handling event), we go into an endless loop, with nothing comming out
the UART. With this patch, we actually see that we have commited a
double fault event
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
This allows debugging of problems which happen eary in the kernel
boot process (after bootargs are parsed, but before serial subsystem
is fully initialized)
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Now that there is a generic GPIO driver framework
remove GPIO register unified name space legacy support.
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
- Update gpio_request to allow multiple request with the same signature (label)
- Use generic GPIO API where applicable
- Update generic board support form stamp board
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
rewrite our reboot code in C rather than assembly to be like
other architectures and to allow board maintainers to define
custom behavior
Signed-off-by: Mike Frysinger <michael.frysinger@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Break up our .init into separate section like all other ports do and
so that we dont mix text and data (causes disassembly headaches as
pointed out by Robin)
Cc: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <michael.frysinger@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
add an exception request/free api similar to the interrupt request/fre
api so people can utilize the free software based exceptions for their
own purposes
Signed-off-by: Mike Frysinger <michael.frysinger@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
- allow people to select the feature that is unavailable to the kernel: NMI, JTAG, or CYCLES.
- change default NMI handler to simply dump hardware trace buffer.
- remove default NMI handler completely as calling into kernel code is not safe
move example handler to wiki so people dont haphazardly copy and paste this stuff thinking its safe
Signed-off-by: Mike Frysinger <michael.frysinger@analog.com>
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
