Make sure our interrupt entry code with exact hardware errors handles
anomaly 05000283 (infinite stall in system MMR kill) so we don't stall
while under load.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The majority of the time we are returning to user space, it is not in the
fixed atomic code region. So rather than branch to a function where we
check the PC and return, do the check inline and branch only when needed.
Also, tweak some of the fixed code handling based on assumptions we are
aware of but cannot be expressed in C.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
No one uses these functions, and some are duplicate of existing C code. So
just punt the whole thing.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The handling of updating the [DI]MEM_CONTROL MMRs does not follow proper
sync procedures as laid out in the Blackfin programming manual. So rather
than audit/fix every call location, create helper functions that do the
right things in order to safely update these MMRs. Then convert all call
sites to use these new helper functions.
While we're fixing the code, drop the workaround for anomaly 05000125 as
that anomaly applies to old versions of silicon that we do not support.
Signed-off-by: Yi Li <yi.li@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Since the hardware only provides reporting for the last exception handled,
and the values are valid only when executing the exception handler, we
need to save the context for reporting at a later point. While we do this
for one exception, it doesn't work properly when handling a second one as
the original exception is clobbered by the double fault. So when double
fault debugging is enabled, create a dedicated shadow of these values and
save/restore out of there. Now the crash report properly displays the
first exception as well as the second one.
Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Allow hardware errors to be caught during early portions of booting, and
leave something in the shadow console that people can use to debug their
system with (to be printed out by the bootloader on next reset).
This enables the hardare error interrupts in head.S, allowing us to find
hardware errors when they happen (well, as much as you can with a hardware
error) and prints out the trace if it is enabled. This will catch errors
(like booting the wrong image on a 533) which previously resulted in a
infinite loop/hang, as well as random hardware errors before before
setup_arch().
To disable this debug only feature - turn off EARLY_PRINTK.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Cleanup is performed in two ways:
- remove extraneous updates of IPEND[4] w/ CONFIG_IPIPE,
and document remaining use.
- substitute pop-reg-from-stack instructions with plain SP fixups in
all save-RETI-then-discard patterns.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The purpose of the EVT14 handler may depend on whether CONFIG_IPIPE is
enabled, albeit its implementation can be the same in both cases. When
the interrupt pipeline is enabled, EVT14 can be used to raise the core
priority level for the running code; when CONFIG_IPIPE is off, EVT14
can be used to lower this level before running softirq handlers.
Rename evt14_softirq to evt_evt14 to pick an identifier that fits
both, which allows to reuse the same vector setup code as well.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
We handle many exceptions at EVT5 (hardware error level) so that we can
catch exceptions in our exception handling code. Today - if the global
interrupt enable bit (IPEND[4]) is set (interrupts disabled) our trap
handling code goes into a infinite loop, since we need interrupts to be
on to defer things to EVT5.
Normal kernel code should not trigger this for any reason as IPEND[4] gets
cleared early (when doing an interrupt context save) and the kernel stack
there should be sane (or something much worse is happening in the system).
But there have been a few times where this has happened, so this change
makes sure we dump a proper crash message even when things have gone south.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Check that the result of kmalloc is not NULL before passing it to other
functions.
In the first two cases, the new code returns -ENOMEM, which seems
compatible with what is done for similar functions for other architectures.
In the last two cases, the new code fails silently, ie just returns,
because the function has void return type.
The semantic match that finds this problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)
// <smpl>
@@
expression *x;
identifier f;
constant char *C;
@@
x = \(kmalloc\|kcalloc\|kzalloc\)(...);
... when != x == NULL
when != x != NULL
when != (x || ...)
(
kfree(x)
|
f(...,C,...,x,...)
|
*f(...,x,...)
|
*x->f
)
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
On Blackfin SMP, a per-cpu loops_per_jiffy is pointless since both cores
always run at the same CCLK. In addition, the current implementation has
flaws since the main consumer for loops_per_jiffy (asm/delay.h) uses the
global kernel loops_per_jiffy and not the per_cpu one. So punt all of the
per-cpu handling and go back to the global shared one.
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Improve the assembly with a few explanatory comments and use symbolic
defines rather than numeric values for bit positions.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The current cache options don't really represent the hardware features.
They end up setting different aspects of the hardware so that the end
result is to turn on/off the cache. Unfortunately, when we hit cache
problems with the hardware, it's difficult to test different settings to
root cause the problem. The current settings also don't cleanly allow for
different caching behaviors with different regions of memory.
So split the configure options such that they properly reflect the settings
that are applied to the hardware.
Signed-off-by: Jie Zhang <jie.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The irq_panic function is only used when CONFIG_DEBUG_ICACHE_CHECK is
enabled, so move the conditional build to the Makefile rather than
wrapping all of the contents of the file.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Convert to test_bit() as that is what pretty much everyone uses and allows
us to migrate asm/bitops.h to the asm-generic version.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
When a low priority interrupt (like ethernet) is triggered between 2 high
priority IPI messages, a deadlock in disable_irq() is hit by the second
IPI handler. This is because the second IPI message is queued within the
first IPI handler, but the handler doesn't process all messages, and new
ones are inserted rather than appended. So now we process all the pending
messages, and append new ones to the pending list.
URL: http://blackfin.uclinux.org/gf/tracker/5226
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The KGDB code uses this when switching processors to make sure the icache
is in a valid state.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
We don't need to handle CPLB protection violations unless we are running
with the MPU on. Fix the entry code to call common trap_c, and remove the
code which is never run. This allows the traps test suite to run on older
boards with the MPU disabled.
URL: http://blackfin.uclinux.org/gf/tracker/5129
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Rather than having to maintain a hard coded list of Blackfin variants, use
the SIC defines themselves. This fixes build problems on BF51x/BF538 under
some configurations as they were missing from one of the lists.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Commit 6b3087c6 (which introduced Blackfin SMP) broke command line passing
when the DEBUG_DOUBLEFAULT config option was enabled. Switch the code to
using a scratch register and not R7 which holds the command line.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The Per-processor Data Area isn't actually reserved by this function, and
all it ended up doing was issuing a printk(), so punt it.
Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Hardware errors on the Blackfin architecture are queued by nature of the
hardware design. Things that could generate a hardware level queue up at
the system interface and might not process until much later, at which
point the system would send a notification back to the core.
As such, it is possible for user space code to do something that would
trigger a hardware error, but have it delay long enough for the process
context to switch. So when the hardware error does signal, we mistakenly
evaluate it as a different process or as kernel context and panic (erp!).
This makes it pretty difficult to find the offending context. But wait,
there is good news somewhere.
By forcing a SSYNC in the interrupt entry, we force all pending queues at
the system level to be processed and all hardware errors to be signaled.
Then we check the current interrupt state to see if the hardware error is
now signaled. If so, we re-queue the current interrupt and return thus
allowing the higher priority hardware error interrupt to process properly.
Since we haven't done any other context processing yet, the right context
will be selected and killed. There is still the possibility that the
exact offending instruction will be unknown, but at least we'll have a
much better idea of where to look.
The downside of course is that this causes system-wide syncs at every
interrupt point which results in significant performance degradation.
Since this situation should not occur in any properly configured system
(as hardware errors are triggered by things like bad pointers), make it a
debug configuration option and disable it by default.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
When possible, work around anomaly 05000220 (external memory is write
back cached, but L2 is not cached). If not possible, detect the
conditions at build time and reject any qualifying configurations.
Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
For systems where the core cycles are not a usable tick source (like SMP
or cycles gets updated), enable gptimer0 as an alternative.
Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Add some notes for anomaly 05000120 to make sure we work around it.
Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Detect and reject operating conditions for anomaly 05000274 since the
problem cannot be worked around in software.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Workaround anomaly 05000227 by only using the scratch pad for stack when
absolutely necessary. The core code which reprograms clocks really only
touches MMRs directly with constants.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Simplify the do_flush macro now that we don't need to take into account
a second instruction being used together.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
ipipe-2.6.28.9-blackfin-git95aafe6.patch
Singed-off-by: Philippe Gerum <rpm@xenomai.org>
Signed-off-by: Yi Li <yi.li@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Make sure we flush all data caches and their write buffers before flushing
icache, otherwise random edge cases could crop up where stale data is read
into icache from external memory. As fallout, punt the combined icache +
dcache flush function since we cannot safely do them back to back -- the
SSYNC is needed between the dcache flush and the icache flush.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Move exception stack mess from entry.S to init.c to fix link failure when
CONFIG_EXCEPTION_L1_SCRATCH is in use.
Signed-off-by: Graf Yang <graf.yang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Random read/write errors are a bad thing - so don't let anyone
(including the test bench) run on something we know is bad.
Signed-off-by: Robin Getz <rgetz@blackfin.uclinux.org>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Run IFLUSH twice to avoid loading wrong instruction
after invalidating icache and following sequence is met.
1) The one instruction address is cached in the icache.
2) This instruction in SDRAM is changed.
3) IFLASH[P0] is executed only once in lackfin_icache_flush_range().
4) This instruction is executed again, but not the changed new one.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
IMHO the setting should depend on ANOMALY_05000305 which is about the
availability of the bit, not ANOMALY_05000265 which only describes the
SPORT sensitivity to noise (checked for BF561 only, though).
If that's not true for other BF variants, maybe the definition of
ANOMALY_05000265 for BF561 should be changed to '(1)' instead.
Signed-off-by: Enrik Berkhan <Enrik.Berkhan@ge.com>
Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Robin Getz