Commit Graph

11 Commits

Author SHA1 Message Date
Uwe Kleine-König ddf5a25c5f ARM: unwinder: fix bisection to find origin in .idx section
The bisection implemented in unwind_find_origin() stopped to early.  If
there is only a single entry left to check the original code just took
the end point as origin which might be wrong.

This was introduced in commit de66a97901 ("ARM: 7187/1: fix unwinding
for XIP kernels").

Reported-and-tested-by: Nick Bowler <nbowler@elliptictech.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-12-15 14:02:19 -08:00
Uwe Kleine-König de66a97901 ARM: 7187/1: fix unwinding for XIP kernels
The linker places the unwind tables in readonly sections. So when using
an XIP kernel these are located in ROM and cannot be modified.
For that reason the current approach to convert the relative offsets in
the unwind index to absolute addresses early in the boot process doesn't
work with XIP.

The offsets in the unwind index section are signed 31 bit numbers and
the structs are sorted by this offset. So it first has offsets between
0x40000000 and 0x7fffffff (i.e. the negative offsets) and then offsets
between 0x00000000 and 0x3fffffff. When seperating these two blocks the
numbers are sorted even when interpreting the offsets as unsigned longs.

So determine the first non-negative entry once and track that using the
new origin pointer. The actual bisection can then use a plain unsigned
long comparison. The only thing that makes the new bisection more
complicated is that the offsets are relative to their position in the
index section, so the key to search needs to be adapted accordingly in
each step.

Moreover several consts are added to catch future writes and rename the
member "addr" of struct unwind_idx to "addr_offset" to better match the
new semantic. (This has the additional benefit of breaking eventual
users at compile time to make them aware of the change.)

In my tests the new algorithm was a tad faster than the original and has
the additional upside of not needing the initial conversion and so saves
some boot time and it's possible to unwind even earlier.

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-12-06 11:16:13 +00:00
Paul Gortmaker ecea4ab6d3 arm: convert core files from module.h to export.h
Many of the core ARM kernel files are not modules, but just
including module.h for exporting symbols.  Now these files can
use the lighter footprint export.h for this role.

There are probably lots more, but ARM files of mach-* and plat-*
don't get coverage via a simple yesconfig build.  They will have
to be cleaned up and tested via using their respective configs.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
2011-10-31 19:30:49 -04:00
Will Deacon d33aadbf8e ARM: 6468/1: backtrace: fix calculation of thread stack base
When unwinding stack frames we must take care not to unwind
areas of memory that lie outside of the known extent of the stack.

This patch fixes an incorrect calculation of the stack base where
THREAD_SIZE is added to the stack pointer after it has already
been aligned to this value. Since the ALIGN macro performs this
addition internally, we end up overshooting the base by 8k.

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-11-07 16:12:37 +00:00
Phil Carmody 5333a3de3c ARM: 6341/1: unwind - optimise linked-list searches for modules
With several sections per module, and dozens of modules, the
searches down the linked list of sections would dominate the
lookup time, dwarfing any savings from the binary search
within the section.

A simple move-to-front optimisation exploits the commonality
of the code paths taken, and in simple real-world tests reduces
the number of steps in the search to barely more than 1.

Signed-off-by: Phil Carmody <ext-phil.2.carmody@nokia.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-09-02 15:31:36 +01:00
Alexander Shishkin 830703c766 ARM: 6140/1: silence a bogus sparse warning in unwind.c
The check for compiler which is supposed to miscompile unwind tables
clearly has nothing to do with sparse (which does not define necessary
macros anyway), so simply silence it.

Signed-off-by: Alexander Shishkin <virtuoso@slind.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-05-24 20:42:03 +01:00
Laurent Pinchart 51d47999b9 ARM: 5977/1: arm: Enable backtrace printing on oops when PC is corrupted
If PC points outside kernel text, start printing the backtrace at LR
instead.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-03-07 10:22:00 +00:00
Claudio Scordino 6603a4fd51 ARM: 5776/1: Check compiler version and EABI support when adding ARM unwind support.
ARM unwind is known to compile only with EABI and not-buggy compilers.
The problem is not the unwinding information but the -fno-frame-pointer
option added as a result of !CONFIG_FRAME_POINTER.  Now we check the
compiler and raise a #warning in case of wrong compiler.

Signed-off-by: Claudio Scordino <claudio@evidence.eu.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-11-02 16:59:45 +00:00
Catalin Marinas b86040a59f Thumb-2: Implementation of the unified start-up and exceptions code
This patch implements the ARM/Thumb-2 unified kernel start-up and
exception handling code.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 12:32:54 +01:00
Catalin Marinas c894ed6956 [ARM] 5558/1: Add extra checks to ARM unwinder to avoid tracing corrupt stacks
There are situations where the unwinder goes beyond stack boundaries and
unwinds random data. This patch moves the stack boundaries check after
the unwind_exec_insn() call and adds an extra check for possible
infinite loops (like "mov pc, lr" with pc == lr).

The patch also fixes a bug in the unwind instructions interpreter. The
0xb0 instruction can only set PC to LR if this wasn't already set by
a previous instruction (this is used on exceptions taken while in kernel
mode where svc_entry is annotated with ".save {r0 - pc}").

Tested-by: Tony Lindgren <tony@atomide.com>

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-06-19 16:44:22 +01:00
Catalin Marinas bff595c15c [ARM] 5383/2: unwind: Add core support for ARM stack unwinding
This patch adds the main functionality for parsing the stack unwinding
information generated by the ARM EABI toolchains. The unwinding
information consists of an index with a pair of words per function and a
table with unwinding instructions. For more information, see "Exception
Handling ABI for the ARM Architecture" at:

http://infocenter.arm.com/help/topic/com.arm.doc.subset.swdev.abi/index.html

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-02-19 11:26:24 +00:00