kernel_optimize_test/arch/arm/kernel/unwind.c

492 lines
12 KiB
C
Raw Normal View History

/*
* arch/arm/kernel/unwind.c
*
* Copyright (C) 2008 ARM Limited
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
*
* Stack unwinding support for ARM
*
* An ARM EABI version of gcc is required to generate the unwind
* tables. For information about the structure of the unwind tables,
* see "Exception Handling ABI for the ARM Architecture" at:
*
* http://infocenter.arm.com/help/topic/com.arm.doc.subset.swdev.abi/index.html
*/
#ifndef __CHECKER__
#if !defined (__ARM_EABI__)
#warning Your compiler does not have EABI support.
#warning ARM unwind is known to compile only with EABI compilers.
#warning Change compiler or disable ARM_UNWIND option.
#elif (__GNUC__ == 4 && __GNUC_MINOR__ <= 2)
#warning Your compiler is too buggy; it is known to not compile ARM unwind support.
#warning Change compiler or disable ARM_UNWIND option.
#endif
#endif /* __CHECKER__ */
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/export.h>
#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/list.h>
#include <asm/stacktrace.h>
#include <asm/traps.h>
#include <asm/unwind.h>
/* Dummy functions to avoid linker complaints */
void __aeabi_unwind_cpp_pr0(void)
{
};
EXPORT_SYMBOL(__aeabi_unwind_cpp_pr0);
void __aeabi_unwind_cpp_pr1(void)
{
};
EXPORT_SYMBOL(__aeabi_unwind_cpp_pr1);
void __aeabi_unwind_cpp_pr2(void)
{
};
EXPORT_SYMBOL(__aeabi_unwind_cpp_pr2);
struct unwind_ctrl_block {
unsigned long vrs[16]; /* virtual register set */
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>
2011-12-05 16:39:59 +08:00
const unsigned long *insn; /* pointer to the current instructions word */
int entries; /* number of entries left to interpret */
int byte; /* current byte number in the instructions word */
};
enum regs {
#ifdef CONFIG_THUMB2_KERNEL
FP = 7,
#else
FP = 11,
#endif
SP = 13,
LR = 14,
PC = 15
};
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>
2011-12-05 16:39:59 +08:00
extern const struct unwind_idx __start_unwind_idx[];
static const struct unwind_idx *__origin_unwind_idx;
extern const struct unwind_idx __stop_unwind_idx[];
static DEFINE_SPINLOCK(unwind_lock);
static LIST_HEAD(unwind_tables);
/* Convert a prel31 symbol to an absolute address */
#define prel31_to_addr(ptr) \
({ \
/* sign-extend to 32 bits */ \
long offset = (((long)*(ptr)) << 1) >> 1; \
(unsigned long)(ptr) + offset; \
})
/*
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>
2011-12-05 16:39:59 +08:00
* Binary search in the unwind index. The entries are
* guaranteed to be sorted in ascending order by the linker.
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>
2011-12-05 16:39:59 +08:00
*
* start = first entry
* origin = first entry with positive offset (or stop if there is no such entry)
* stop - 1 = last entry
*/
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>
2011-12-05 16:39:59 +08:00
static const struct unwind_idx *search_index(unsigned long addr,
const struct unwind_idx *start,
const struct unwind_idx *origin,
const struct unwind_idx *stop)
{
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>
2011-12-05 16:39:59 +08:00
unsigned long addr_prel31;
pr_debug("%s(%08lx, %p, %p, %p)\n",
__func__, addr, start, origin, stop);
/*
* only search in the section with the matching sign. This way the
* prel31 numbers can be compared as unsigned longs.
*/
if (addr < (unsigned long)start)
/* negative offsets: [start; origin) */
stop = origin;
else
/* positive offsets: [origin; stop) */
start = origin;
/* prel31 for address relavive to start */
addr_prel31 = (addr - (unsigned long)start) & 0x7fffffff;
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>
2011-12-05 16:39:59 +08:00
while (start < stop - 1) {
const struct unwind_idx *mid = start + ((stop - start) >> 1);
/*
* As addr_prel31 is relative to start an offset is needed to
* make it relative to mid.
*/
if (addr_prel31 - ((unsigned long)mid - (unsigned long)start) <
mid->addr_offset)
stop = mid;
else {
/* keep addr_prel31 relative to start */
addr_prel31 -= ((unsigned long)mid -
(unsigned long)start);
start = mid;
}
}
if (likely(start->addr_offset <= addr_prel31))
return start;
else {
pr_warning("unwind: Unknown symbol address %08lx\n", addr);
return NULL;
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>
2011-12-05 16:39:59 +08:00
}
}
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>
2011-12-05 16:39:59 +08:00
static const struct unwind_idx *unwind_find_origin(
const struct unwind_idx *start, const struct unwind_idx *stop)
{
pr_debug("%s(%p, %p)\n", __func__, start, stop);
while (start < stop) {
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>
2011-12-05 16:39:59 +08:00
const struct unwind_idx *mid = start + ((stop - start) >> 1);
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>
2011-12-05 16:39:59 +08:00
if (mid->addr_offset >= 0x40000000)
/* negative offset */
start = mid + 1;
else
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>
2011-12-05 16:39:59 +08:00
/* positive offset */
stop = mid;
}
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>
2011-12-05 16:39:59 +08:00
pr_debug("%s -> %p\n", __func__, stop);
return stop;
}
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>
2011-12-05 16:39:59 +08:00
static const struct unwind_idx *unwind_find_idx(unsigned long addr)
{
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>
2011-12-05 16:39:59 +08:00
const struct unwind_idx *idx = NULL;
unsigned long flags;
pr_debug("%s(%08lx)\n", __func__, addr);
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>
2011-12-05 16:39:59 +08:00
if (core_kernel_text(addr)) {
if (unlikely(!__origin_unwind_idx))
__origin_unwind_idx =
unwind_find_origin(__start_unwind_idx,
__stop_unwind_idx);
/* main unwind table */
idx = search_index(addr, __start_unwind_idx,
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>
2011-12-05 16:39:59 +08:00
__origin_unwind_idx,
__stop_unwind_idx);
} else {
/* module unwind tables */
struct unwind_table *table;
spin_lock_irqsave(&unwind_lock, flags);
list_for_each_entry(table, &unwind_tables, list) {
if (addr >= table->begin_addr &&
addr < table->end_addr) {
idx = search_index(addr, table->start,
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>
2011-12-05 16:39:59 +08:00
table->origin,
table->stop);
/* Move-to-front to exploit common traces */
list_move(&table->list, &unwind_tables);
break;
}
}
spin_unlock_irqrestore(&unwind_lock, flags);
}
pr_debug("%s: idx = %p\n", __func__, idx);
return idx;
}
static unsigned long unwind_get_byte(struct unwind_ctrl_block *ctrl)
{
unsigned long ret;
if (ctrl->entries <= 0) {
pr_warning("unwind: Corrupt unwind table\n");
return 0;
}
ret = (*ctrl->insn >> (ctrl->byte * 8)) & 0xff;
if (ctrl->byte == 0) {
ctrl->insn++;
ctrl->entries--;
ctrl->byte = 3;
} else
ctrl->byte--;
return ret;
}
/*
* Execute the current unwind instruction.
*/
static int unwind_exec_insn(struct unwind_ctrl_block *ctrl)
{
unsigned long insn = unwind_get_byte(ctrl);
pr_debug("%s: insn = %08lx\n", __func__, insn);
if ((insn & 0xc0) == 0x00)
ctrl->vrs[SP] += ((insn & 0x3f) << 2) + 4;
else if ((insn & 0xc0) == 0x40)
ctrl->vrs[SP] -= ((insn & 0x3f) << 2) + 4;
else if ((insn & 0xf0) == 0x80) {
unsigned long mask;
unsigned long *vsp = (unsigned long *)ctrl->vrs[SP];
int load_sp, reg = 4;
insn = (insn << 8) | unwind_get_byte(ctrl);
mask = insn & 0x0fff;
if (mask == 0) {
pr_warning("unwind: 'Refuse to unwind' instruction %04lx\n",
insn);
return -URC_FAILURE;
}
/* pop R4-R15 according to mask */
load_sp = mask & (1 << (13 - 4));
while (mask) {
if (mask & 1)
ctrl->vrs[reg] = *vsp++;
mask >>= 1;
reg++;
}
if (!load_sp)
ctrl->vrs[SP] = (unsigned long)vsp;
} else if ((insn & 0xf0) == 0x90 &&
(insn & 0x0d) != 0x0d)
ctrl->vrs[SP] = ctrl->vrs[insn & 0x0f];
else if ((insn & 0xf0) == 0xa0) {
unsigned long *vsp = (unsigned long *)ctrl->vrs[SP];
int reg;
/* pop R4-R[4+bbb] */
for (reg = 4; reg <= 4 + (insn & 7); reg++)
ctrl->vrs[reg] = *vsp++;
if (insn & 0x80)
ctrl->vrs[14] = *vsp++;
ctrl->vrs[SP] = (unsigned long)vsp;
} else if (insn == 0xb0) {
if (ctrl->vrs[PC] == 0)
ctrl->vrs[PC] = ctrl->vrs[LR];
/* no further processing */
ctrl->entries = 0;
} else if (insn == 0xb1) {
unsigned long mask = unwind_get_byte(ctrl);
unsigned long *vsp = (unsigned long *)ctrl->vrs[SP];
int reg = 0;
if (mask == 0 || mask & 0xf0) {
pr_warning("unwind: Spare encoding %04lx\n",
(insn << 8) | mask);
return -URC_FAILURE;
}
/* pop R0-R3 according to mask */
while (mask) {
if (mask & 1)
ctrl->vrs[reg] = *vsp++;
mask >>= 1;
reg++;
}
ctrl->vrs[SP] = (unsigned long)vsp;
} else if (insn == 0xb2) {
unsigned long uleb128 = unwind_get_byte(ctrl);
ctrl->vrs[SP] += 0x204 + (uleb128 << 2);
} else {
pr_warning("unwind: Unhandled instruction %02lx\n", insn);
return -URC_FAILURE;
}
pr_debug("%s: fp = %08lx sp = %08lx lr = %08lx pc = %08lx\n", __func__,
ctrl->vrs[FP], ctrl->vrs[SP], ctrl->vrs[LR], ctrl->vrs[PC]);
return URC_OK;
}
/*
* Unwind a single frame starting with *sp for the symbol at *pc. It
* updates the *pc and *sp with the new values.
*/
int unwind_frame(struct stackframe *frame)
{
unsigned long high, low;
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>
2011-12-05 16:39:59 +08:00
const struct unwind_idx *idx;
struct unwind_ctrl_block ctrl;
/* only go to a higher address on the stack */
low = frame->sp;
high = ALIGN(low, THREAD_SIZE);
pr_debug("%s(pc = %08lx lr = %08lx sp = %08lx)\n", __func__,
frame->pc, frame->lr, frame->sp);
if (!kernel_text_address(frame->pc))
return -URC_FAILURE;
idx = unwind_find_idx(frame->pc);
if (!idx) {
pr_warning("unwind: Index not found %08lx\n", frame->pc);
return -URC_FAILURE;
}
ctrl.vrs[FP] = frame->fp;
ctrl.vrs[SP] = frame->sp;
ctrl.vrs[LR] = frame->lr;
ctrl.vrs[PC] = 0;
if (idx->insn == 1)
/* can't unwind */
return -URC_FAILURE;
else if ((idx->insn & 0x80000000) == 0)
/* prel31 to the unwind table */
ctrl.insn = (unsigned long *)prel31_to_addr(&idx->insn);
else if ((idx->insn & 0xff000000) == 0x80000000)
/* only personality routine 0 supported in the index */
ctrl.insn = &idx->insn;
else {
pr_warning("unwind: Unsupported personality routine %08lx in the index at %p\n",
idx->insn, idx);
return -URC_FAILURE;
}
/* check the personality routine */
if ((*ctrl.insn & 0xff000000) == 0x80000000) {
ctrl.byte = 2;
ctrl.entries = 1;
} else if ((*ctrl.insn & 0xff000000) == 0x81000000) {
ctrl.byte = 1;
ctrl.entries = 1 + ((*ctrl.insn & 0x00ff0000) >> 16);
} else {
pr_warning("unwind: Unsupported personality routine %08lx at %p\n",
*ctrl.insn, ctrl.insn);
return -URC_FAILURE;
}
while (ctrl.entries > 0) {
int urc = unwind_exec_insn(&ctrl);
if (urc < 0)
return urc;
if (ctrl.vrs[SP] < low || ctrl.vrs[SP] >= high)
return -URC_FAILURE;
}
if (ctrl.vrs[PC] == 0)
ctrl.vrs[PC] = ctrl.vrs[LR];
/* check for infinite loop */
if (frame->pc == ctrl.vrs[PC])
return -URC_FAILURE;
frame->fp = ctrl.vrs[FP];
frame->sp = ctrl.vrs[SP];
frame->lr = ctrl.vrs[LR];
frame->pc = ctrl.vrs[PC];
return URC_OK;
}
void unwind_backtrace(struct pt_regs *regs, struct task_struct *tsk)
{
struct stackframe frame;
register unsigned long current_sp asm ("sp");
pr_debug("%s(regs = %p tsk = %p)\n", __func__, regs, tsk);
if (!tsk)
tsk = current;
if (regs) {
frame.fp = regs->ARM_fp;
frame.sp = regs->ARM_sp;
frame.lr = regs->ARM_lr;
/* PC might be corrupted, use LR in that case. */
frame.pc = kernel_text_address(regs->ARM_pc)
? regs->ARM_pc : regs->ARM_lr;
} else if (tsk == current) {
frame.fp = (unsigned long)__builtin_frame_address(0);
frame.sp = current_sp;
frame.lr = (unsigned long)__builtin_return_address(0);
frame.pc = (unsigned long)unwind_backtrace;
} else {
/* task blocked in __switch_to */
frame.fp = thread_saved_fp(tsk);
frame.sp = thread_saved_sp(tsk);
/*
* The function calling __switch_to cannot be a leaf function
* so LR is recovered from the stack.
*/
frame.lr = 0;
frame.pc = thread_saved_pc(tsk);
}
while (1) {
int urc;
unsigned long where = frame.pc;
urc = unwind_frame(&frame);
if (urc < 0)
break;
dump_backtrace_entry(where, frame.pc, frame.sp - 4);
}
}
struct unwind_table *unwind_table_add(unsigned long start, unsigned long size,
unsigned long text_addr,
unsigned long text_size)
{
unsigned long flags;
struct unwind_table *tab = kmalloc(sizeof(*tab), GFP_KERNEL);
pr_debug("%s(%08lx, %08lx, %08lx, %08lx)\n", __func__, start, size,
text_addr, text_size);
if (!tab)
return tab;
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>
2011-12-05 16:39:59 +08:00
tab->start = (const struct unwind_idx *)start;
tab->stop = (const struct unwind_idx *)(start + size);
tab->origin = unwind_find_origin(tab->start, tab->stop);
tab->begin_addr = text_addr;
tab->end_addr = text_addr + text_size;
spin_lock_irqsave(&unwind_lock, flags);
list_add_tail(&tab->list, &unwind_tables);
spin_unlock_irqrestore(&unwind_lock, flags);
return tab;
}
void unwind_table_del(struct unwind_table *tab)
{
unsigned long flags;
if (!tab)
return;
spin_lock_irqsave(&unwind_lock, flags);
list_del(&tab->list);
spin_unlock_irqrestore(&unwind_lock, flags);
kfree(tab);
}