kernel_optimize_test/arch/arm
Ivan Djelic 455bd4c430 ARM: 7668/1: fix memset-related crashes caused by recent GCC (4.7.2) optimizations
Recent GCC versions (e.g. GCC-4.7.2) perform optimizations based on
assumptions about the implementation of memset and similar functions.
The current ARM optimized memset code does not return the value of
its first argument, as is usually expected from standard implementations.

For instance in the following function:

void debug_mutex_lock_common(struct mutex *lock, struct mutex_waiter *waiter)
{
	memset(waiter, MUTEX_DEBUG_INIT, sizeof(*waiter));
	waiter->magic = waiter;
	INIT_LIST_HEAD(&waiter->list);
}

compiled as:

800554d0 <debug_mutex_lock_common>:
800554d0:       e92d4008        push    {r3, lr}
800554d4:       e1a00001        mov     r0, r1
800554d8:       e3a02010        mov     r2, #16 ; 0x10
800554dc:       e3a01011        mov     r1, #17 ; 0x11
800554e0:       eb04426e        bl      80165ea0 <memset>
800554e4:       e1a03000        mov     r3, r0
800554e8:       e583000c        str     r0, [r3, #12]
800554ec:       e5830000        str     r0, [r3]
800554f0:       e5830004        str     r0, [r3, #4]
800554f4:       e8bd8008        pop     {r3, pc}

GCC assumes memset returns the value of pointer 'waiter' in register r0; causing
register/memory corruptions.

This patch fixes the return value of the assembly version of memset.
It adds a 'mov' instruction and merges an additional load+store into
existing load/store instructions.
For ease of review, here is a breakdown of the patch into 4 simple steps:

Step 1
======
Perform the following substitutions:
ip -> r8, then
r0 -> ip,
and insert 'mov ip, r0' as the first statement of the function.
At this point, we have a memset() implementation returning the proper result,
but corrupting r8 on some paths (the ones that were using ip).

Step 2
======
Make sure r8 is saved and restored when (! CALGN(1)+0) == 1:

save r8:
-       str     lr, [sp, #-4]!
+       stmfd   sp!, {r8, lr}

and restore r8 on both exit paths:
-       ldmeqfd sp!, {pc}               @ Now <64 bytes to go.
+       ldmeqfd sp!, {r8, pc}           @ Now <64 bytes to go.
(...)
        tst     r2, #16
        stmneia ip!, {r1, r3, r8, lr}
-       ldr     lr, [sp], #4
+       ldmfd   sp!, {r8, lr}

Step 3
======
Make sure r8 is saved and restored when (! CALGN(1)+0) == 0:

save r8:
-       stmfd   sp!, {r4-r7, lr}
+       stmfd   sp!, {r4-r8, lr}

and restore r8 on both exit paths:
        bgt     3b
-       ldmeqfd sp!, {r4-r7, pc}
+       ldmeqfd sp!, {r4-r8, pc}
(...)
        tst     r2, #16
        stmneia ip!, {r4-r7}
-       ldmfd   sp!, {r4-r7, lr}
+       ldmfd   sp!, {r4-r8, lr}

Step 4
======
Rewrite register list "r4-r7, r8" as "r4-r8".

Signed-off-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Dirk Behme <dirk.behme@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2013-03-07 16:14:22 +00:00
..
boot ARM: 7666/1: decompressor: add -mno-single-pic-base for building the decompressor 2013-03-06 23:45:42 +00:00
common ARM: GIC: fix GIC cpumask initialization 2013-02-08 10:17:22 +00:00
configs arm-soc: mvebu platform changes 2013-02-28 20:09:24 -08:00
crypto
include ARM: 7665/1: Wire up kcmp syscall 2013-03-03 22:54:18 +00:00
kernel ARM: 7667/1: perf: Fix section mismatch on armpmu_init() 2013-03-06 23:45:43 +00:00
kvm ARM: KVM: fix compilation after removal of user_alloc from struct kvm_memory_slot 2013-02-25 11:48:41 +02:00
lib ARM: 7668/1: fix memset-related crashes caused by recent GCC (4.7.2) optimizations 2013-03-07 16:14:22 +00:00
mach-at91 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-bcm Merge remote-tracking branches 'arm-soc/irqchip/gic-vic-move' and 'arm-soc/timer/cleanup' into soc 2013-01-24 17:57:20 +09:00
mach-bcm2835 arm-soc: soc-specific updates 2013-02-21 15:27:22 -08:00
mach-clps711x The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-cns3xxx The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-davinci Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-dove arm-soc: mvebu platform changes 2013-02-28 20:09:24 -08:00
mach-ebsa110 The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-ep93xx The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-exynos Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-footbridge The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-gemini The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-h720x The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-highbank Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-imx arm-soc: i.MX DT changes 2013-02-28 19:59:34 -08:00
mach-integrator Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-iop13xx Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-iop32x The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-iop33x The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-ixp4xx The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-kirkwood arm-soc: mvebu platform changes 2013-02-28 20:09:24 -08:00
mach-ks8695 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-l7200/include/mach
mach-lpc32xx The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-mmp Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-msm Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-mv78xx0 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-mvebu arm-soc: mvebu platform changes 2013-02-28 20:09:24 -08:00
mach-mxs arm-soc: device tree conversions 2013-02-21 15:38:49 -08:00
mach-netx ARM: 7642/1: netx: bump IRQ offset to 64 2013-03-03 22:54:11 +00:00
mach-nomadik Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-omap1 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-omap2 arm-soc: late OMAP changes 2013-02-28 20:00:40 -08:00
mach-orion5x Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-picoxcell Merge remote-tracking branches 'arm-soc/irqchip/gic-vic-move' and 'arm-soc/timer/cleanup' into soc 2013-01-24 17:57:20 +09:00
mach-prima2 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-pxa Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-realview arm-soc: cleanups 2013-02-21 14:58:40 -08:00
mach-rpc The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-s3c24xx Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-s3c64xx Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-s5p64x0 arm-soc: cleanups 2013-02-21 14:58:40 -08:00
mach-s5pc100 Merge remote-tracking branches 'arm-soc/irqchip/gic-vic-move' and 'arm-soc/timer/cleanup' into soc 2013-01-24 17:57:20 +09:00
mach-s5pv210 Merge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media 2013-02-24 17:35:10 -08:00
mach-sa1100 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-shark The series cleans up ARCH_HAS_DECOMP_WDOG and arch_decomp_wdog which 2013-02-05 10:26:20 -08:00
mach-shmobile IOMMU Updates for Linux v3.9 2013-02-26 11:09:17 -08:00
mach-socfpga Merge branch 'socfpga/hw' into next/soc 2013-02-11 19:37:51 -08:00
mach-spear3xx Merge branch 'next' of git://git.infradead.org/users/vkoul/slave-dma 2013-02-26 09:24:48 -08:00
mach-spear6xx Merge branch 'next' of git://git.infradead.org/users/vkoul/slave-dma 2013-02-26 09:24:48 -08:00
mach-spear13xx Revert "ARM: SPEAr13xx: Pass DW DMAC platform data from DT" 2013-02-20 15:48:02 +05:30
mach-sunxi Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-tegra Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-u300 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-ux500 Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-versatile Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-vexpress Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
mach-virt ARM: mach-virt: fixup machine descriptor after removal of sys_timer 2013-02-11 10:39:30 -08:00
mach-vt8500 arm-soc: soc-specific updates 2013-02-21 15:27:22 -08:00
mach-w90x900 arm-soc: cleanups 2013-02-21 14:58:40 -08:00
mach-zynq arm: zynq: Add missing irqchip.h to common.c 2013-01-29 10:55:28 -08:00
mm ARM: 7661/1: mm: perform explicit branch predictor maintenance when required 2013-03-03 22:54:16 +00:00
net ARM: net: bpf_jit: fix emit_swap16() for non ARMv6+. 2013-02-14 13:26:44 -05:00
nwfpe
oprofile
plat-iop
plat-omap arm-soc: late OMAP changes 2013-02-28 20:00:40 -08:00
plat-orion arm-soc: cleanups 2013-02-21 14:58:40 -08:00
plat-pxa
plat-samsung Merge branch 'timer/cleanup' into late/mvebu2 2013-02-28 18:54:15 +01:00
plat-spear arm-soc: driver specific changes 2013-02-21 15:12:18 -08:00
plat-versatile Merge branch 'for-arm-soc/arch-timers' of git://git.kernel.org/pub/scm/linux/kernel/git/will/linux into kvm-arm/timer 2013-02-11 19:03:52 +00:00
tools
vfp ARM: Fix broken commit 0cc41e4a21 corrupting kernel messages 2013-03-01 21:09:59 +00:00
xen xen: implement updated XENMEM_add_to_physmap_range ABI 2013-02-19 22:01:55 -05:00
Kconfig arm-soc: late OMAP changes 2013-02-28 20:00:40 -08:00
Kconfig-nommu
Kconfig.debug arm-soc: soc-specific updates 2013-02-21 15:27:22 -08:00
Makefile arm-soc: virtualization changes 2013-02-21 15:40:16 -08:00