linux-stable/arch/arm/mach-exynos/headsmp.S

40 lines
847 B
ArmAsm
Raw Normal View History

/* SPDX-License-Identifier: GPL-2.0 */
/*
* Cloned from linux/arch/arm/mach-realview/headsmp.S
*
* Copyright (c) 2003 ARM Limited
* All Rights Reserved
*/
#include <linux/linkage.h>
#include <linux/init.h>
#include <asm/assembler.h>
/*
* exynos4 specific entry point for secondary CPUs. This provides
* a "holding pen" into which all secondary cores are held until we're
* ready for them to initialise.
*/
ENTRY(exynos4_secondary_startup)
ARM_BE8(setend be)
mrc p15, 0, r0, c0, c0, 5
and r0, r0, #15
adr r4, 1f
ldmia r4, {r5, r6}
sub r4, r4, r5
add r6, r6, r4
pen: ldr r7, [r6]
cmp r7, r0
bne pen
/*
* we've been released from the holding pen: secondary_stack
* should now contain the SVC stack for this core
*/
b secondary_startup
ENDPROC(exynos4_secondary_startup)
.align 2
1: .long .
ARM: smp: remove arch-provided "pen_release" Consolidating the "pen_release" stuff amongst the various SoC implementations gives credence to having a CPU holding pen for secondary CPUs. However, this is far from the truth. Many SoC implementations cargo-cult copied various bits of the pen release implementation from the initial Realview/Versatile Express implementation without understanding what it was or why it existed. The reason it existed is because these are _development_ platforms, and some board firmware is unable to individually control the startup of secondary CPUs. Moreover, they do not have a way to power down or reset secondary CPUs for hot-unplug. Hence, the pen_release implementation was designed for ARM Ltd's development platforms to provide a working implementation, even though it is very far from what is required. It was decided a while back to reduce the duplication by consolidating the "pen_release" variable, but this only made the situation worse - we have ended up with several implementations that read this variable but do not write it - again, showing the cargo-cult mentality at work, lack of proper review of new code, and in some cases a lack of testing. While it would be preferable to remove pen_release entirely from the kernel, this is not possible without help from the SoC maintainers, which seems to be lacking. However, I want to remove pen_release from arch code to remove the credence that having it gives. This patch removes pen_release from the arch code entirely, adding private per-SoC definitions for it instead, and explicitly stating that write_pen_release() is cargo-cult copied and should not be copied any further. Rename write_pen_release() in a similar fashion as well. Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
2018-12-13 14:02:48 +00:00
.long exynos_pen_release