On Tue, Aug 17, 2010 at 01:53:12PM +0300, Tony Lindgren wrote:
> Here are some experimental patches to allow booting ARMv7 SMP
> kernel on UP to some extent. Posting these early in case it is
> of any help as I know at least Bryan Wu is working on similar
> issues.
I think these are compeltely the wrong direction. First thing to realise
is that XIP in the "SMP and UP in one kernel" is not really practical -
I'm not sure that many people who want that kind of flexibility also want
XIP too.
So let's forget about the kernel text being read-only.
The second thing to realise is that most of the SMP dependencies are in
assembly - and we can make lists of instructions and their modified
versions that would be necessary to boot a SMP kernel on UP.
So something like this will do (though note that not everywhere has been
fixed up - such as the page table flags - or this patch tested yet.)
If we don't want the SMP-on-UP support for SMP kernels (it's not actually
all that big - around 512 bytes) then we can discard the .smpalt.init
section and the __fixup_smp code.
diff --git a/arch/arm/include/asm/assembler.h b/arch/arm/include/asm/assembler.h
index 6e8f05c..55974d2 100644
--- a/arch/arm/include/asm/assembler.h
+++ b/arch/arm/include/asm/assembler.h
@@ -154,16 +154,32 @@
.long 9999b,9001f; \
.popsection
+#ifdef CONFIG_SMP
+#define SMP(instr...) \
+9998: instr
+#define UP(instr...) \
+ .pushsection ".smpalt.init", "a" ;\
+ .word 9998b ;\
+ instr ;\
+ .popsection
+#else
+#define SMP(instr...)
+#define UP(instr...) instr
+#endif
+
/*
* SMP data memory barrier
*/
.macro smp_dmb
#ifdef CONFIG_SMP
#if __LINUX_ARM_ARCH__ >= 7
- dmb
+ SMP(dmb)
#elif __LINUX_ARM_ARCH__ == 6
- mcr p15, 0, r0, c7, c10, 5 @ dmb
+ SMP(mcr p15, 0, r0, c7, c10, 5) @ dmb
+#else
+#error Incompatible SMP platform
#endif
+ UP(nop)
#endif
.endm
diff --git a/arch/arm/include/asm/smp_midr.h b/arch/arm/include/asm/smp_midr.h
index e69de29..4538ba4 100644
--- a/arch/arm/include/asm/smp_midr.h
+++ b/arch/arm/include/asm/smp_midr.h
@@ -0,0 +1,17 @@
+#ifndef ASMARM_SMP_MIDR_H
+#define ASMARM_SMP_MIDR_H
+
+#define hard_smp_processor_id()
\
+ ({ \
+ unsigned int cpunum; \
+ __asm__("\n" \
+ "1: mrc p15, 0, %0, c0, c0, 5\n" \
+ " .pushsection \".smpalt.init\", \"a\"\n" \
+ " .word 1b\n" \
+ " mov %0, #0\n" \
+ " .popsection" \
+ : "=r" (cpunum)); \
+ cpunum &= 0x0F; \
+ })
+
+#endif
diff --git a/arch/arm/include/asm/tlbflush.h b/arch/arm/include/asm/tlbflush.h
index 33b546a..0644860 100644
--- a/arch/arm/include/asm/tlbflush.h
+++ b/arch/arm/include/asm/tlbflush.h
@@ -185,12 +185,15 @@
# define v6wbi_always_flags (-1UL)
#endif
-#ifdef CONFIG_SMP
-#define v7wbi_tlb_flags (TLB_WB | TLB_DCLEAN | TLB_V7_IS_BTB | \
+#define v7wbi_tlb_flags_smp (TLB_WB | TLB_DCLEAN | TLB_V7_IS_BTB | \
TLB_V7_UIS_FULL | TLB_V7_UIS_PAGE | TLB_V7_UIS_ASID)
-#else
-#define v7wbi_tlb_flags (TLB_WB | TLB_DCLEAN | TLB_BTB | \
+#define v7wbi_tlb_flags_up (TLB_WB | TLB_DCLEAN | TLB_BTB | \
TLB_V6_U_FULL | TLB_V6_U_PAGE | TLB_V6_U_ASID)
+
+#ifdef CONFIG_SMP
+#define v7wbi_tlb_flags v7wbi_tlb_flags_smp
+#else
+#define v7wbi_tlb_flags v7wbi_tlb_flags_up
#endif
#ifdef CONFIG_CPU_TLB_V7
diff --git a/arch/arm/kernel/entry-armv.S b/arch/arm/kernel/entry-armv.S
index bb8e93a..bb2ef60 100644
--- a/arch/arm/kernel/entry-armv.S
+++ b/arch/arm/kernel/entry-armv.S
@@ -965,11 +965,8 @@ kuser_cmpxchg_fixup:
beq 1b
rsbs r0, r3, #0
/* beware -- each __kuser slot must be 8 instructions max */
-#ifdef CONFIG_SMP
- b __kuser_memory_barrier
-#else
- usr_ret lr
-#endif
+ SMP(b __kuser_memory_barrier)
+ UP(usr_ret lr)
#endif
diff --git a/arch/arm/kernel/head.S b/arch/arm/kernel/head.S
index eb62bf9..feabbf0 100644
--- a/arch/arm/kernel/head.S
+++ b/arch/arm/kernel/head.S
@@ -86,6 +86,9 @@ ENTRY(stext)
movs r8, r5 @ invalid machine (r5=0)?
beq __error_a @ yes, error 'a'
bl __vet_atags
+#ifdef CONFIG_SMP
+ bl __fixup_smp
+#endif
bl __create_page_tables
/*
@@ -333,4 +336,35 @@ __create_page_tables:
ENDPROC(__create_page_tables)
.ltorg
+#ifdef CONFIG_SMP
+__fixup_smp:
+ and r0, r9, #0xff000000
+ teq r0, #0x41000000 @ ARM CPU?
+ bne smp_on_up @ no, assume UP
+ and r0, r9, #0x00070000
+ teq r0, #0x00070000 @ ARMv6/v7?
+ bne smp_on_up @ no, assume UP
+ mrc p15, 0, r0, c0, c0, 5 @ read MIDR
+ movs r0, r0, lsr #30
+ teqne r0, #3 @ check top two bits 00 or 11
+ moveq pc, lr
+
+smp_on_up:
+ adr r0, 1f
+ ldmia r0, {r1, r2, r3}
+ sub r1, r0, r1
+ add r2, r2, r1
+ add r3, r3, r1
+2: cmp r2, r3
+ ldmia r2!, {r0, r4}
+ movhs pc, lr
+ str r4, [r0, r1]
+ b 2b
+ENDPROC(__fixup_smp)
+
+1: .word .
+ .word __smpalt_begin
+ .word __smpalt_end
+#endif
+
#include "head-common.S"
diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
index b16c079..89858be 100644
--- a/arch/arm/kernel/vmlinux.lds.S
+++ b/arch/arm/kernel/vmlinux.lds.S
@@ -40,6 +40,9 @@ SECTIONS
__tagtable_begin = .;
*(.taglist.init)
__tagtable_end = .;
+ __smpalt_begin = .;
+ *(.smpalt.init)
+ __smpalt_end = .;
INIT_SETUP(16)
diff --git a/arch/arm/mach-realview/include/mach/smp.h
b/arch/arm/mach-realview/include/mach/smp.h
index dd53892..833ee85 100644
--- a/arch/arm/mach-realview/include/mach/smp.h
+++ b/arch/arm/mach-realview/include/mach/smp.h
@@ -1,16 +1,8 @@
#ifndef ASMARM_ARCH_SMP_H
#define ASMARM_ARCH_SMP_H
-
#include <asm/hardware/gic.h>
-
-#define hard_smp_processor_id() \
- ({ \
- unsigned int cpunum; \
- __asm__("mrc p15, 0, %0, c0, c0, 5" \
- : "=r" (cpunum)); \
- cpunum &= 0x0F; \
- })
+#include <asm/smp_midr.h>
/*
* We use IRQ1 as the IPI
diff --git a/arch/arm/mach-s5pv310/include/mach/smp.h
b/arch/arm/mach-s5pv310/include/mach/smp.h
index 990f3ba..1b8b637 100644
--- a/arch/arm/mach-s5pv310/include/mach/smp.h
+++ b/arch/arm/mach-s5pv310/include/mach/smp.h
@@ -7,17 +7,10 @@
#define ASM_ARCH_SMP_H __FILE__
#include <asm/hardware/gic.h>
+#include <asm/smp_midr.h>
extern void __iomem *gic_cpu_base_addr;
-#define hard_smp_processor_id() \
- ({ \
- unsigned int cpunum; \
- __asm__("mrc p15, 0, %0, c0, c0, 5" \
- : "=r" (cpunum)); \
- cpunum &= 0x03; \
- })
-
/*
* We use IRQ1 as the IPI
*/
diff --git a/arch/arm/mach-tegra/include/mach/smp.h
b/arch/arm/mach-tegra/include/mach/smp.h
index 8b42dab..d5c4030 100644
--- a/arch/arm/mach-tegra/include/mach/smp.h
+++ b/arch/arm/mach-tegra/include/mach/smp.h
@@ -1,16 +1,8 @@
#ifndef ASMARM_ARCH_SMP_H
#define ASMARM_ARCH_SMP_H
-
#include <asm/hardware/gic.h>
-
-#define hard_smp_processor_id() \
- ({ \
- unsigned int cpunum; \
- __asm__("mrc p15, 0, %0, c0, c0, 5" \
- : "=r" (cpunum)); \
- cpunum &= 0x0F; \
- })
+#include <asm/smp_midr.h>
/*
* We use IRQ1 as the IPI
diff --git a/arch/arm/mach-ux500/include/mach/smp.h
b/arch/arm/mach-ux500/include/mach/smp.h
index b59f7bc..87a9cf3 100644
--- a/arch/arm/mach-ux500/include/mach/smp.h
+++ b/arch/arm/mach-ux500/include/mach/smp.h
@@ -10,18 +10,11 @@
#define ASMARM_ARCH_SMP_H
#include <asm/hardware/gic.h>
+#include <asm/smp_midr.h>
/* This is required to wakeup the secondary core */
extern void u8500_secondary_startup(void);
-#define hard_smp_processor_id() \
- ({ \
- unsigned int cpunum; \
- __asm__("mrc p15, 0, %0, c0, c0, 5" \
- : "=r" (cpunum)); \
- cpunum &= 0x0F; \
- })
-
/*
* We use IRQ1 as the IPI
*/
diff --git a/arch/arm/mach-vexpress/include/mach/smp.h
b/arch/arm/mach-vexpress/include/mach/smp.h
index 72a9621..e02bc7d 100644
--- a/arch/arm/mach-vexpress/include/mach/smp.h
+++ b/arch/arm/mach-vexpress/include/mach/smp.h
@@ -2,14 +2,7 @@
#define __MACH_SMP_H
#include <asm/hardware/gic.h>
-
-#define hard_smp_processor_id() \
- ({ \
- unsigned int cpunum; \
- __asm__("mrc p15, 0, %0, c0, c0, 5" \
- : "=r" (cpunum)); \
- cpunum &= 0x0F; \
- })
+#include <asm/smp_midr.h>
/*
* We use IRQ1 as the IPI
diff --git a/arch/arm/mm/cache-v7.S b/arch/arm/mm/cache-v7.S
index 37c8157..2aa59d5 100644
--- a/arch/arm/mm/cache-v7.S
+++ b/arch/arm/mm/cache-v7.S
@@ -91,11 +91,8 @@ ENTRY(v7_flush_kern_cache_all)
THUMB( stmfd sp!, {r4-r7, r9-r11, lr} )
bl v7_flush_dcache_all
mov r0, #0
-#ifdef CONFIG_SMP
- mcr p15, 0, r0, c7, c1, 0 @ invalidate I-cache inner
shareable
-#else
- mcr p15, 0, r0, c7, c5, 0 @ I+BTB cache invalidate
-#endif
+ SMP(mcr p15, 0, r0, c7, c1, 0) @ invalidate I-cache inner
shareable
+ UP(mcr p15, 0, r0, c7, c5, 0) @ I+BTB cache invalidate
ARM( ldmfd sp!, {r4-r5, r7, r9-r11, lr} )
THUMB( ldmfd sp!, {r4-r7, r9-r11, lr} )
mov pc, lr
@@ -171,11 +168,8 @@ ENTRY(v7_coherent_user_range)
cmp r0, r1
blo 1b
mov r0, #0
-#ifdef CONFIG_SMP
- mcr p15, 0, r0, c7, c1, 6 @ invalidate BTB Inner Shareable
-#else
- mcr p15, 0, r0, c7, c5, 6 @ invalidate BTB
-#endif
+ SMP(mcr p15, 0, r0, c7, c1, 6) @ invalidate BTB Inner Shareable
+ UP(mcr p15, 0, r0, c7, c5, 6) @ invalidate BTB
dsb
isb
mov pc, lr
diff --git a/arch/arm/mm/proc-v6.S b/arch/arm/mm/proc-v6.S
index 22aac85..692d22e 100644
--- a/arch/arm/mm/proc-v6.S
+++ b/arch/arm/mm/proc-v6.S
@@ -30,13 +30,10 @@
#define TTB_RGN_WT (2 << 3)
#define TTB_RGN_WB (3 << 3)
-#ifndef CONFIG_SMP
-#define TTB_FLAGS TTB_RGN_WBWA
-#define PMD_FLAGS PMD_SECT_WB
-#else
-#define TTB_FLAGS TTB_RGN_WBWA|TTB_S
-#define PMD_FLAGS PMD_SECT_WBWA|PMD_SECT_S
-#endif
+#define TTB_FLAGS_UP TTB_RGN_WBWA
+#define PMD_FLAGS_UP PMD_SECT_WB
+#define TTB_FLAGS_SMP TTB_RGN_WBWA|TTB_S
+#define PMD_FLAGS_SMP PMD_SECT_WBWA|PMD_SECT_S
ENTRY(cpu_v6_proc_init)
mov pc, lr
@@ -97,7 +94,8 @@ ENTRY(cpu_v6_switch_mm)
#ifdef CONFIG_MMU
mov r2, #0
ldr r1, [r1, #MM_CONTEXT_ID] @ get mm->context.id
- orr r0, r0, #TTB_FLAGS
+ SMP(orr r0, r0, #TTB_FLAGS_SMP)
+ UP(orr r0, r0, #TTB_FLAGS_UP)
mcr p15, 0, r2, c7, c5, 6 @ flush BTAC/BTB
mcr p15, 0, r2, c7, c10, 4 @ drain write buffer
mcr p15, 0, r0, c2, c0, 0 @ set TTB 0
@@ -169,7 +167,8 @@ __v6_setup:
#ifdef CONFIG_MMU
mcr p15, 0, r0, c8, c7, 0 @ invalidate I + D TLBs
mcr p15, 0, r0, c2, c0, 2 @ TTB control register
- orr r4, r4, #TTB_FLAGS
+ SMP(orr r4, r4, #TTB_FLAGS_SMP)
+ UP(orr r4, r4, #TTB_FLAGS_UP)
mcr p15, 0, r4, c2, c0, 1 @ load TTB1
#endif /* CONFIG_MMU */
adr r5, v6_crval
@@ -225,10 +224,16 @@ cpu_elf_name:
__v6_proc_info:
.long 0x0007b000
.long 0x0007f000
- .long PMD_TYPE_SECT | \
+ SMP(.long \
+ PMD_TYPE_SECT | \
+ PMD_SECT_AP_WRITE | \
+ PMD_SECT_AP_READ | \
+ PMD_FLAGS_SMP)
+ UP(.long \
+ PMD_TYPE_SECT | \
PMD_SECT_AP_WRITE | \
PMD_SECT_AP_READ | \
- PMD_FLAGS
+ PMD_FLAGS_UP)
.long PMD_TYPE_SECT | \
PMD_SECT_XN | \
PMD_SECT_AP_WRITE | \
@@ -249,10 +254,16 @@ __v6_proc_info:
__pj4_v6_proc_info:
.long 0x560f5810
.long 0xff0ffff0
- .long PMD_TYPE_SECT | \
+ SMP(.long \
+ PMD_TYPE_SECT | \
+ PMD_SECT_AP_WRITE | \
+ PMD_SECT_AP_READ | \
+ PMD_FLAGS_SMP)
+ UP(.long \
+ PMD_TYPE_SECT | \
PMD_SECT_AP_WRITE | \
PMD_SECT_AP_READ | \
- PMD_FLAGS
+ PMD_FLAGS_UP)
.long PMD_TYPE_SECT | \
PMD_SECT_XN | \
PMD_SECT_AP_WRITE | \
diff --git a/arch/arm/mm/proc-v7.S b/arch/arm/mm/proc-v7.S
index 6a8506d..c04f9c1 100644
--- a/arch/arm/mm/proc-v7.S
+++ b/arch/arm/mm/proc-v7.S
@@ -30,15 +30,13 @@
#define TTB_IRGN_WT ((1 << 0) | (0 << 6))
#define TTB_IRGN_WB ((1 << 0) | (1 << 6))
-#ifndef CONFIG_SMP
/* PTWs cacheable, inner WB not shareable, outer WB not shareable */
-#define TTB_FLAGS TTB_IRGN_WB|TTB_RGN_OC_WB
-#define PMD_FLAGS PMD_SECT_WB
-#else
+#define TTB_FLAGS_UP TTB_IRGN_WB|TTB_RGN_OC_WB
+#define PMD_FLAGS_UP PMD_SECT_WB
+
/* PTWs cacheable, inner WBWA shareable, outer WBWA not shareable */
-#define TTB_FLAGS TTB_IRGN_WBWA|TTB_S|TTB_NOS|TTB_RGN_OC_WBWA
-#define PMD_FLAGS PMD_SECT_WBWA|PMD_SECT_S
-#endif
+#define TTB_FLAGS_SMP TTB_IRGN_WBWA|TTB_S|TTB_NOS|TTB_RGN_OC_WBWA
+#define PMD_FLAGS_SMP PMD_SECT_WBWA|PMD_SECT_S
ENTRY(cpu_v7_proc_init)
mov pc, lr
@@ -105,7 +103,8 @@ ENTRY(cpu_v7_switch_mm)
#ifdef CONFIG_MMU
mov r2, #0
ldr r1, [r1, #MM_CONTEXT_ID] @ get mm->context.id
- orr r0, r0, #TTB_FLAGS
+ SMP(orr r0, r0, #TTB_FLAGS_SMP)
+ UP(orr r0, r0, #TTB_FLAGS_UP)
#ifdef CONFIG_ARM_ERRATA_430973
mcr p15, 0, r2, c7, c5, 6 @ flush BTAC/BTB
#endif
@@ -235,7 +234,8 @@ __v7_setup:
#ifdef CONFIG_MMU
mcr p15, 0, r10, c8, c7, 0 @ invalidate I + D TLBs
mcr p15, 0, r10, c2, c0, 2 @ TTB control register
- orr r4, r4, #TTB_FLAGS
+ SMP(orr r4, r4, #TTB_FLAGS_SMP)
+ UP(orr r4, r4, #TTB_FLAGS_UP)
mcr p15, 0, r4, c2, c0, 1 @ load TTB1
mov r10, #0x1f @ domains 0, 1 = manager
mcr p15, 0, r10, c3, c0, 0 @ load domain access register
@@ -330,10 +330,16 @@ cpu_elf_name:
__v7_proc_info:
.long 0x000f0000 @ Required ID value
.long 0x000f0000 @ Mask for ID
- .long PMD_TYPE_SECT | \
+ SMP(.long \
+ PMD_TYPE_SECT | \
+ PMD_SECT_AP_WRITE | \
+ PMD_SECT_AP_READ | \
+ PMD_FLAGS_SMP)
+ UP(.long \
+ PMD_TYPE_SECT | \
PMD_SECT_AP_WRITE | \
PMD_SECT_AP_READ | \
- PMD_FLAGS
+ PMD_FLAGS_UP)
.long PMD_TYPE_SECT | \
PMD_SECT_XN | \
PMD_SECT_AP_WRITE | \
diff --git a/arch/arm/mm/tlb-v7.S b/arch/arm/mm/tlb-v7.S
index f3f288a..26247c9 100644
--- a/arch/arm/mm/tlb-v7.S
+++ b/arch/arm/mm/tlb-v7.S
@@ -13,6 +13,7 @@
*/
#include <linux/init.h>
#include <linux/linkage.h>
+#include <asm/assembler.h>
#include <asm/asm-offsets.h>
#include <asm/page.h>
#include <asm/tlbflush.h>
@@ -41,20 +42,15 @@ ENTRY(v7wbi_flush_user_tlb_range)
orr r0, r3, r0, lsl #PAGE_SHIFT @ Create initial MVA
mov r1, r1, lsl #PAGE_SHIFT
1:
-#ifdef CONFIG_SMP
- mcr p15, 0, r0, c8, c3, 1 @ TLB invalidate U MVA
(shareable)
-#else
- mcr p15, 0, r0, c8, c7, 1 @ TLB invalidate U MVA
-#endif
+ SMP(mcr p15, 0, r0, c8, c3, 1) @ TLB invalidate U MVA
(shareable)
+ UP(mcr p15, 0, r0, c8, c7, 1) @ TLB invalidate U MVA
+
add r0, r0, #PAGE_SZ
cmp r0, r1
blo 1b
mov ip, #0
-#ifdef CONFIG_SMP
- mcr p15, 0, ip, c7, c1, 6 @ flush BTAC/BTB Inner Shareable
-#else
- mcr p15, 0, ip, c7, c5, 6 @ flush BTAC/BTB
-#endif
+ SMP(mcr p15, 0, ip, c7, c1, 6) @ flush BTAC/BTB Inner Shareable
+ UP(mcr p15, 0, ip, c7, c5, 6) @ flush BTAC/BTB
dsb
mov pc, lr
ENDPROC(v7wbi_flush_user_tlb_range)
@@ -74,20 +70,14 @@ ENTRY(v7wbi_flush_kern_tlb_range)
mov r0, r0, lsl #PAGE_SHIFT
mov r1, r1, lsl #PAGE_SHIFT
1:
-#ifdef CONFIG_SMP
- mcr p15, 0, r0, c8, c3, 1 @ TLB invalidate U MVA
(shareable)
-#else
- mcr p15, 0, r0, c8, c7, 1 @ TLB invalidate U MVA
-#endif
+ SMP(mcr p15, 0, r0, c8, c3, 1) @ TLB invalidate U MVA
(shareable)
+ UP(mcr p15, 0, r0, c8, c7, 1) @ TLB invalidate U MVA
add r0, r0, #PAGE_SZ
cmp r0, r1
blo 1b
mov r2, #0
-#ifdef CONFIG_SMP
- mcr p15, 0, r2, c7, c1, 6 @ flush BTAC/BTB Inner Shareable
-#else
- mcr p15, 0, r2, c7, c5, 6 @ flush BTAC/BTB
-#endif
+ SMP(mcr p15, 0, r2, c7, c1, 6) @ flush BTAC/BTB Inner Shareable
+ UP(mcr p15, 0, r2, c7, c5, 6) @ flush BTAC/BTB
dsb
isb
mov pc, lr
@@ -99,5 +89,6 @@ ENDPROC(v7wbi_flush_kern_tlb_range)
ENTRY(v7wbi_tlb_fns)
.long v7wbi_flush_user_tlb_range
.long v7wbi_flush_kern_tlb_range
- .long v7wbi_tlb_flags
+ SMP(.long v7wbi_tlb_flags_smp)
+ UP(.long v7wbi_tlb_flags_up)
.size v7wbi_tlb_fns, . - v7wbi_tlb_fns
diff --git a/arch/arm/plat-omap/include/plat/smp.h
b/arch/arm/plat-omap/include/plat/smp.h
index 6a3ff65..ed96907 100644
--- a/arch/arm/plat-omap/include/plat/smp.h
+++ b/arch/arm/plat-omap/include/plat/smp.h
@@ -18,6 +18,7 @@
#define OMAP_ARCH_SMP_H
#include <asm/hardware/gic.h>
+#include <asm/smp_midr.h>
/*
* set_event() is used to wake up secondary core from wfe using sev. ROM
@@ -40,15 +41,4 @@ static inline void smp_cross_call(const struct cpumask *mask)
gic_raise_softirq(mask, 1);
}
-/*
- * Read MPIDR: Multiprocessor affinity register
- */
-#define hard_smp_processor_id() \
- ({ \
- unsigned int cpunum; \
- __asm__("mrc p15, 0, %0, c0, c0, 5" \
- : "=r" (cpunum)); \
- cpunum &= 0x0F; \
- })
-
#endif
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
