From 8fca92705ef462f39e7db5a0f7100bcaae91bfd3 Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Thu, 16 Feb 2006 14:13:51 +1100 Subject: [PATCH] powerpc: Make UP -> SMP kexec work again For UP to SMP kexec to work we need to jump into pSeries_secondary_smp_init event on a UP + KEXEC kernel. The secondary cpus will not find their hw_cpu_id in the paca and so they'll jump into kexec_wait, ready for a kexec. Signed-off-by: Michael Ellerman Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/head_64.S | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'arch/powerpc/kernel/head_64.S') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index 415659629394..2b03a09fe5e9 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -157,8 +157,7 @@ _GLOBAL(__secondary_hold) SET_REG_IMMEDIATE(r4, .hmt_init) mtctr r4 bctr -#else -#ifdef CONFIG_SMP +#elif defined(CONFIG_SMP) || defined(CONFIG_KEXEC) LOAD_REG_IMMEDIATE(r4, .pSeries_secondary_smp_init) mtctr r4 mr r3,r24 @@ -166,7 +165,6 @@ _GLOBAL(__secondary_hold) #else BUG_OPCODE #endif -#endif /* This value is used to mark exception frames on the stack. */ .section ".toc","aw" -- cgit v1.2.1 From cb2c9b2741346eb23b177187a51ff5abf08295bd Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Mon, 13 Feb 2006 14:48:35 +1100 Subject: [PATCH] powerpc: Fix runlatch performance issues The runlatch SPR can take a lot of time to write. My original runlatch code would set it on every exception entry even though most of the time this was not required. It would also continually set it in the idle loop, which is an issue on an SMT capable processor. Now we cache the runlatch value in a threadinfo bit, and only check for it in decrementer and hardware interrupt exceptions as well as the idle loop. Boot on POWER3, POWER5 and iseries, and compile tested on pmac32. Signed-off-by: Anton Blanchard Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/head_64.S | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-) (limited to 'arch/powerpc/kernel/head_64.S') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index 2b03a09fe5e9..bb845eed0e98 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -319,7 +319,6 @@ exception_marker: label##_pSeries: \ HMT_MEDIUM; \ mtspr SPRN_SPRG1,r13; /* save r13 */ \ - RUNLATCH_ON(r13); \ EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, label##_common) #define STD_EXCEPTION_ISERIES(n, label, area) \ @@ -327,7 +326,6 @@ label##_pSeries: \ label##_iSeries: \ HMT_MEDIUM; \ mtspr SPRN_SPRG1,r13; /* save r13 */ \ - RUNLATCH_ON(r13); \ EXCEPTION_PROLOG_ISERIES_1(area); \ EXCEPTION_PROLOG_ISERIES_2; \ b label##_common @@ -337,7 +335,6 @@ label##_iSeries: \ label##_iSeries: \ HMT_MEDIUM; \ mtspr SPRN_SPRG1,r13; /* save r13 */ \ - RUNLATCH_ON(r13); \ EXCEPTION_PROLOG_ISERIES_1(PACA_EXGEN); \ lbz r10,PACAPROCENABLED(r13); \ cmpwi 0,r10,0; \ @@ -390,6 +387,7 @@ label##_common: \ label##_common: \ EXCEPTION_PROLOG_COMMON(trap, PACA_EXGEN); \ DISABLE_INTS; \ + bl .ppc64_runlatch_on; \ addi r3,r1,STACK_FRAME_OVERHEAD; \ bl hdlr; \ b .ret_from_except_lite @@ -407,7 +405,6 @@ __start_interrupts: _machine_check_pSeries: HMT_MEDIUM mtspr SPRN_SPRG1,r13 /* save r13 */ - RUNLATCH_ON(r13) EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common) . = 0x300 @@ -434,7 +431,6 @@ END_FTR_SECTION_IFCLR(CPU_FTR_SLB) data_access_slb_pSeries: HMT_MEDIUM mtspr SPRN_SPRG1,r13 - RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) mfspr r3,SPRN_DAR @@ -460,7 +456,6 @@ data_access_slb_pSeries: instruction_access_slb_pSeries: HMT_MEDIUM mtspr SPRN_SPRG1,r13 - RUNLATCH_ON(r13) mfspr r13,SPRN_SPRG3 /* get paca address into r13 */ std r3,PACA_EXSLB+EX_R3(r13) mfspr r3,SPRN_SRR0 /* SRR0 is faulting address */ @@ -491,7 +486,6 @@ instruction_access_slb_pSeries: .globl system_call_pSeries system_call_pSeries: HMT_MEDIUM - RUNLATCH_ON(r9) mr r9,r13 mfmsr r10 mfspr r13,SPRN_SPRG3 @@ -575,7 +569,6 @@ slb_miss_user_pseries: system_reset_fwnmi: HMT_MEDIUM mtspr SPRN_SPRG1,r13 /* save r13 */ - RUNLATCH_ON(r13) EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, system_reset_common) .globl machine_check_fwnmi @@ -583,7 +576,6 @@ system_reset_fwnmi: machine_check_fwnmi: HMT_MEDIUM mtspr SPRN_SPRG1,r13 /* save r13 */ - RUNLATCH_ON(r13) EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common) #ifdef CONFIG_PPC_ISERIES @@ -894,7 +886,6 @@ unrecov_fer: .align 7 .globl data_access_common data_access_common: - RUNLATCH_ON(r10) /* It wont fit in the 0x300 handler */ mfspr r10,SPRN_DAR std r10,PACA_EXGEN+EX_DAR(r13) mfspr r10,SPRN_DSISR @@ -1042,6 +1033,7 @@ hardware_interrupt_common: EXCEPTION_PROLOG_COMMON(0x500, PACA_EXGEN) hardware_interrupt_entry: DISABLE_INTS + bl .ppc64_runlatch_on addi r3,r1,STACK_FRAME_OVERHEAD bl .do_IRQ b .ret_from_except_lite -- cgit v1.2.1 From f1870f772c5e884862b4dd8f1ec2147247dda0ef Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Mon, 13 Feb 2006 18:11:13 +1100 Subject: [PATCH] powerpc64: remove broken/bitrotted HMT support HMT support is currently broken and needs to be reworked to play nicely with the SMT scheduler. Remove the bit rotten bits for the time being. I also updated an incorrect comment, we enter __secondary_hold with the physical cpu id in r3. Signed-off-by: Anton Blanchard Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/head_64.S | 97 ++----------------------------------------- 1 file changed, 4 insertions(+), 93 deletions(-) (limited to 'arch/powerpc/kernel/head_64.S') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index bb845eed0e98..11f2cd5af7dc 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -139,7 +139,7 @@ _GLOBAL(__secondary_hold) ori r24,r24,MSR_RI mtmsrd r24 /* RI on */ - /* Grab our linux cpu number */ + /* Grab our physical cpu number */ mr r24,r3 /* Tell the master cpu we're here */ @@ -153,11 +153,7 @@ _GLOBAL(__secondary_hold) cmpdi 0,r4,1 bne 100b -#ifdef CONFIG_HMT - SET_REG_IMMEDIATE(r4, .hmt_init) - mtctr r4 - bctr -#elif defined(CONFIG_SMP) || defined(CONFIG_KEXEC) +#if defined(CONFIG_SMP) || defined(CONFIG_KEXEC) LOAD_REG_IMMEDIATE(r4, .pSeries_secondary_smp_init) mtctr r4 mr r3,r24 @@ -1808,22 +1804,6 @@ _STATIC(start_here_multiplatform) ori r6,r6,MSR_RI mtmsrd r6 /* RI on */ -#ifdef CONFIG_HMT - /* Start up the second thread on cpu 0 */ - mfspr r3,SPRN_PVR - srwi r3,r3,16 - cmpwi r3,0x34 /* Pulsar */ - beq 90f - cmpwi r3,0x36 /* Icestar */ - beq 90f - cmpwi r3,0x37 /* SStar */ - beq 90f - b 91f /* HMT not supported */ -90: li r3,0 - bl .hmt_start_secondary -91: -#endif - /* The following gets the stack and TOC set up with the regs */ /* pointing to the real addr of the kernel stack. This is */ /* all done to support the C function call below which sets */ @@ -1937,77 +1917,8 @@ _STATIC(start_here_common) bl .start_kernel -_GLOBAL(hmt_init) -#ifdef CONFIG_HMT - LOAD_REG_IMMEDIATE(r5, hmt_thread_data) - mfspr r7,SPRN_PVR - srwi r7,r7,16 - cmpwi r7,0x34 /* Pulsar */ - beq 90f - cmpwi r7,0x36 /* Icestar */ - beq 91f - cmpwi r7,0x37 /* SStar */ - beq 91f - b 101f -90: mfspr r6,SPRN_PIR - andi. r6,r6,0x1f - b 92f -91: mfspr r6,SPRN_PIR - andi. r6,r6,0x3ff -92: sldi r4,r24,3 - stwx r6,r5,r4 - bl .hmt_start_secondary - b 101f - -__hmt_secondary_hold: - LOAD_REG_IMMEDIATE(r5, hmt_thread_data) - clrldi r5,r5,4 - li r7,0 - mfspr r6,SPRN_PIR - mfspr r8,SPRN_PVR - srwi r8,r8,16 - cmpwi r8,0x34 - bne 93f - andi. r6,r6,0x1f - b 103f -93: andi. r6,r6,0x3f - -103: lwzx r8,r5,r7 - cmpw r8,r6 - beq 104f - addi r7,r7,8 - b 103b - -104: addi r7,r7,4 - lwzx r9,r5,r7 - mr r24,r9 -101: -#endif - mr r3,r24 - b .pSeries_secondary_smp_init - -#ifdef CONFIG_HMT -_GLOBAL(hmt_start_secondary) - LOAD_REG_IMMEDIATE(r4,__hmt_secondary_hold) - clrldi r4,r4,4 - mtspr SPRN_NIADORM, r4 - mfspr r4, SPRN_MSRDORM - li r5, -65 - and r4, r4, r5 - mtspr SPRN_MSRDORM, r4 - lis r4,0xffef - ori r4,r4,0x7403 - mtspr SPRN_TSC, r4 - li r4,0x1f4 - mtspr SPRN_TST, r4 - mfspr r4, SPRN_HID0 - ori r4, r4, 0x1 - mtspr SPRN_HID0, r4 - mfspr r4, SPRN_CTRLF - oris r4, r4, 0x40 - mtspr SPRN_CTRLT, r4 - blr -#endif + /* Not reached */ + BUG_OPCODE /* * We put a few things here that have to be page-aligned. -- cgit v1.2.1 From c6622f63db86fcbd41bf6fe05ddf2e00c1e51ced Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Fri, 24 Feb 2006 10:06:59 +1100 Subject: powerpc: Implement accurate task and CPU time accounting This implements accurate task and cpu time accounting for 64-bit powerpc kernels. Instead of accounting a whole jiffy of time to a task on a timer interrupt because that task happened to be running at the time, we now account time in units of timebase ticks according to the actual time spent by the task in user mode and kernel mode. We also count the time spent processing hardware and software interrupts accurately. This is conditional on CONFIG_VIRT_CPU_ACCOUNTING. If that is not set, we do tick-based approximate accounting as before. To get this accurate information, we read either the PURR (processor utilization of resources register) on POWER5 machines, or the timebase on other machines on * each entry to the kernel from usermode * each exit to usermode * transitions between process context, hard irq context and soft irq context in kernel mode * context switches. On POWER5 systems with shared-processor logical partitioning we also read both the PURR and the timebase at each timer interrupt and context switch in order to determine how much time has been taken by the hypervisor to run other partitions ("steal" time). Unfortunately, since we need values of the PURR on both threads at the same time to accurately calculate the steal time, and since we can only calculate steal time on a per-core basis, the apportioning of the steal time between idle time (time which we ceded to the hypervisor in the idle loop) and actual stolen time is somewhat approximate at the moment. This is all based quite heavily on what s390 does, and it uses the generic interfaces that were added by the s390 developers, i.e. account_system_time(), account_user_time(), etc. This patch doesn't add any new interfaces between the kernel and userspace, and doesn't change the units in which time is reported to userspace by things such as /proc/stat, /proc//stat, getrusage(), times(), etc. Internally the various task and cpu times are stored in timebase units, but they are converted to USER_HZ units (1/100th of a second) when reported to userspace. Some precision is therefore lost but there should not be any accumulating error, since the internal accumulation is at full precision. Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/head_64.S | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'arch/powerpc/kernel/head_64.S') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index 2b21ec499285..be3ae7733577 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -277,6 +277,7 @@ exception_marker: std r10,0(r1); /* make stack chain pointer */ \ std r0,GPR0(r1); /* save r0 in stackframe */ \ std r10,GPR1(r1); /* save r1 in stackframe */ \ + ACCOUNT_CPU_USER_ENTRY(r9, r10); \ std r2,GPR2(r1); /* save r2 in stackframe */ \ SAVE_4GPRS(3, r1); /* save r3 - r6 in stackframe */ \ SAVE_2GPRS(7, r1); /* save r7, r8 in stackframe */ \ @@ -844,6 +845,14 @@ fast_exception_return: ld r11,_NIP(r1) andi. r3,r12,MSR_RI /* check if RI is set */ beq- unrecov_fer + +#ifdef CONFIG_VIRT_CPU_ACCOUNTING + andi. r3,r12,MSR_PR + beq 2f + ACCOUNT_CPU_USER_EXIT(r3, r4) +2: +#endif + ld r3,_CCR(r1) ld r4,_LINK(r1) ld r5,_CTR(r1) -- cgit v1.2.1 From c05b47704570b015134522c36142cd17bd48640a Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sat, 4 Mar 2006 15:00:45 -0800 Subject: ppc64: make sure to align stack pointer to 16 bytes at boot yaboot is scrogged and calls us with an invalid stack alignment, it seems. Thanks to David Woodhouse to pointing me to the problem. Signed-off-by: Linus Torvalds --- arch/powerpc/kernel/head_64.S | 3 +++ 1 file changed, 3 insertions(+) (limited to 'arch/powerpc/kernel/head_64.S') diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S index 11f2cd5af7dc..9b65029dd2a3 100644 --- a/arch/powerpc/kernel/head_64.S +++ b/arch/powerpc/kernel/head_64.S @@ -1537,6 +1537,9 @@ _STATIC(__boot_from_prom) mr r28,r6 mr r27,r7 + /* Align the stack to 16-byte boundary for broken yaboot */ + rldicr r1,r1,0,59 + /* Make sure we are running in 64 bits mode */ bl .enable_64b_mode -- cgit v1.2.1