linux/include/asm-powerpc/paca.h

114 lines
3.6 KiB
C
Raw Normal View History

/*
* include/asm-powerpc/paca.h
*
* This control block defines the PACA which defines the processor
* specific data for each logical processor on the system.
* There are some pointers defined that are utilized by PLIC.
*
* C 2001 PPC 64 Team, IBM Corp
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*/
#ifndef _ASM_POWERPC_PACA_H
#define _ASM_POWERPC_PACA_H
#ifdef __KERNEL__
#include <asm/types.h>
#include <asm/lppaca.h>
#include <asm/mmu.h>
register struct paca_struct *local_paca asm("r13");
#if defined(CONFIG_DEBUG_PREEMPT) && defined(CONFIG_SMP)
extern unsigned int debug_smp_processor_id(void); /* from linux/smp.h */
/*
* Add standard checks that preemption cannot occur when using get_paca():
* otherwise the paca_struct it points to may be the wrong one just after.
*/
#define get_paca() ((void) debug_smp_processor_id(), local_paca)
#else
#define get_paca() local_paca
#endif
#define get_lppaca() (get_paca()->lppaca_ptr)
#define get_slb_shadow() (get_paca()->slb_shadow_ptr)
struct task_struct;
/*
* Defines the layout of the paca.
*
* This structure is not directly accessed by firmware or the service
* processor.
*/
struct paca_struct {
/*
* Because hw_cpu_id, unlike other paca fields, is accessed
* routinely from other CPUs (from the IRQ code), we stick to
* read-only (after boot) fields in the first cacheline to
* avoid cacheline bouncing.
*/
struct lppaca *lppaca_ptr; /* Pointer to LpPaca for PLIC */
/*
* MAGIC: the spinlock functions in arch/powerpc/lib/locks.c
* load lock_token and paca_index with a single lwz
* instruction. They must travel together and be properly
* aligned.
*/
u16 lock_token; /* Constant 0x8000, used in locks */
u16 paca_index; /* Logical processor number */
u64 kernel_toc; /* Kernel TOC address */
u64 stab_real; /* Absolute address of segment table */
u64 stab_addr; /* Virtual address of segment table */
void *emergency_sp; /* pointer to emergency stack */
[PATCH] powerpc/64: per cpu data optimisations The current ppc64 per cpu data implementation is quite slow. eg: lhz 11,18(13) /* smp_processor_id() */ ld 9,.LC63-.LCTOC1(30) /* per_cpu__variable_name */ ld 8,.LC61-.LCTOC1(30) /* __per_cpu_offset */ sldi 11,11,3 /* form index into __per_cpu_offset */ mr 10,9 ldx 9,11,8 /* __per_cpu_offset[smp_processor_id()] */ ldx 0,10,9 /* load per cpu data */ 5 loads for something that is supposed to be fast, pretty awful. One reason for the large number of loads is that we have to synthesize 2 64bit constants (per_cpu__variable_name and __per_cpu_offset). By putting __per_cpu_offset into the paca we can avoid the 2 loads associated with it: ld 11,56(13) /* paca->data_offset */ ld 9,.LC59-.LCTOC1(30) /* per_cpu__variable_name */ ldx 0,9,11 /* load per cpu data Longer term we can should be able to do even better than 3 loads. If per_cpu__variable_name wasnt a 64bit constant and paca->data_offset was in a register we could cut it down to one load. A suggestion from Rusty is to use gcc's __thread extension here. In order to do this we would need to free up r13 (the __thread register and where the paca currently is). So far Ive had a few unsuccessful attempts at doing that :) The patch also allocates per cpu memory node local on NUMA machines. This patch from Rusty has been sitting in my queue _forever_ but stalled when I hit the compiler bug. Sorry about that. Finally I also only allocate per cpu data for possible cpus, which comes straight out of the x86-64 port. On a pseries kernel (with NR_CPUS == 128) and 4 possible cpus we see some nice gains: total used free shared buffers cached Mem: 4012228 212860 3799368 0 0 162424 total used free shared buffers cached Mem: 4016200 212984 3803216 0 0 162424 A saving of 3.75MB. Quite nice for smaller machines. Note: we now have to be careful of per cpu users that touch data for !possible cpus. At this stage it might be worth making the NUMA and possible cpu optimisations generic, but per cpu init is done so early we have to be careful that all architectures have their possible map setup correctly. Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2006-01-11 02:16:44 +00:00
u64 data_offset; /* per cpu data offset */
s16 hw_cpu_id; /* Physical processor number */
u8 cpu_start; /* At startup, processor spins until */
/* this becomes non-zero. */
struct slb_shadow *slb_shadow_ptr;
/*
* Now, starting in cacheline 2, the exception save areas
*/
/* used for most interrupts/exceptions */
u64 exgen[10] __attribute__((aligned(0x80)));
u64 exmc[10]; /* used for machine checks */
u64 exslb[10]; /* used for SLB/segment table misses
* on the linear mapping */
mm_context_t context;
powerpc: Use 64k pages without needing cache-inhibited large pages Some POWER5+ machines can do 64k hardware pages for normal memory but not for cache-inhibited pages. This patch lets us use 64k hardware pages for most user processes on such machines (assuming the kernel has been configured with CONFIG_PPC_64K_PAGES=y). User processes start out using 64k pages and get switched to 4k pages if they use any non-cacheable mappings. With this, we use 64k pages for the vmalloc region and 4k pages for the imalloc region. If anything creates a non-cacheable mapping in the vmalloc region, the vmalloc region will get switched to 4k pages. I don't know of any driver other than the DRM that would do this, though, and these machines don't have AGP. When a region gets switched from 64k pages to 4k pages, we do not have to clear out all the 64k HPTEs from the hash table immediately. We use the _PAGE_COMBO bit in the Linux PTE to indicate whether the page was hashed in as a 64k page or a set of 4k pages. If hash_page is trying to insert a 4k page for a Linux PTE and it sees that it has already been inserted as a 64k page, it first invalidates the 64k HPTE before inserting the 4k HPTE. The hash invalidation routines also use the _PAGE_COMBO bit, to determine whether to look for a 64k HPTE or a set of 4k HPTEs to remove. With those two changes, we can tolerate a mix of 4k and 64k HPTEs in the hash table, and they will all get removed when the address space is torn down. Signed-off-by: Paul Mackerras <paulus@samba.org>
2006-06-15 00:45:18 +00:00
u16 vmalloc_sllp;
u16 slb_cache_ptr;
[POWERPC] Introduce address space "slices" The basic issue is to be able to do what hugetlbfs does but with different page sizes for some other special filesystems; more specifically, my need is: - Huge pages - SPE local store mappings using 64K pages on a 4K base page size kernel on Cell - Some special 4K segments in 64K-page kernels for mapping a dodgy type of powerpc-specific infiniband hardware that requires 4K MMU mappings for various reasons I won't explain here. The main issues are: - To maintain/keep track of the page size per "segment" (as we can only have one page size per segment on powerpc, which are 256MB divisions of the address space). - To make sure special mappings stay within their allotted "segments" (including MAP_FIXED crap) - To make sure everybody else doesn't mmap/brk/grow_stack into a "segment" that is used for a special mapping Some of the necessary mechanisms to handle that were present in the hugetlbfs code, but mostly in ways not suitable for anything else. The patch relies on some changes to the generic get_unmapped_area() that just got merged. It still hijacks hugetlb callbacks here or there as the generic code hasn't been entirely cleaned up yet but that shouldn't be a problem. So what is a slice ? Well, I re-used the mechanism used formerly by our hugetlbfs implementation which divides the address space in "meta-segments" which I called "slices". The division is done using 256MB slices below 4G, and 1T slices above. Thus the address space is divided currently into 16 "low" slices and 16 "high" slices. (Special case: high slice 0 is the area between 4G and 1T). Doing so simplifies significantly the tracking of segments and avoids having to keep track of all the 256MB segments in the address space. While I used the "concepts" of hugetlbfs, I mostly re-implemented everything in a more generic way and "ported" hugetlbfs to it. Slices can have an associated page size, which is encoded in the mmu context and used by the SLB miss handler to set the segment sizes. The hash code currently doesn't care, it has a specific check for hugepages, though I might add a mechanism to provide per-slice hash mapping functions in the future. The slice code provide a pair of "generic" get_unmapped_area() (bottomup and topdown) functions that should work with any slice size. There is some trickiness here so I would appreciate people to have a look at the implementation of these and let me know if I got something wrong. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-08 06:27:27 +00:00
u16 slb_cache[SLB_CACHE_ENTRIES];
/*
* then miscellaneous read-write fields
*/
struct task_struct *__current; /* Pointer to current */
u64 kstack; /* Saved Kernel stack addr */
u64 stab_rr; /* stab/slb round-robin counter */
u64 saved_r1; /* r1 save for RTAS calls */
u64 saved_msr; /* MSR saved here by enter_rtas */
u16 trap_save; /* Used when bad stack is encountered */
u8 soft_enabled; /* irq soft-enable flag */
u8 hard_enabled; /* set if irqs are enabled in MSR */
u8 io_sync; /* writel() needs spin_unlock sync */
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/<pid>/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 <paulus@samba.org>
2006-02-23 23:06:59 +00:00
/* Stuff for accurate time accounting */
u64 user_time; /* accumulated usermode TB ticks */
u64 system_time; /* accumulated system TB ticks */
u64 startpurr; /* PURR/TB value snapshot */
powerpc: add scaled time accounting This adds POWERPC specific hooks for scaled time accounting. POWER6 includes a SPURR register. The SPURR is based off the PURR register but is scaled based on CPU frequency and issue rates. This gives a more accurate account of the instructions used per task. The PURR and timebase will be constant relative to the wall clock, irrespective of the CPU frequency. This implementation reads the SPURR register in account_system_vtime which is only call called on context witch and hard and soft irq entry and exit. The percentage of user and system time is then estimated using the ratio of these accounted by the PURR. If the SPURR is not present, the PURR read. An earlier implementation of this patch read the SPURR whenever the PURR was read, which included the system call entry and exit path. Unfortunately this showed a performance regression on lmbench runs, so was re-implemented. I've included the lmbench results here when run bare metal on POWER6. 1st column is the unpatch results. 2nd column is the results using the below patch and the 3rd is the % diff of these results from the base. 4th and 5th columns are the results and % differnce from the base using the older patch (SPURR read in syscall entry/exit path). Base Scaled-Acct SPURR-in-syscall Result Result % diff Result % diff Simple syscall: 0.3086 0.3086 0.0000 0.3452 11.8600 Simple read: 0.4591 0.4671 1.7425 0.5044 9.86713 Simple write: 0.4364 0.4366 0.0458 0.4731 8.40971 Simple stat: 2.0055 2.0295 1.1967 2.0669 3.06158 Simple fstat: 0.5962 0.5876 -1.442 0.6368 6.80979 Simple open/close: 3.1283 3.1009 -0.875 3.2088 2.57328 Select on 10 fd's: 0.8554 0.8457 -1.133 0.8667 1.32101 Select on 100 fd's: 3.5292 3.6329 2.9383 3.6664 3.88756 Select on 250 fd's: 7.9097 8.1881 3.5197 8.2242 3.97613 Select on 500 fd's: 15.2659 15.836 3.7357 15.873 3.97814 Select on 10 tcp fd's: 0.9576 0.9416 -1.670 0.9752 1.83792 Select on 100 tcp fd's: 7.248 7.2254 -0.311 7.2685 0.28283 Select on 250 tcp fd's: 17.7742 17.707 -0.375 17.749 -0.1406 Select on 500 tcp fd's: 35.4258 35.25 -0.496 35.286 -0.3929 Signal handler installation: 0.6131 0.6075 -0.913 0.647 5.52927 Signal handler overhead: 2.0919 2.1078 0.7600 2.1831 4.35967 Protection fault: 0.7345 0.7478 1.8107 0.8031 9.33968 Pipe latency: 33.006 16.398 -50.31 33.475 1.42368 AF_UNIX sock stream latency: 14.5093 30.910 113.03 30.715 111.692 Process fork+exit: 219.8 222.8 1.3648 229.37 4.35623 Process fork+execve: 876.14 873.28 -0.32 868.66 -0.8533 Process fork+/bin/sh -c: 2830 2876.5 1.6431 2958 4.52296 File /var/tmp/XXX write bw: 1193497 1195536 0.1708 118657 -0.5799 Pagefaults on /var/tmp/XXX: 3.1272 3.2117 2.7020 3.2521 3.99398 Also, kernel compile times show no difference with this patch applied. [pbadari@us.ibm.com: Avoid unnecessary PURR reading] Signed-off-by: Michael Neuling <mikey@neuling.org> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Jay Lan <jlan@engr.sgi.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-18 10:06:37 +00:00
u64 startspurr; /* SPURR value snapshot */
};
extern struct paca_struct paca[];
#endif /* __KERNEL__ */
#endif /* _ASM_POWERPC_PACA_H */