From e0e8eb54d8ae0c4cfd1d297f6351b08a7f635c5f Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Thu, 16 Mar 2006 10:31:38 -0700 Subject: [PATCH 01/31] [PATCH] unshare: Use rcu_assign_pointer when setting sighand The sighand pointer only needs the rcu_read_lock on the read side. So only depending on task_lock protection when setting this pointer is not enough. We also need a memory barrier to ensure the initialization is seen first. Use rcu_assign_pointer as it does this for us, and clearly documents that we are setting an rcu readable pointer. Signed-off-by: Eric W. Biederman Acked-by: Paul E. McKenney Signed-off-by: Linus Torvalds --- kernel/fork.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/fork.c b/kernel/fork.c index ccdfbb16c86d..46060cb24af0 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1569,7 +1569,7 @@ asmlinkage long sys_unshare(unsigned long unshare_flags) if (new_sigh) { sigh = current->sighand; - current->sighand = new_sigh; + rcu_assign_pointer(current->sighand, new_sigh); new_sigh = sigh; } From 90036ee5938d89638e80f4d0d0700d0f2dbd4a6a Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Thu, 16 Mar 2006 23:03:59 -0800 Subject: [PATCH 02/31] [PATCH] page migration: Fail with error if swap not setup Currently the migration of anonymous pages will silently fail if no swap is setup. This patch makes page migration functions check for available swap and fail with -ENODEV if no swap space is available. Signed-off-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/mempolicy.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/mm/mempolicy.c b/mm/mempolicy.c index 2a8206009422..b21869a39f0b 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -330,9 +330,19 @@ check_range(struct mm_struct *mm, unsigned long start, unsigned long end, int err; struct vm_area_struct *first, *vma, *prev; - /* Clear the LRU lists so pages can be isolated */ - if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) + if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) { + /* Must have swap device for migration */ + if (nr_swap_pages <= 0) + return ERR_PTR(-ENODEV); + + /* + * Clear the LRU lists so pages can be isolated. + * Note that pages may be moved off the LRU after we have + * drained them. Those pages will fail to migrate like other + * pages that may be busy. + */ lru_add_drain_all(); + } first = find_vma(mm, start); if (!first) From 67890d7084085e29c51afa2514036d42643fd3cf Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Thu, 16 Mar 2006 23:04:00 -0800 Subject: [PATCH 03/31] [PATCH] time_interpolator: add __read_mostly The pointer to the current time interpolator and the current list of time interpolators are typically only changed during bootup. Adding __read_mostly takes them away from possibly hot cachelines. Signed-off-by: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/timer.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/timer.c b/kernel/timer.c index bf7c4193b936..2410c18dbeb1 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -1354,8 +1354,8 @@ void __init init_timers(void) #ifdef CONFIG_TIME_INTERPOLATION -struct time_interpolator *time_interpolator; -static struct time_interpolator *time_interpolator_list; +struct time_interpolator *time_interpolator __read_mostly; +static struct time_interpolator *time_interpolator_list __read_mostly; static DEFINE_SPINLOCK(time_interpolator_lock); static inline u64 time_interpolator_get_cycles(unsigned int src) From a0a0c28c1a7109d7955815074c52cac079ab3ba5 Mon Sep 17 00:00:00 2001 From: Roman Zippel Date: Thu, 16 Mar 2006 23:04:01 -0800 Subject: [PATCH 04/31] [PATCH] posix-timers: fix requeue accounting when signal is ignored When the posix-timer signal is ignored then the timer is rearmed by the callback function. The requeue pending accounting has to be fixed up else the state might be wrong. Signed-off-by: Roman Zippel Signed-off-by: Thomas Gleixner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/posix-timers.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index 216f574b5ffb..fa895fc2ecf5 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -353,6 +353,7 @@ static int posix_timer_fn(void *data) hrtimer_forward(&timr->it.real.timer, timr->it.real.interval); ret = HRTIMER_RESTART; + ++timr->it_requeue_pending; } } From 8532159f5521ba24e697f0d25970ae89ff62a1f2 Mon Sep 17 00:00:00 2001 From: Eric Van Hensbergen Date: Thu, 16 Mar 2006 23:04:04 -0800 Subject: [PATCH 05/31] [PATCH] v9fs: fix overzealous dropping of dentry which breaks dcache There is a d_drop in dir_release which caused problems as it invalidates dcache entries too soon. This was likely a part of the wierd cwd behavior folks were seeing. Signed-off-by: Eric Van Hensbergen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/9p/vfs_dir.c | 1 - 1 file changed, 1 deletion(-) diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c index ae6d032b9b59..cd5eeb032d64 100644 --- a/fs/9p/vfs_dir.c +++ b/fs/9p/vfs_dir.c @@ -202,7 +202,6 @@ int v9fs_dir_release(struct inode *inode, struct file *filp) filp->private_data = NULL; } - d_drop(filp->f_dentry); return 0; } From 82c3c03a4096badd026c6e337f3c5dde020e9ec6 Mon Sep 17 00:00:00 2001 From: Srivatsa Vaddagiri Date: Thu, 16 Mar 2006 23:04:06 -0800 Subject: [PATCH 06/31] [PATCH] x86: check for online cpus before bringing them up Bryce reported a bug wherein offlining CPU0 (on x86 box) and then subsequently onlining it resulted in a lockup. On x86, CPU0 is never offlined. The subsequent attempt to online CPU0 doesn't take that into account. It actually tries to bootup the already booted CPU. Following patch fixes the problem (as acknowledged by Bryce). Please consider for inclusion in 2.6.16. Check if cpu is already online. Signed-off-by: Srivatsa Vaddagiri Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/smpboot.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c index eba7f53f8b4a..7007e1783797 100644 --- a/arch/i386/kernel/smpboot.c +++ b/arch/i386/kernel/smpboot.c @@ -1029,6 +1029,16 @@ int __devinit smp_prepare_cpu(int cpu) int apicid, ret; lock_cpu_hotplug(); + + /* + * On x86, CPU0 is never offlined. Trying to bring up an + * already-booted CPU will hang. So check for that case. + */ + if (cpu_online(cpu)) { + ret = -EINVAL; + goto exit; + } + apicid = x86_cpu_to_apicid[cpu]; if (apicid == BAD_APICID) { ret = -ENODEV; From 8ba32fde2c5be52865b2fd7e5e3752a46971fabe Mon Sep 17 00:00:00 2001 From: Kevin Corry Date: Thu, 16 Mar 2006 23:04:03 -0800 Subject: [PATCH 07/31] [PATCH] dm stripe: Fix bounds The dm-stripe target currently does not enforce that the size of a stripe device be a multiple of the chunk-size. Under certain conditions, this can lead to I/O requests going off the end of an underlying device. This test-case shows one example. echo "0 100 linear /dev/hdb1 0" | dmsetup create linear0 echo "0 100 linear /dev/hdb1 100" | dmsetup create linear1 echo "0 200 striped 2 32 /dev/mapper/linear0 0 /dev/mapper/linear1 0" | \ dmsetup create stripe0 dd if=/dev/zero of=/dev/mapper/stripe0 bs=1k This will produce the output: dd: writing '/dev/mapper/stripe0': Input/output error 97+0 records in 96+0 records out And in the kernel log will be: attempt to access beyond end of device dm-0: rw=0, want=104, limit=100 The patch will check that the table size is a multiple of the stripe chunk-size when the table is created, which will prevent the above striped device from being created. This should not affect tools like LVM or EVMS, since in all the cases I can think of, striped devices are always created with the sizes being a multiple of the chunk-size. The size of a stripe device must be a multiple of its chunk-size. (akpm: that typecast is quite gratuitous) Signed-off-by: Kevin Corry Signed-off-by: Alasdair G Kergon Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/md/dm-stripe.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c index ab89278a56bf..697aacafb02a 100644 --- a/drivers/md/dm-stripe.c +++ b/drivers/md/dm-stripe.c @@ -103,9 +103,15 @@ static int stripe_ctr(struct dm_target *ti, unsigned int argc, char **argv) return -EINVAL; } + if (((uint32_t)ti->len) & (chunk_size - 1)) { + ti->error = "dm-stripe: Target length not divisible by " + "chunk size"; + return -EINVAL; + } + width = ti->len; if (sector_div(width, stripes)) { - ti->error = "dm-stripe: Target length not divisable by " + ti->error = "dm-stripe: Target length not divisible by " "number of stripes"; return -EINVAL; } From 5b40dc780ed996162f3af8712eb03beb24dcdbef Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Thu, 16 Mar 2006 23:04:07 -0800 Subject: [PATCH 08/31] [PATCH] fix race in pagevec_strip? We can call try_to_release_page() with PagePrivate off and a valid page->mapping This may cause all sorts of trouble for the filesystem *_releasepage() handlers. XFS bombs out in that case. Lock the page before checking for page private. Signed-off-by: Christoph Lameter Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/swap.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/mm/swap.c b/mm/swap.c index e9ec06d845e8..b524ea90bddb 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -393,7 +393,8 @@ void pagevec_strip(struct pagevec *pvec) struct page *page = pvec->pages[i]; if (PagePrivate(page) && !TestSetPageLocked(page)) { - try_to_release_page(page, 0); + if (PagePrivate(page)) + try_to_release_page(page, 0); unlock_page(page); } } From 85c6932ef0c7a82c309f8728ddf29768001d794e Mon Sep 17 00:00:00 2001 From: Peter Staubach Date: Thu, 16 Mar 2006 23:04:02 -0800 Subject: [PATCH 09/31] [PATCH] nfsservctl(): remove user-triggerable printk A user can use nfsservctl() to spam the logs. This can happen because the arguments to the nfsservctl() system call are versioned. This is a good thing. However, when a bad version is detected, the kernel prints a message and then returns an error. Signed-off-by: Peter Staubach Cc: Trond Myklebust Cc: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nfsctl.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/fs/nfsctl.c b/fs/nfsctl.c index 0d4cf9486068..1c72c7f85ddc 100644 --- a/fs/nfsctl.c +++ b/fs/nfsctl.c @@ -98,10 +98,8 @@ asmlinkage sys_nfsservctl(int cmd, struct nfsctl_arg __user *arg, void __user *r if (copy_from_user(&version, &arg->ca_version, sizeof(int))) return -EFAULT; - if (version != NFSCTL_VERSION) { - printk(KERN_WARNING "nfsd: incompatible version in syscall.\n"); + if (version != NFSCTL_VERSION) return -EINVAL; - } if (cmd < 0 || cmd >= sizeof(map)/sizeof(map[0]) || !map[cmd].name) return -EINVAL; From 7670f023aabd976c25862e4c6fb9f6d9d2758153 Mon Sep 17 00:00:00 2001 From: Sam Ravnborg Date: Thu, 16 Mar 2006 23:04:08 -0800 Subject: [PATCH 10/31] [PATCH] kbuild: fix buffer overflow in modpost Jiri Benc reported that modpost would stop with SIGABRT if used with long filepaths. The error looked like: > Building modules, stage 2. > MODPOST > *** glibc detected *** scripts/mod/modpost: realloc(): invalid next size: +0x0809f588 *** > [...] Fix this by allocating at least the required memory + SZ bytes each time. Before we sometimes ended up allocating too little memory resuting in the glibc detected bug above. Based on patch originally submitted by: Jiri Benc Signed-off-by: Sam Ravnborg Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- scripts/mod/modpost.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index f70ff13d4818..b8b2a560b26b 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -508,12 +508,7 @@ buf_printf(struct buffer *buf, const char *fmt, ...) va_start(ap, fmt); len = vsnprintf(tmp, SZ, fmt, ap); - if (buf->size - buf->pos < len + 1) { - buf->size += 128; - buf->p = realloc(buf->p, buf->size); - } - strncpy(buf->p + buf->pos, tmp, len + 1); - buf->pos += len; + buf_write(buf, tmp, len); va_end(ap); } @@ -521,7 +516,7 @@ void buf_write(struct buffer *buf, const char *s, int len) { if (buf->size - buf->pos < len) { - buf->size += len; + buf->size += len + SZ; buf->p = realloc(buf->p, buf->size); } strncpy(buf->p + buf->pos, s, len); From 6f5e6b9e69bf043074a0edabe3d271899c34eb79 Mon Sep 17 00:00:00 2001 From: Hugh Dickins Date: Thu, 16 Mar 2006 23:04:09 -0800 Subject: [PATCH 11/31] [PATCH] fix free swap cache latency Lee Revell reported 28ms latency when process with lots of swapped memory exits. 2.6.15 introduced a latency regression when unmapping: in accounting the zap_work latency breaker, pte_none counted 1, pte_present PAGE_SIZE, but a swap entry counted nothing at all. We think of pages present as the slow case, but Lee's trace shows that free_swap_and_cache's radix tree lookup can make a lot of work - and we could have been doing it many thousands of times without a latency break. Move the zap_work update up to account swap entries like pages present. This does account non-linear pte_file entries, and unmap_mapping_range skipping over swap entries, by the same amount even though they're quick: but neither of those cases deserves complicating the code (and they're treated no worse than they were in 2.6.14). Signed-off-by: Hugh Dickins Acked-by: Nick Piggin Acked-by: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/memory.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/mm/memory.c b/mm/memory.c index 9abc6008544b..85e80a57db29 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -623,11 +623,12 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb, (*zap_work)--; continue; } + + (*zap_work) -= PAGE_SIZE; + if (pte_present(ptent)) { struct page *page; - (*zap_work) -= PAGE_SIZE; - page = vm_normal_page(vma, addr, ptent); if (unlikely(details) && page) { /* From 265a92856b17524c87da0258ac0d3cec80ae1d35 Mon Sep 17 00:00:00 2001 From: Alexey Kuznetsov Date: Fri, 17 Mar 2006 16:05:43 -0800 Subject: [PATCH 12/31] [NET]: Fix race condition in sk_wait_event(). It is broken, the condition is checked out of socket lock. It is wonderful the bug survived for so long time. [ This fixes bugzilla #6233: race condition in tcp_sendmsg when connection became established ] Signed-off-by: David S. Miller --- include/net/sock.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/net/sock.h b/include/net/sock.h index 30758035d616..f63d0d56712c 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -478,9 +478,9 @@ static inline void sk_add_backlog(struct sock *sk, struct sk_buff *skb) rc = __condition; \ if (!rc) { \ *(__timeo) = schedule_timeout(*(__timeo)); \ - rc = __condition; \ } \ lock_sock(__sk); \ + rc = __condition; \ rc; \ }) From 7425b3403131d652c24f5047574e6c21034813ed Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Fri, 10 Mar 2006 13:47:21 +0000 Subject: [PATCH 13/31] Update MAINTAINERS entry for MIPS. Signed-off-by: Ralf Baechle --- MAINTAINERS | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index 3d7d30dc5439..8db5c339845d 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1752,7 +1752,8 @@ P: Ralf Baechle M: ralf@linux-mips.org W: http://www.linux-mips.org/ L: linux-mips@linux-mips.org -S: Maintained +T: git www.linux-mips.org:/pub/scm/linux.git +S: Supported MISCELLANEOUS MCA-SUPPORT P: James Bottomley From 3a2f735700332621274aca752be3b6f839fa47e7 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Fri, 10 Mar 2006 19:47:17 +0000 Subject: [PATCH 14/31] [MIPS] Get rid of the IP22-specific code in arclib. This breaks the kernel build if sgiwd93 was configured as a module. Signed-off-by: Ralf Baechle --- arch/mips/arc/misc.c | 19 ------------------- 1 file changed, 19 deletions(-) diff --git a/arch/mips/arc/misc.c b/arch/mips/arc/misc.c index 84867de22028..b2e10b9e9452 100644 --- a/arch/mips/arc/misc.c +++ b/arch/mips/arc/misc.c @@ -9,7 +9,6 @@ * Copyright (C) 1999 Ralf Baechle (ralf@gnu.org) * Copyright (C) 1999 Silicon Graphics, Inc. */ -#include #include #include @@ -20,17 +19,11 @@ #include #include -extern void *sgiwd93_host; -extern void reset_wd33c93(void *instance); - VOID ArcHalt(VOID) { bc_disable(); local_irq_disable(); -#ifdef CONFIG_SCSI_SGIWD93 - reset_wd33c93(sgiwd93_host); -#endif ARC_CALL0(halt); never: goto never; } @@ -40,9 +33,6 @@ ArcPowerDown(VOID) { bc_disable(); local_irq_disable(); -#ifdef CONFIG_SCSI_SGIWD93 - reset_wd33c93(sgiwd93_host); -#endif ARC_CALL0(pdown); never: goto never; } @@ -53,9 +43,6 @@ ArcRestart(VOID) { bc_disable(); local_irq_disable(); -#ifdef CONFIG_SCSI_SGIWD93 - reset_wd33c93(sgiwd93_host); -#endif ARC_CALL0(restart); never: goto never; } @@ -65,9 +52,6 @@ ArcReboot(VOID) { bc_disable(); local_irq_disable(); -#ifdef CONFIG_SCSI_SGIWD93 - reset_wd33c93(sgiwd93_host); -#endif ARC_CALL0(reboot); never: goto never; } @@ -77,9 +61,6 @@ ArcEnterInteractiveMode(VOID) { bc_disable(); local_irq_disable(); -#ifdef CONFIG_SCSI_SGIWD93 - reset_wd33c93(sgiwd93_host); -#endif ARC_CALL0(imode); never: goto never; } From a3c4946db4fe64cb21b66a09e89890678aac6d65 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Mon, 13 Mar 2006 16:16:29 +0000 Subject: [PATCH 15/31] [MIPS] SB1: Fix interrupt disable hazard. The SB1 core has a three cycle interrupt disable hazard but we were wrongly treating it as fully interlocked. Signed-off-by: Ralf Baechle --- include/asm-mips/hazards.h | 186 +++++++++++++++++++++---------------- 1 file changed, 106 insertions(+), 80 deletions(-) diff --git a/include/asm-mips/hazards.h b/include/asm-mips/hazards.h index 6111a0ce58c4..feb29a793888 100644 --- a/include/asm-mips/hazards.h +++ b/include/asm-mips/hazards.h @@ -3,7 +3,9 @@ * License. See the file "COPYING" in the main directory of this archive * for more details. * - * Copyright (C) 2003, 2004 Ralf Baechle + * Copyright (C) 2003, 2004 Ralf Baechle + * Copyright (C) MIPS Technologies, Inc. + * written by Ralf Baechle */ #ifndef _ASM_HAZARDS_H #define _ASM_HAZARDS_H @@ -74,8 +76,7 @@ #define irq_disable_hazard _ehb -#elif defined(CONFIG_CPU_R10000) || defined(CONFIG_CPU_RM9000) || \ - defined(CONFIG_CPU_SB1) +#elif defined(CONFIG_CPU_R10000) || defined(CONFIG_CPU_RM9000) /* * R10000 rocks - all hazards handled in hardware, so this becomes a nobrainer. @@ -99,13 +100,13 @@ #else /* __ASSEMBLY__ */ __asm__( - " .macro _ssnop \n\t" - " sll $0, $0, 1 \n\t" - " .endm \n\t" - " \n\t" - " .macro _ehb \n\t" - " sll $0, $0, 3 \n\t" - " .endm \n\t"); + " .macro _ssnop \n" + " sll $0, $0, 1 \n" + " .endm \n" + " \n" + " .macro _ehb \n" + " sll $0, $0, 3 \n" + " .endm \n"); #ifdef CONFIG_CPU_RM9000 @@ -117,17 +118,21 @@ __asm__( #define mtc0_tlbw_hazard() \ __asm__ __volatile__( \ - ".set\tmips32\n\t" \ - "_ssnop; _ssnop; _ssnop; _ssnop\n\t" \ - ".set\tmips0") + " .set mips32 \n" \ + " _ssnop \n" \ + " _ssnop \n" \ + " _ssnop \n" \ + " _ssnop \n" \ + " .set mips0 \n") #define tlbw_use_hazard() \ __asm__ __volatile__( \ - ".set\tmips32\n\t" \ - "_ssnop; _ssnop; _ssnop; _ssnop\n\t" \ - ".set\tmips0") - -#define back_to_back_c0_hazard() do { } while (0) + " .set mips32 \n" \ + " _ssnop \n" \ + " _ssnop \n" \ + " _ssnop \n" \ + " _ssnop \n" \ + " .set mips0 \n") #else @@ -136,15 +141,25 @@ __asm__( */ #define mtc0_tlbw_hazard() \ __asm__ __volatile__( \ - ".set noreorder\n\t" \ - "nop; nop; nop; nop; nop; nop;\n\t" \ - ".set reorder\n\t") + " .set noreorder \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " .set reorder \n") #define tlbw_use_hazard() \ __asm__ __volatile__( \ - ".set noreorder\n\t" \ - "nop; nop; nop; nop; nop; nop;\n\t" \ - ".set reorder\n\t") + " .set noreorder \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " nop \n" \ + " .set reorder \n") #endif @@ -156,49 +171,26 @@ __asm__( #ifdef CONFIG_CPU_MIPSR2 -__asm__( - " .macro\tirq_enable_hazard \n\t" - " _ehb \n\t" - " .endm \n\t" - " \n\t" - " .macro\tirq_disable_hazard \n\t" - " _ehb \n\t" - " .endm \n\t" - " \n\t" - " .macro\tback_to_back_c0_hazard \n\t" - " _ehb \n\t" - " .endm"); +__asm__(" .macro irq_enable_hazard \n" + " _ehb \n" + " .endm \n" + " \n" + " .macro irq_disable_hazard \n" + " _ehb \n" + " .endm \n"); -#define irq_enable_hazard() \ - __asm__ __volatile__( \ - "irq_enable_hazard") - -#define irq_disable_hazard() \ - __asm__ __volatile__( \ - "irq_disable_hazard") - -#define back_to_back_c0_hazard() \ - __asm__ __volatile__( \ - "back_to_back_c0_hazard") - -#elif defined(CONFIG_CPU_R10000) || defined(CONFIG_CPU_RM9000) || \ - defined(CONFIG_CPU_SB1) +#elif defined(CONFIG_CPU_R10000) || defined(CONFIG_CPU_RM9000) /* * R10000 rocks - all hazards handled in hardware, so this becomes a nobrainer. */ __asm__( - " .macro\tirq_enable_hazard \n\t" - " .endm \n\t" - " \n\t" - " .macro\tirq_disable_hazard \n\t" - " .endm"); - -#define irq_enable_hazard() do { } while (0) -#define irq_disable_hazard() do { } while (0) - -#define back_to_back_c0_hazard() do { } while (0) + " .macro irq_enable_hazard \n" + " .endm \n" + " \n" + " .macro irq_disable_hazard \n" + " .endm \n"); #else @@ -209,29 +201,63 @@ __asm__( */ __asm__( - " # \n\t" - " # There is a hazard but we do not care \n\t" - " # \n\t" - " .macro\tirq_enable_hazard \n\t" - " .endm \n\t" - " \n\t" - " .macro\tirq_disable_hazard \n\t" - " _ssnop; _ssnop; _ssnop \n\t" - " .endm"); - -#define irq_enable_hazard() do { } while (0) -#define irq_disable_hazard() \ - __asm__ __volatile__( \ - "irq_disable_hazard") - -#define back_to_back_c0_hazard() \ - __asm__ __volatile__( \ - " .set noreorder \n" \ - " nop; nop; nop \n" \ - " .set reorder \n") + " # \n" + " # There is a hazard but we do not care \n" + " # \n" + " .macro\tirq_enable_hazard \n" + " .endm \n" + " \n" + " .macro\tirq_disable_hazard \n" + " _ssnop \n" + " _ssnop \n" + " _ssnop \n" + " .endm \n"); #endif +#define irq_enable_hazard() \ + __asm__ __volatile__("irq_enable_hazard") +#define irq_disable_hazard() \ + __asm__ __volatile__("irq_disable_hazard") + + +/* + * Back-to-back hazards - + * + * What is needed to separate a move to cp0 from a subsequent read from the + * same cp0 register? + */ +#ifdef CONFIG_CPU_MIPSR2 + +__asm__(" .macro back_to_back_c0_hazard \n" + " _ehb \n" + " .endm \n"); + +#elif defined(CONFIG_CPU_R10000) || defined(CONFIG_CPU_RM9000) || \ + defined(CONFIG_CPU_SB1) + +__asm__(" .macro back_to_back_c0_hazard \n" + " .endm \n"); + +#else + +__asm__(" .macro back_to_back_c0_hazard \n" + " .set noreorder \n" + " _ssnop \n" + " _ssnop \n" + " _ssnop \n" + " .set reorder \n" + " .endm"); + +#endif + +#define back_to_back_c0_hazard() \ + __asm__ __volatile__("back_to_back_c0_hazard") + + +/* + * Instruction execution hazard + */ #ifdef CONFIG_CPU_MIPSR2 /* * gcc has a tradition of misscompiling the previous construct using the From de62893bc0725f8b5f0445250577cd7a10b2d8f8 Mon Sep 17 00:00:00 2001 From: Atsushi Nemoto Date: Mon, 13 Mar 2006 18:23:03 +0900 Subject: [PATCH 16/31] [MIPS] local_r4k_flush_cache_page fix If dcache_size != icache_size or dcache_size != scache_size, or set-associative cache, icache/scache does not flushed properly. Make blast_?cache_page_indexed() masks its index value correctly. Also, use physical address for physically indexed pcache/scache. Signed-off-by: Atsushi Nemoto Signed-off-by: Ralf Baechle --- arch/mips/mm/c-r4k.c | 13 +++++++++---- arch/mips/mm/c-tx39.c | 1 - include/asm-mips/cpu-features.h | 3 +++ include/asm-mips/cpu-info.h | 1 + include/asm-mips/r4kcache.h | 3 ++- 5 files changed, 15 insertions(+), 6 deletions(-) diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c index 0668e9bfce41..9572ed44f0d5 100644 --- a/arch/mips/mm/c-r4k.c +++ b/arch/mips/mm/c-r4k.c @@ -375,6 +375,7 @@ static void r4k_flush_cache_mm(struct mm_struct *mm) struct flush_cache_page_args { struct vm_area_struct *vma; unsigned long addr; + unsigned long pfn; }; static inline void local_r4k_flush_cache_page(void *args) @@ -382,6 +383,7 @@ static inline void local_r4k_flush_cache_page(void *args) struct flush_cache_page_args *fcp_args = args; struct vm_area_struct *vma = fcp_args->vma; unsigned long addr = fcp_args->addr; + unsigned long paddr = fcp_args->pfn << PAGE_SHIFT; int exec = vma->vm_flags & VM_EXEC; struct mm_struct *mm = vma->vm_mm; pgd_t *pgdp; @@ -431,11 +433,12 @@ static inline void local_r4k_flush_cache_page(void *args) * Do indexed flush, too much work to get the (possible) TLB refills * to work correctly. */ - addr = INDEX_BASE + (addr & (dcache_size - 1)); if (cpu_has_dc_aliases || (exec && !cpu_has_ic_fills_f_dc)) { - r4k_blast_dcache_page_indexed(addr); - if (exec && !cpu_icache_snoops_remote_store) - r4k_blast_scache_page_indexed(addr); + r4k_blast_dcache_page_indexed(cpu_has_pindexed_dcache ? + paddr : addr); + if (exec && !cpu_icache_snoops_remote_store) { + r4k_blast_scache_page_indexed(paddr); + } } if (exec) { if (cpu_has_vtag_icache) { @@ -455,6 +458,7 @@ static void r4k_flush_cache_page(struct vm_area_struct *vma, args.vma = vma; args.addr = addr; + args.pfn = pfn; on_each_cpu(local_r4k_flush_cache_page, &args, 1, 1); } @@ -956,6 +960,7 @@ static void __init probe_pcache(void) switch (c->cputype) { case CPU_20KC: case CPU_25KF: + c->dcache.flags |= MIPS_CACHE_PINDEX; case CPU_R10000: case CPU_R12000: case CPU_SB1: diff --git a/arch/mips/mm/c-tx39.c b/arch/mips/mm/c-tx39.c index 7c572bea4a98..fe232e3988e3 100644 --- a/arch/mips/mm/c-tx39.c +++ b/arch/mips/mm/c-tx39.c @@ -210,7 +210,6 @@ static void tx39_flush_cache_page(struct vm_area_struct *vma, unsigned long page * Do indexed flush, too much work to get the (possible) TLB refills * to work correctly. */ - page = (KSEG0 + (page & (dcache_size - 1))); if (cpu_has_dc_aliases || exec) tx39_blast_dcache_page_indexed(page); if (exec) diff --git a/include/asm-mips/cpu-features.h b/include/asm-mips/cpu-features.h index 78c9cc2735d5..3f2b6d9ac45e 100644 --- a/include/asm-mips/cpu-features.h +++ b/include/asm-mips/cpu-features.h @@ -96,6 +96,9 @@ #ifndef cpu_has_ic_fills_f_dc #define cpu_has_ic_fills_f_dc (cpu_data[0].icache.flags & MIPS_CACHE_IC_F_DC) #endif +#ifndef cpu_has_pindexed_dcache +#define cpu_has_pindexed_dcache (cpu_data[0].dcache.flags & MIPS_CACHE_PINDEX) +#endif /* * I-Cache snoops remote store. This only matters on SMP. Some multiprocessors diff --git a/include/asm-mips/cpu-info.h b/include/asm-mips/cpu-info.h index d5cf519f8fcc..140be1c67da7 100644 --- a/include/asm-mips/cpu-info.h +++ b/include/asm-mips/cpu-info.h @@ -39,6 +39,7 @@ struct cache_desc { #define MIPS_CACHE_ALIASES 0x00000004 /* Cache could have aliases */ #define MIPS_CACHE_IC_F_DC 0x00000008 /* Ic can refill from D-cache */ #define MIPS_IC_SNOOPS_REMOTE 0x00000010 /* Ic snoops remote stores */ +#define MIPS_CACHE_PINDEX 0x00000020 /* Physically indexed cache */ struct cpuinfo_mips { unsigned long udelay_val; diff --git a/include/asm-mips/r4kcache.h b/include/asm-mips/r4kcache.h index 9632c27dad15..0bcb79a58ee9 100644 --- a/include/asm-mips/r4kcache.h +++ b/include/asm-mips/r4kcache.h @@ -257,7 +257,8 @@ static inline void blast_##pfx##cache##lsize##_page(unsigned long page) \ \ static inline void blast_##pfx##cache##lsize##_page_indexed(unsigned long page) \ { \ - unsigned long start = page; \ + unsigned long indexmask = current_cpu_data.desc.waysize - 1; \ + unsigned long start = INDEX_BASE + (page & indexmask); \ unsigned long end = start + PAGE_SIZE; \ unsigned long ws_inc = 1UL << current_cpu_data.desc.waybit; \ unsigned long ws_end = current_cpu_data.desc.ways << \ From 86dde15b3dc36dda4bbc32a8da607675ce63ef23 Mon Sep 17 00:00:00 2001 From: Sergei Shtylylov Date: Tue, 14 Mar 2006 07:20:00 +0300 Subject: [PATCH 17/31] [MIPS] Fix DBAu1550 software power off. Signed-off-by: Sergei Shtylyov Signed-off-by: Ralf Baechle --- arch/mips/au1000/common/reset.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/arch/mips/au1000/common/reset.c b/arch/mips/au1000/common/reset.c index 4ffccedf5967..96a211400f34 100644 --- a/arch/mips/au1000/common/reset.c +++ b/arch/mips/au1000/common/reset.c @@ -164,13 +164,13 @@ void au1000_restart(char *command) void au1000_halt(void) { -#if defined(CONFIG_MIPS_PB1550) +#if defined(CONFIG_MIPS_PB1550) || defined(CONFIG_MIPS_DB1550) /* power off system */ - printk("\n** Powering off Pb1550\n"); + printk("\n** Powering off...\n"); au_writew(au_readw(0xAF00001C) | (3<<14), 0xAF00001C); au_sync(); while(1); /* should not get here */ -#endif +#else printk(KERN_NOTICE "\n** You can safely turn off the power\n"); #ifdef CONFIG_MIPS_MIRAGE au_writel((1 << 26) | (1 << 10), GPIO2_OUTPUT); @@ -187,6 +187,7 @@ void au1000_halt(void) "wait\n\t" ".set\tmips0"); #endif +#endif /* defined(CONFIG_MIPS_PB1550) || defined(CONFIG_MIPS_DB1550) */ } void au1000_power_off(void) From 66a9a4ffda3474b193f36ed579cee06c597952f3 Mon Sep 17 00:00:00 2001 From: Matej Kupljen Date: Wed, 30 Nov 2005 10:20:01 +0100 Subject: [PATCH 18/31] [MIPS] Simple patch to power off DBAU1200 Signed-off-by: Matej Kupljen Signed-off-by: Ralf Baechle --- arch/mips/au1000/common/reset.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/arch/mips/au1000/common/reset.c b/arch/mips/au1000/common/reset.c index 96a211400f34..c93af224c1b3 100644 --- a/arch/mips/au1000/common/reset.c +++ b/arch/mips/au1000/common/reset.c @@ -175,6 +175,9 @@ void au1000_halt(void) #ifdef CONFIG_MIPS_MIRAGE au_writel((1 << 26) | (1 << 10), GPIO2_OUTPUT); #endif +#ifdef CONFIG_MIPS_DB1200 + au_writew(au_readw(0xB980001C) | (1<<14), 0xB980001C); +#endif #ifdef CONFIG_PM au_sleep(); From 966f4406d903a4214fdc74bec54710c6232a95b8 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Wed, 15 Mar 2006 11:36:31 +0000 Subject: [PATCH 19/31] [MIPS] Work around bad code generation for . If a call to set_io_port_base() was being followed by usage of mips_io_port_base in the same function gcc was possibly using the old value due to some clever abuse of const. Adding a barrier will keep the optimization and result in correct code with latest gcc. Signed-off-by: Ralf Baechle --- include/asm-mips/io.h | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) diff --git a/include/asm-mips/io.h b/include/asm-mips/io.h index 8c011aa61afa..ba1d7bbc15d2 100644 --- a/include/asm-mips/io.h +++ b/include/asm-mips/io.h @@ -4,7 +4,7 @@ * for more details. * * Copyright (C) 1994, 1995 Waldorf GmbH - * Copyright (C) 1994 - 2000 Ralf Baechle + * Copyright (C) 1994 - 2000, 06 Ralf Baechle * Copyright (C) 1999, 2000 Silicon Graphics, Inc. * Copyright (C) 2004, 2005 MIPS Technologies, Inc. All rights reserved. * Author: Maciej W. Rozycki @@ -103,8 +103,20 @@ */ extern const unsigned long mips_io_port_base; -#define set_io_port_base(base) \ - do { * (unsigned long *) &mips_io_port_base = (base); } while (0) +/* + * Gcc will generate code to load the value of mips_io_port_base after each + * function call which may be fairly wasteful in some cases. So we don't + * play quite by the book. We tell gcc mips_io_port_base is a long variable + * which solves the code generation issue. Now we need to violate the + * aliasing rules a little to make initialization possible and finally we + * will need the barrier() to fight side effects of the aliasing chat. + * This trickery will eventually collapse under gcc's optimizer. Oh well. + */ +static inline void set_io_port_base(unsigned long base) +{ + * (unsigned long *) &mips_io_port_base = base; + barrier(); +} /* * Thanks to James van Artsdalen for a better timing-fix than From d6bd0e6b32ea72be91e5789b1e838c244f8a05d9 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Tue, 14 Mar 2006 23:46:58 +0000 Subject: [PATCH 20/31] [MIPS] Protect more of timer_interrupt() by xtime_lock. From Dave Johnson : * do_timer() expects the arch-specific handler to take the lock as it modifies jiffies[_64] and xtime. * writing timerhi/lo in timer_interrupt() will mess up fixed_rate_gettimeoffset() which reads timerhi/lo. Signed-off-by: Ralf Baechle --- arch/mips/kernel/time.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/arch/mips/kernel/time.c b/arch/mips/kernel/time.c index 42c94c771afb..51273b7297a7 100644 --- a/arch/mips/kernel/time.c +++ b/arch/mips/kernel/time.c @@ -424,6 +424,8 @@ irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) unsigned long j; unsigned int count; + write_seqlock(&xtime_lock); + count = mips_hpt_read(); mips_timer_ack(); @@ -441,7 +443,6 @@ irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) * CMOS clock accordingly every ~11 minutes. rtc_set_time() has to be * called as close as possible to 500 ms before the new second starts. */ - write_seqlock(&xtime_lock); if (ntp_synced() && xtime.tv_sec > last_rtc_update + 660 && (xtime.tv_nsec / 1000) >= 500000 - ((unsigned) TICK_SIZE) / 2 && @@ -453,7 +454,6 @@ irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) last_rtc_update = xtime.tv_sec - 600; } } - write_sequnlock(&xtime_lock); /* * If jiffies has overflown in this timer_interrupt, we must @@ -496,6 +496,8 @@ irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) } } + write_sequnlock(&xtime_lock); + /* * In UP mode, we call local_timer_interrupt() to do profiling * and process accouting. From a77f124294822203660b0926392b963cfa72fcf4 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Tue, 14 Mar 2006 23:47:35 +0000 Subject: [PATCH 21/31] [MIPS] Sibyte: Fix M_SCD_TIMER_INIT and M_SCD_TIMER_CNT wrong field width. From Dave Johnson : Field width should be 23 bits not 20 bits. Signed-off-by: Ralf Baechle --- include/asm-mips/sibyte/sb1250_scd.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/include/asm-mips/sibyte/sb1250_scd.h b/include/asm-mips/sibyte/sb1250_scd.h index a667bc14a7cd..f4178bdcfcb0 100644 --- a/include/asm-mips/sibyte/sb1250_scd.h +++ b/include/asm-mips/sibyte/sb1250_scd.h @@ -359,14 +359,15 @@ */ #define V_SCD_TIMER_FREQ 1000000 +#define V_SCD_TIMER_WIDTH 23 #define S_SCD_TIMER_INIT 0 -#define M_SCD_TIMER_INIT _SB_MAKEMASK(20,S_SCD_TIMER_INIT) +#define M_SCD_TIMER_INIT _SB_MAKEMASK(V_SCD_TIMER_WIDTH,S_SCD_TIMER_INIT) #define V_SCD_TIMER_INIT(x) _SB_MAKEVALUE(x,S_SCD_TIMER_INIT) #define G_SCD_TIMER_INIT(x) _SB_GETVALUE(x,S_SCD_TIMER_INIT,M_SCD_TIMER_INIT) #define S_SCD_TIMER_CNT 0 -#define M_SCD_TIMER_CNT _SB_MAKEMASK(20,S_SCD_TIMER_CNT) +#define M_SCD_TIMER_CNT _SB_MAKEMASK(V_SCD_TIMER_WIDTH,S_SCD_TIMER_CNT) #define V_SCD_TIMER_CNT(x) _SB_MAKEVALUE(x,S_SCD_TIMER_CNT) #define G_SCD_TIMER_CNT(x) _SB_GETVALUE(x,S_SCD_TIMER_CNT,M_SCD_TIMER_CNT) From 4308cb16286c96d980570cc5319173b524220c06 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Tue, 14 Mar 2006 23:52:47 +0000 Subject: [PATCH 22/31] [MIPS] Sibyte: Fix interrupt timer off by one bug. From Dave Johnson : The timers need to be loaded with 1 less than the desired interval not the interval itself. Signed-off-by: Ralf Baechle --- arch/mips/sibyte/sb1250/time.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/mips/sibyte/sb1250/time.c b/arch/mips/sibyte/sb1250/time.c index 511c89d65f38..adc0b5271a06 100644 --- a/arch/mips/sibyte/sb1250/time.c +++ b/arch/mips/sibyte/sb1250/time.c @@ -75,10 +75,10 @@ void sb1250_time_init(void) /* Disable the timer and set up the count */ __raw_writeq(0, IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_CFG))); #ifdef CONFIG_SIMULATION - __raw_writeq(50000 / HZ, + __raw_writeq((50000 / HZ) - 1, IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_INIT))); #else - __raw_writeq(1000000 / HZ, + __raw_writeq((V_SCD_TIMER_FREQ / HZ) - 1, IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_INIT))); #endif From a904f7478561464f9fe74929b81fec237b6ff4c3 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Wed, 15 Mar 2006 00:03:29 +0000 Subject: [PATCH 23/31] [MIPS] Sibyte: Fix race in sb1250_gettimeoffset(). From Dave Johnson : sb1250_gettimeoffset() simply reads the current cpu 0 timer remaining value, however once this counter reaches 0 and the interrupt is raised, it immediately resets and begins to count down again. If sb1250_gettimeoffset() is called on cpu 1 via do_gettimeofday() after the timer has reset but prior to cpu 0 processing the interrupt and taking write_seqlock() in timer_interrupt() it will return a full value (or close to it) causing time to jump backwards 1ms. Once cpu 0 handles the interrupt and timer_interrupt() gets far enough along it will jump forward 1ms. Fix this problem by implementing mips_hpt_*() on sb1250 using a spare timer unrelated to the existing periodic interrupt timers. It runs at 1Mhz with a full 23bit counter. This eliminated the custom do_gettimeoffset() for sb1250 and allowed use of the generic fixed_rate_gettimeoffset() using mips_hpt_*() and timerhi/timerlo. Signed-off-by: Ralf Baechle --- arch/mips/sibyte/sb1250/time.c | 73 ++++++++++++++++++++++++-------- arch/mips/sibyte/swarm/setup.c | 7 +++ include/asm-mips/sibyte/sb1250.h | 2 +- 3 files changed, 64 insertions(+), 18 deletions(-) diff --git a/arch/mips/sibyte/sb1250/time.c b/arch/mips/sibyte/sb1250/time.c index adc0b5271a06..1588f6debd90 100644 --- a/arch/mips/sibyte/sb1250/time.c +++ b/arch/mips/sibyte/sb1250/time.c @@ -47,23 +47,51 @@ #define IMR_IP3_VAL K_INT_MAP_I1 #define IMR_IP4_VAL K_INT_MAP_I2 +#define SB1250_HPT_NUM 3 +#define SB1250_HPT_VALUE M_SCD_TIMER_CNT /* max value */ +#define SB1250_HPT_SHIFT ((sizeof(unsigned int)*8)-V_SCD_TIMER_WIDTH) + + extern int sb1250_steal_irq(int irq); +static unsigned int sb1250_hpt_read(void); +static void sb1250_hpt_init(unsigned int); + +static unsigned int hpt_offset; + +void __init sb1250_hpt_setup(void) +{ + int cpu = smp_processor_id(); + + if (!cpu) { + /* Setup hpt using timer #3 but do not enable irq for it */ + __raw_writeq(0, IOADDR(A_SCD_TIMER_REGISTER(SB1250_HPT_NUM, R_SCD_TIMER_CFG))); + __raw_writeq(SB1250_HPT_VALUE, + IOADDR(A_SCD_TIMER_REGISTER(SB1250_HPT_NUM, R_SCD_TIMER_INIT))); + __raw_writeq(M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS, + IOADDR(A_SCD_TIMER_REGISTER(SB1250_HPT_NUM, R_SCD_TIMER_CFG))); + + /* + * we need to fill 32 bits, so just use the upper 23 bits and pretend + * the timer is going 512Mhz instead of 1Mhz + */ + mips_hpt_frequency = V_SCD_TIMER_FREQ << SB1250_HPT_SHIFT; + mips_hpt_init = sb1250_hpt_init; + mips_hpt_read = sb1250_hpt_read; + } +} + + void sb1250_time_init(void) { int cpu = smp_processor_id(); int irq = K_INT_TIMER_0+cpu; - /* Only have 4 general purpose timers */ - if (cpu > 3) { + /* Only have 4 general purpose timers, and we use last one as hpt */ + if (cpu > 2) { BUG(); } - if (!cpu) { - /* Use our own gettimeoffset() routine */ - do_gettimeoffset = sb1250_gettimeoffset; - } - sb1250_mask_irq(cpu, irq); /* Map the timer interrupt to ip[4] of this cpu */ @@ -103,7 +131,7 @@ void sb1250_timer_interrupt(struct pt_regs *regs) int cpu = smp_processor_id(); int irq = K_INT_TIMER_0 + cpu; - /* Reset the timer */ + /* ACK interrupt */ ____raw_writeq(M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS, IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_CFG))); @@ -122,15 +150,26 @@ void sb1250_timer_interrupt(struct pt_regs *regs) } /* - * We use our own do_gettimeoffset() instead of the generic one, - * because the generic one does not work for SMP case. - * In addition, since we use general timer 0 for system time, - * we can get accurate intra-jiffy offset without calibration. + * The HPT is free running from SB1250_HPT_VALUE down to 0 then starts over + * again. There's no easy way to set to a specific value so store init value + * in hpt_offset and subtract each time. + * + * Note: Timer isn't full 32bits so shift it into the upper part making + * it appear to run at a higher frequency. */ -unsigned long sb1250_gettimeoffset(void) +static unsigned int sb1250_hpt_read(void) { - unsigned long count = - __raw_readq(IOADDR(A_SCD_TIMER_REGISTER(0, R_SCD_TIMER_CNT))); + unsigned int count; - return 1000000/HZ - count; - } + count = G_SCD_TIMER_CNT(__raw_readq(IOADDR(A_SCD_TIMER_REGISTER(SB1250_HPT_NUM, R_SCD_TIMER_CNT)))); + + count = (SB1250_HPT_VALUE - count) << SB1250_HPT_SHIFT; + + return count - hpt_offset; +} + +static void sb1250_hpt_init(unsigned int count) +{ + hpt_offset = count; + return; +} diff --git a/arch/mips/sibyte/swarm/setup.c b/arch/mips/sibyte/swarm/setup.c index b614ca0ddb69..b661d2425a36 100644 --- a/arch/mips/sibyte/swarm/setup.c +++ b/arch/mips/sibyte/swarm/setup.c @@ -70,6 +70,12 @@ const char *get_system_type(void) return "SiByte " SIBYTE_BOARD_NAME; } +void __init swarm_time_init(void) +{ + /* Setup HPT */ + sb1250_hpt_setup(); +} + void __init swarm_timer_setup(struct irqaction *irq) { /* @@ -109,6 +115,7 @@ void __init plat_setup(void) panic_timeout = 5; /* For debug. */ + board_time_init = swarm_time_init; board_timer_setup = swarm_timer_setup; board_be_handler = swarm_be_handler; diff --git a/include/asm-mips/sibyte/sb1250.h b/include/asm-mips/sibyte/sb1250.h index a474c29cd701..b09e16c93ca0 100644 --- a/include/asm-mips/sibyte/sb1250.h +++ b/include/asm-mips/sibyte/sb1250.h @@ -45,8 +45,8 @@ extern unsigned int soc_type; extern unsigned int periph_rev; extern unsigned int zbbus_mhz; +extern void sb1250_hpt_setup(void); extern void sb1250_time_init(void); -extern unsigned long sb1250_gettimeoffset(void); extern void sb1250_mask_irq(int cpu, int irq); extern void sb1250_unmask_irq(int cpu, int irq); extern void sb1250_smp_finish(void); From 9007c9a2b03ea325ee593a161dbf01dbb8222d7e Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Sat, 18 Mar 2006 16:59:31 +0000 Subject: [PATCH 24/31] [MIPS] SB1: Check for -mno-sched-prolog if building corelis debug kernel. Signed-off-by: Ralf Baechle --- arch/mips/Makefile | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/mips/Makefile b/arch/mips/Makefile index fe9da16f3a40..3d8dac681c63 100644 --- a/arch/mips/Makefile +++ b/arch/mips/Makefile @@ -108,7 +108,8 @@ MODFLAGS += -mlong-calls cflags-$(CONFIG_CPU_BIG_ENDIAN) += $(shell $(CC) -dumpmachine |grep -q 'mips.*el-.*' && echo -EB) cflags-$(CONFIG_CPU_LITTLE_ENDIAN) += $(shell $(CC) -dumpmachine |grep -q 'mips.*el-.*' || echo -EL) -cflags-$(CONFIG_SB1XXX_CORELIS) += -mno-sched-prolog -fno-omit-frame-pointer +cflags-$(CONFIG_SB1XXX_CORELIS) += $(call cc-option,-mno-sched-prolog) \ + -fno-omit-frame-pointer # # Use: $(call set_gccflags,,,,,) From 2d61b86775a5676a8fba2ba2f0f869564e35c630 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Sat, 18 Mar 2006 20:41:10 +0300 Subject: [PATCH 25/31] [PATCH] disable unshare(CLONE_VM) for now sys_unshare() does mmput(new_mm). This is not enough if we have mm->core_waiters. This patch is a temporary fix for soon to be released 2.6.16. Signed-off-by: Oleg Nesterov [ Checked with Uli: "I'm not planning to use unshare(CLONE_VM). It's not needed for any functionality planned so far. What we (as in Red Hat) need unshare() for now is the filesystem side." ] Signed-off-by: Linus Torvalds --- kernel/fork.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/kernel/fork.c b/kernel/fork.c index 46060cb24af0..b373322ca497 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1478,9 +1478,7 @@ static int unshare_vm(unsigned long unshare_flags, struct mm_struct **new_mmp) if ((unshare_flags & CLONE_VM) && (mm && atomic_read(&mm->mm_users) > 1)) { - *new_mmp = dup_mm(current); - if (!*new_mmp) - return -ENOMEM; + return -EINVAL; } return 0; From 7215fdb214ad8d76c9d667be07f870f0f0544959 Mon Sep 17 00:00:00 2001 From: Markus Rechberger Date: Tue, 7 Feb 2006 08:49:13 -0200 Subject: [PATCH 26/31] [PATCH] Fixed em28xx based system lockup Fixed em28xx based system lockup, device needs to be initialized before starting the isoc transfer otherwise the system will completly lock up. Signed-off-by: Markus Rechberger Signed-off-by: Mauro Carvalho Chehab Signed-off-by: Linus Torvalds --- drivers/media/video/em28xx/em28xx-video.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/media/video/em28xx/em28xx-video.c b/drivers/media/video/em28xx/em28xx-video.c index 94a14a2bb6d6..5b267808a9d4 100644 --- a/drivers/media/video/em28xx/em28xx-video.c +++ b/drivers/media/video/em28xx/em28xx-video.c @@ -367,6 +367,9 @@ static int em28xx_v4l2_open(struct inode *inode, struct file *filp) em28xx_capture_start(dev, 1); em28xx_resolution_set(dev); + /* device needs to be initialized before isoc transfer */ + video_mux(dev, 0); + /* start the transfer */ errCode = em28xx_init_isoc(dev); if (errCode) From 60a6dc55b93b3321afa52f650a149fb7e87fa85a Mon Sep 17 00:00:00 2001 From: Michael Krufky Date: Wed, 15 Mar 2006 02:36:13 -0300 Subject: [PATCH 27/31] [PATCH] Kconfig: swap VIDEO_CX88_ALSA and VIDEO_CX88_DVB VIDEO_CX88_ALSA should not be between VIDEO_CX88_DVB and VIDEO_CX88_DVB_ALL_FRONTENDS When cx88-alsa was added to cx88/Kconfig, it was added in between VIDEO_CX88_DVB and VIDEO_CX88_DVB_ALL_FRONTENDS. This caused undesireable effects to the appearance of the menu options in menuconfig. This fix reorders cx88-alsa and cx88-dvb in Kconfig, to match saa7134, and restore the correct menuconfig appearance. Signed-off-by: Michael Krufky Signed-off-by: Mauro Carvalho Chehab Signed-off-by: Linus Torvalds --- drivers/media/video/cx88/Kconfig | 28 ++++++++++++++-------------- 1 file changed, 14 insertions(+), 14 deletions(-) diff --git a/drivers/media/video/cx88/Kconfig b/drivers/media/video/cx88/Kconfig index e99dfbbf3e95..87d79df05336 100644 --- a/drivers/media/video/cx88/Kconfig +++ b/drivers/media/video/cx88/Kconfig @@ -15,20 +15,6 @@ config VIDEO_CX88 To compile this driver as a module, choose M here: the module will be called cx8800 -config VIDEO_CX88_DVB - tristate "DVB/ATSC Support for cx2388x based TV cards" - depends on VIDEO_CX88 && DVB_CORE - select VIDEO_BUF_DVB - ---help--- - This adds support for DVB/ATSC cards based on the - Connexant 2388x chip. - - To compile this driver as a module, choose M here: the - module will be called cx88-dvb. - - You must also select one or more DVB/ATSC demodulators. - If you are unsure which you need, choose all of them. - config VIDEO_CX88_ALSA tristate "ALSA DMA audio support" depends on VIDEO_CX88 && SND && EXPERIMENTAL @@ -44,6 +30,20 @@ config VIDEO_CX88_ALSA To compile this driver as a module, choose M here: the module will be called cx88-alsa. +config VIDEO_CX88_DVB + tristate "DVB/ATSC Support for cx2388x based TV cards" + depends on VIDEO_CX88 && DVB_CORE + select VIDEO_BUF_DVB + ---help--- + This adds support for DVB/ATSC cards based on the + Connexant 2388x chip. + + To compile this driver as a module, choose M here: the + module will be called cx88-dvb. + + You must also select one or more DVB/ATSC demodulators. + If you are unsure which you need, choose all of them. + config VIDEO_CX88_DVB_ALL_FRONTENDS bool "Build all supported frontends for cx2388x based TV cards" default y From c7c694d196a39af6e644e24279953d04f30362db Mon Sep 17 00:00:00 2001 From: Ralf Baechle DL5RB Date: Sun, 19 Mar 2006 13:20:06 -0800 Subject: [PATCH 28/31] [AX.25]: Fix potencial memory hole. If the AX.25 dialect chosen by the sysadmin is set to DAMA master / 3 (or DAMA slave / 2, if CONFIG_AX25_DAMA_SLAVE=n) ax25_kick() will fall through the switch statement without calling ax25_send_iframe() or any other function that would eventually free skbn thus leaking the packet. Fix by restricting the sysctl inferface to allow only actually supported AX.25 dialects. The system administration mistake needed for this to happen is rather unlikely, so this is an uncritical hole. Coverity #651. Signed-off-by: Ralf Baechle DL5RB Signed-off-by: David S. Miller --- include/net/ax25.h | 9 ++++++++- net/ax25/sysctl_net_ax25.c | 2 +- 2 files changed, 9 insertions(+), 2 deletions(-) diff --git a/include/net/ax25.h b/include/net/ax25.h index 2250a18b0cbb..d052b221dbcd 100644 --- a/include/net/ax25.h +++ b/include/net/ax25.h @@ -110,8 +110,15 @@ enum { enum { AX25_PROTO_STD_SIMPLEX, AX25_PROTO_STD_DUPLEX, +#ifdef CONFIG_AX25_DAMA_SLAVE AX25_PROTO_DAMA_SLAVE, - AX25_PROTO_DAMA_MASTER +#ifdef CONFIG_AX25_DAMA_MASTER + AX25_PROTO_DAMA_MASTER, +#define AX25_PROTO_MAX AX25_PROTO_DAMA_MASTER +#endif +#endif + __AX25_PROTO_MAX, + AX25_PROTO_MAX = __AX25_PROTO_MAX -1 }; enum { diff --git a/net/ax25/sysctl_net_ax25.c b/net/ax25/sysctl_net_ax25.c index f67711f2ee96..894a22558d9d 100644 --- a/net/ax25/sysctl_net_ax25.c +++ b/net/ax25/sysctl_net_ax25.c @@ -24,7 +24,7 @@ static int min_t3[1], max_t3[] = {3600 * HZ}; static int min_idle[1], max_idle[] = {65535 * HZ}; static int min_n2[] = {1}, max_n2[] = {31}; static int min_paclen[] = {1}, max_paclen[] = {512}; -static int min_proto[1], max_proto[] = {3}; +static int min_proto[1], max_proto[] = { AX25_PROTO_MAX }; static int min_ds_timeout[1], max_ds_timeout[] = {65535 * HZ}; static struct ctl_table_header *ax25_table_header; From 4a29cc2e503b33a1e96db4c3f9a94165f153f259 Mon Sep 17 00:00:00 2001 From: Michael Chan Date: Sun, 19 Mar 2006 13:21:12 -0800 Subject: [PATCH 29/31] [TG3]: 40-bit DMA workaround part 2 The 40-bit DMA workaround recently implemented for 5714, 5715, and 5780 needs to be expanded because there may be other tg3 devices behind the EPB Express to PCIX bridge in the 5780 class device. For example, some 4-port card or mother board designs have 5704 behind the 5714. All devices behind the EPB require the 40-bit DMA workaround. Thanks to Chris Elmquist again for reporting the problem and testing the patch. Signed-off-by: Michael Chan Signed-off-by: David S. Miller --- drivers/net/tg3.c | 52 ++++++++++++++++++++++++++++++++++------- drivers/net/tg3.h | 1 + include/linux/pci_ids.h | 1 + 3 files changed, 45 insertions(+), 9 deletions(-) diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c index b8f1524da557..caf4102b54ce 100644 --- a/drivers/net/tg3.c +++ b/drivers/net/tg3.c @@ -9552,12 +9552,36 @@ static int __devinit tg3_get_invariants(struct tg3 *tp) } } - /* Find msi capability. */ + /* The EPB bridge inside 5714, 5715, and 5780 cannot support + * DMA addresses > 40-bit. This bridge may have other additional + * 57xx devices behind it in some 4-port NIC designs for example. + * Any tg3 device found behind the bridge will also need the 40-bit + * DMA workaround. + */ if (GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5780 || GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5714) { tp->tg3_flags2 |= TG3_FLG2_5780_CLASS; + tp->tg3_flags |= TG3_FLAG_40BIT_DMA_BUG; tp->msi_cap = pci_find_capability(tp->pdev, PCI_CAP_ID_MSI); } + else { + struct pci_dev *bridge = NULL; + + do { + bridge = pci_get_device(PCI_VENDOR_ID_SERVERWORKS, + PCI_DEVICE_ID_SERVERWORKS_EPB, + bridge); + if (bridge && bridge->subordinate && + (bridge->subordinate->number <= + tp->pdev->bus->number) && + (bridge->subordinate->subordinate >= + tp->pdev->bus->number)) { + tp->tg3_flags |= TG3_FLAG_40BIT_DMA_BUG; + pci_dev_put(bridge); + break; + } + } while (bridge); + } /* Initialize misc host control in PCI block. */ tp->misc_host_ctrl |= (misc_ctrl_reg & @@ -10303,7 +10327,14 @@ static int __devinit tg3_test_dma(struct tg3 *tp) GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5704) { u32 ccval = (tr32(TG3PCI_CLOCK_CTRL) & 0x1f); - if (ccval == 0x6 || ccval == 0x7) + /* If the 5704 is behind the EPB bridge, we can + * do the less restrictive ONE_DMA workaround for + * better performance. + */ + if ((tp->tg3_flags & TG3_FLAG_40BIT_DMA_BUG) && + GET_ASIC_REV(tp->pci_chip_rev_id) == ASIC_REV_5704) + tp->dma_rwctrl |= 0x8000; + else if (ccval == 0x6 || ccval == 0x7) tp->dma_rwctrl |= DMA_RWCTRL_ONE_DMA; /* Set bit 23 to enable PCIX hw bug fix */ @@ -10759,19 +10790,20 @@ static int __devinit tg3_init_one(struct pci_dev *pdev, goto err_out_iounmap; } - /* 5714, 5715 and 5780 cannot support DMA addresses > 40-bit. + /* The EPB bridge inside 5714, 5715, and 5780 and any + * device behind the EPB cannot support DMA addresses > 40-bit. * On 64-bit systems with IOMMU, use 40-bit dma_mask. * On 64-bit systems without IOMMU, use 64-bit dma_mask and * do DMA address check in tg3_start_xmit(). */ - if (tp->tg3_flags2 & TG3_FLG2_5780_CLASS) { + if (tp->tg3_flags2 & TG3_FLG2_IS_5788) + persist_dma_mask = dma_mask = DMA_32BIT_MASK; + else if (tp->tg3_flags & TG3_FLAG_40BIT_DMA_BUG) { persist_dma_mask = dma_mask = DMA_40BIT_MASK; #ifdef CONFIG_HIGHMEM dma_mask = DMA_64BIT_MASK; #endif - } else if (tp->tg3_flags2 & TG3_FLG2_IS_5788) - persist_dma_mask = dma_mask = DMA_32BIT_MASK; - else + } else persist_dma_mask = dma_mask = DMA_64BIT_MASK; /* Configure DMA attributes. */ @@ -10908,8 +10940,10 @@ static int __devinit tg3_init_one(struct pci_dev *pdev, (tp->tg3_flags & TG3_FLAG_SPLIT_MODE) != 0, (tp->tg3_flags2 & TG3_FLG2_NO_ETH_WIRE_SPEED) == 0, (tp->tg3_flags2 & TG3_FLG2_TSO_CAPABLE) != 0); - printk(KERN_INFO "%s: dma_rwctrl[%08x]\n", - dev->name, tp->dma_rwctrl); + printk(KERN_INFO "%s: dma_rwctrl[%08x] dma_mask[%d-bit]\n", + dev->name, tp->dma_rwctrl, + (pdev->dma_mask == DMA_32BIT_MASK) ? 32 : + (((u64) pdev->dma_mask == DMA_40BIT_MASK) ? 40 : 64)); return 0; diff --git a/drivers/net/tg3.h b/drivers/net/tg3.h index 7f4b7f6ac40d..7e3b613afb29 100644 --- a/drivers/net/tg3.h +++ b/drivers/net/tg3.h @@ -2163,6 +2163,7 @@ struct tg3 { #define TG3_FLAG_10_100_ONLY 0x01000000 #define TG3_FLAG_PAUSE_AUTONEG 0x02000000 #define TG3_FLAG_IN_RESET_TASK 0x04000000 +#define TG3_FLAG_40BIT_DMA_BUG 0x08000000 #define TG3_FLAG_BROKEN_CHECKSUMS 0x10000000 #define TG3_FLAG_GOT_SERDES_FLOWCTL 0x20000000 #define TG3_FLAG_SPLIT_MODE 0x40000000 diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h index 1709b5009d2e..751eea58bde8 100644 --- a/include/linux/pci_ids.h +++ b/include/linux/pci_ids.h @@ -1365,6 +1365,7 @@ #define PCI_DEVICE_ID_SERVERWORKS_HE 0x0008 #define PCI_DEVICE_ID_SERVERWORKS_LE 0x0009 #define PCI_DEVICE_ID_SERVERWORKS_GCNB_LE 0x0017 +#define PCI_DEVICE_ID_SERVERWORKS_EPB 0x0103 #define PCI_DEVICE_ID_SERVERWORKS_OSB4 0x0200 #define PCI_DEVICE_ID_SERVERWORKS_CSB5 0x0201 #define PCI_DEVICE_ID_SERVERWORKS_CSB6 0x0203 From 2be1aaf988e8b6775eece8374f4fd195b21cc4c4 Mon Sep 17 00:00:00 2001 From: Andrea Arcangeli Date: Sun, 19 Mar 2006 19:04:17 +0100 Subject: [PATCH 30/31] [PATCH] Remove obsolete CREDITS address This address is going to be obsolete, so I should update it. --- CREDITS | 1 - 1 file changed, 1 deletion(-) diff --git a/CREDITS b/CREDITS index 64511e2abc8e..1f171f1033bd 100644 --- a/CREDITS +++ b/CREDITS @@ -120,7 +120,6 @@ D: Author of lil (Linux Interrupt Latency benchmark) D: Fixed the shm swap deallocation at swapoff time (try_to_unuse message) D: VM hacker D: Various other kernel hacks -S: Via Cicalini 26 S: Imola 40026 S: Italy From 7705a8792b0fc82fd7d4dd923724606bbfd9fb20 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 19 Mar 2006 21:53:29 -0800 Subject: [PATCH 31/31] Linux 2.6.16 --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index ac54c03c1b1a..cb5790580fca 100644 --- a/Makefile +++ b/Makefile @@ -1,7 +1,7 @@ VERSION = 2 PATCHLEVEL = 6 SUBLEVEL = 16 -EXTRAVERSION =-rc6 +EXTRAVERSION = NAME=Sliding Snow Leopard # *DOCUMENTATION*