From abbea7187296a7fb316a55f2319438c2bf881f0a Mon Sep 17 00:00:00 2001 From: Andrew Victor Date: Fri, 24 Feb 2006 22:27:50 +0000 Subject: [ARM] 3348/1: Disable GPIO interrupts Patch from Andrew Victor disable_irq() lazily disables the interrupt, so the IRQ is only disabled once the interrupt occurs again. The GPIO interrupt handler therefore must first check disable_depth to see if the IRQ needs to be disabled. Orignal patch by Bill Gatliff. Signed-off-by: Andrew Victor Signed-off-by: Russell King --- arch/arm/mach-at91rm9200/gpio.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/arm/mach-at91rm9200/gpio.c b/arch/arm/mach-at91rm9200/gpio.c index a9f718bf8ba..0e396feec46 100644 --- a/arch/arm/mach-at91rm9200/gpio.c +++ b/arch/arm/mach-at91rm9200/gpio.c @@ -274,8 +274,18 @@ static void gpio_irq_handler(unsigned irq, struct irqdesc *desc, struct pt_regs gpio = &irq_desc[pin]; while (isr) { - if (isr & 1) - gpio->handle(pin, gpio, regs); + if (isr & 1) { + if (unlikely(gpio->disable_depth)) { + /* + * The core ARM interrupt handler lazily disables IRQs so + * another IRQ must be generated before it actually gets + * here to be disabled on the GPIO controller. + */ + gpio_irq_mask(pin); + } + else + gpio->handle(pin, gpio, regs); + } pin++; gpio++; isr >>= 1; -- cgit v1.2.3 From cde05cf2145b0aa06dd61277060bfba5d38acb0b Mon Sep 17 00:00:00 2001 From: Hirokazu Takata Date: Fri, 24 Feb 2006 13:03:50 -0800 Subject: [PATCH] m32r: enable asm code optimization Add -O2 option to AFLAGS to enable asm code optimization for m32r. On m32r gas, "-m32r2 -O" option enables assembler's parallel code generation optimization for M32R2 ISA as a default. So, "-no-parallel" option is required explicitly for a cpu core with single instuction issuing, for example, VDEC2. Signed-off-by: Hirokazu Takata Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/m32r/Makefile | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch') diff --git a/arch/m32r/Makefile b/arch/m32r/Makefile index 983d438b14b..4b3c90ba926 100644 --- a/arch/m32r/Makefile +++ b/arch/m32r/Makefile @@ -12,14 +12,14 @@ CFLAGS_MODULE += -mmodel=large ifdef CONFIG_CHIP_VDEC2 cflags-$(CONFIG_ISA_M32R2) += -DNO_FPU -Wa,-bitinst -aflags-$(CONFIG_ISA_M32R2) += -DNO_FPU -Wa,-bitinst +aflags-$(CONFIG_ISA_M32R2) += -DNO_FPU -O2 -Wa,-bitinst -Wa,-no-parallel else cflags-$(CONFIG_ISA_M32R2) += -DNO_FPU -m32r2 -aflags-$(CONFIG_ISA_M32R2) += -DNO_FPU -m32r2 +aflags-$(CONFIG_ISA_M32R2) += -DNO_FPU -m32r2 -O2 endif cflags-$(CONFIG_ISA_M32R) += -DNO_FPU -aflags-$(CONFIG_ISA_M32R) += -DNO_FPU -Wa,-no-bitinst +aflags-$(CONFIG_ISA_M32R) += -DNO_FPU -O2 -Wa,-no-bitinst CFLAGS += $(cflags-y) AFLAGS += $(aflags-y) -- cgit v1.2.3 From 6ced13cdcab440931b87829b0f2d0dedacfb3f2d Mon Sep 17 00:00:00 2001 From: Hirokazu Takata Date: Fri, 24 Feb 2006 13:03:51 -0800 Subject: [PATCH] m32r: fix and update for gcc-4.0 Fix and update for gcc-4.0. - arch/m32r/kernel/signal.c: Change type of the 8th parameter of sys_rt_sigsuspend() from 'struct pt_regs' to 'struct pt_regs *'. This functions make use of the 'regs' parameter to return status value, but gcc-4.0 optimizes and removes it as a dead code. Functions, sys_sigaltstack() and sys_rt_sigreturn(), have also modified. - arch/m32r/lib/usercopy.c, include/asm-m32r/uaccess.h: Add early-clobber constraints('&') to output values of asm statements; these constraints seems to be required for gcc-4.0 register assignment. Signed-off-by: Hayato Fujiwara Signed-off-by: Hirokazu Takata Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/m32r/kernel/signal.c | 24 ++++++++++-------------- arch/m32r/lib/usercopy.c | 4 ++-- 2 files changed, 12 insertions(+), 16 deletions(-) (limited to 'arch') diff --git a/arch/m32r/kernel/signal.c b/arch/m32r/kernel/signal.c index 71763f7a1d1..cb33097fefc 100644 --- a/arch/m32r/kernel/signal.c +++ b/arch/m32r/kernel/signal.c @@ -36,7 +36,7 @@ int do_signal(struct pt_regs *, sigset_t *); asmlinkage int sys_rt_sigsuspend(sigset_t *unewset, size_t sigsetsize, unsigned long r2, unsigned long r3, unsigned long r4, - unsigned long r5, unsigned long r6, struct pt_regs regs) + unsigned long r5, unsigned long r6, struct pt_regs *regs) { sigset_t saveset, newset; @@ -54,21 +54,21 @@ sys_rt_sigsuspend(sigset_t *unewset, size_t sigsetsize, recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); - regs.r0 = -EINTR; + regs->r0 = -EINTR; while (1) { current->state = TASK_INTERRUPTIBLE; schedule(); - if (do_signal(®s, &saveset)) - return regs.r0; + if (do_signal(regs, &saveset)) + return regs->r0; } } asmlinkage int sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss, unsigned long r2, unsigned long r3, unsigned long r4, - unsigned long r5, unsigned long r6, struct pt_regs regs) + unsigned long r5, unsigned long r6, struct pt_regs *regs) { - return do_sigaltstack(uss, uoss, regs.spu); + return do_sigaltstack(uss, uoss, regs->spu); } @@ -140,11 +140,10 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, asmlinkage int sys_rt_sigreturn(unsigned long r0, unsigned long r1, unsigned long r2, unsigned long r3, unsigned long r4, - unsigned long r5, unsigned long r6, struct pt_regs regs) + unsigned long r5, unsigned long r6, struct pt_regs *regs) { - struct rt_sigframe __user *frame = (struct rt_sigframe __user *)regs.spu; + struct rt_sigframe __user *frame = (struct rt_sigframe __user *)regs->spu; sigset_t set; - stack_t st; int result; if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) @@ -158,14 +157,11 @@ sys_rt_sigreturn(unsigned long r0, unsigned long r1, recalc_sigpending(); spin_unlock_irq(¤t->sighand->siglock); - if (restore_sigcontext(®s, &frame->uc.uc_mcontext, &result)) + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &result)) goto badframe; - if (__copy_from_user(&st, &frame->uc.uc_stack, sizeof(st))) + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->spu) == -EFAULT) goto badframe; - /* It is more difficult to avoid calling this function than to - call it and ignore errors. */ - do_sigaltstack(&st, NULL, regs.spu); return result; diff --git a/arch/m32r/lib/usercopy.c b/arch/m32r/lib/usercopy.c index ce16bbe26a5..2d1dd2106c4 100644 --- a/arch/m32r/lib/usercopy.c +++ b/arch/m32r/lib/usercopy.c @@ -64,7 +64,7 @@ do { \ " .balign 4\n" \ " .long 0b,3b\n" \ ".previous" \ - : "=r"(res), "=r"(count), "=&r" (__d0), "=&r" (__d1), \ + : "=&r"(res), "=&r"(count), "=&r" (__d0), "=&r" (__d1), \ "=&r" (__d2) \ : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), \ "4"(dst) \ @@ -101,7 +101,7 @@ do { \ " .balign 4\n" \ " .long 0b,3b\n" \ ".previous" \ - : "=r"(res), "=r"(count), "=&r" (__d0), "=&r" (__d1), \ + : "=&r"(res), "=&r"(count), "=&r" (__d0), "=&r" (__d1), \ "=&r" (__d2) \ : "i"(-EFAULT), "0"(count), "1"(count), "3"(src), \ "4"(dst) \ -- cgit v1.2.3 From 31bc5a33346b6dd35be219d1416449e0064e9123 Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Fri, 24 Feb 2006 13:03:53 -0800 Subject: [PATCH] uml: correct error messages in COW driver Improve some error messages in the COW driver, and say V3, not V2, when talking about V3 format. Also resync with our userspace code utility a bit more. Signed-off-by: Paolo 'Blaisorblade' Giarrusso Acked-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/um/drivers/cow_sys.h | 2 +- arch/um/drivers/cow_user.c | 21 ++++++++++++--------- 2 files changed, 13 insertions(+), 10 deletions(-) (limited to 'arch') diff --git a/arch/um/drivers/cow_sys.h b/arch/um/drivers/cow_sys.h index c83fc5d6893..df25263d2ad 100644 --- a/arch/um/drivers/cow_sys.h +++ b/arch/um/drivers/cow_sys.h @@ -33,7 +33,7 @@ static inline int cow_file_size(char *file, unsigned long long *size_out) return(os_file_size(file, size_out)); } -static inline int cow_write_file(int fd, char *buf, int size) +static inline int cow_write_file(int fd, void *buf, int size) { return(os_write_file(fd, buf, size)); } diff --git a/arch/um/drivers/cow_user.c b/arch/um/drivers/cow_user.c index fbe2217db5d..d1c86bc377b 100644 --- a/arch/um/drivers/cow_user.c +++ b/arch/um/drivers/cow_user.c @@ -176,7 +176,7 @@ int write_cow_header(char *cow_file, int fd, char *backing_file, err = -ENOMEM; header = cow_malloc(sizeof(*header)); if(header == NULL){ - cow_printf("Failed to allocate COW V3 header\n"); + cow_printf("write_cow_header - failed to allocate COW V3 header\n"); goto out; } header->magic = htonl(COW_MAGIC); @@ -196,15 +196,17 @@ int write_cow_header(char *cow_file, int fd, char *backing_file, err = os_file_modtime(header->backing_file, &modtime); if(err < 0){ - cow_printf("Backing file '%s' mtime request failed, " - "err = %d\n", header->backing_file, -err); + cow_printf("write_cow_header - backing file '%s' mtime " + "request failed, err = %d\n", header->backing_file, + -err); goto out_free; } err = cow_file_size(header->backing_file, size); if(err < 0){ - cow_printf("Couldn't get size of backing file '%s', " - "err = %d\n", header->backing_file, -err); + cow_printf("write_cow_header - couldn't get size of " + "backing file '%s', err = %d\n", + header->backing_file, -err); goto out_free; } @@ -214,10 +216,11 @@ int write_cow_header(char *cow_file, int fd, char *backing_file, header->alignment = htonl(alignment); header->cow_format = COW_BITMAP; - err = os_write_file(fd, header, sizeof(*header)); + err = cow_write_file(fd, header, sizeof(*header)); if(err != sizeof(*header)){ - cow_printf("Write of header to new COW file '%s' failed, " - "err = %d\n", cow_file, -err); + cow_printf("write_cow_header - write of header to " + "new COW file '%s' failed, err = %d\n", cow_file, + -err); goto out_free; } err = 0; @@ -299,7 +302,7 @@ int read_cow_header(int (*reader)(__u64, char *, int, void *), void *arg, } else if(version == 3){ if(n < sizeof(header->v3)){ - cow_printf("read_cow_header - failed to read V2 " + cow_printf("read_cow_header - failed to read V3 " "header\n"); goto out; } -- cgit v1.2.3 From 07f4e2c61c76e8b543c0a2589063aea85c15fb25 Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Fri, 24 Feb 2006 13:03:55 -0800 Subject: [PATCH] uml: fix usage of kernel_errno in place of errno To avoid conflicts, in kernel files errno is expanded to kernel_errno, to distinguish it from glibc errno. In this case, the code wants to use the libc errno but the kernel one is used; in the other usage, we return errno in place of -errno in case of an error. Signed-off-by: Paolo 'Blaisorblade' Giarrusso Acked-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/um/include/os.h | 3 +++ arch/um/os-Linux/process.c | 16 ++++++++++++++++ arch/um/sys-i386/ldt.c | 9 +++------ 3 files changed, 22 insertions(+), 6 deletions(-) (limited to 'arch') diff --git a/arch/um/include/os.h b/arch/um/include/os.h index eb1710b8125..2a1c64d8d0b 100644 --- a/arch/um/include/os.h +++ b/arch/um/include/os.h @@ -179,8 +179,11 @@ extern void os_stop_process(int pid); extern void os_kill_process(int pid, int reap_child); extern void os_kill_ptraced_process(int pid, int reap_child); extern void os_usr1_process(int pid); +extern long os_ptrace_ldt(long pid, long addr, long data); + extern int os_getpid(void); extern int os_getpgrp(void); + extern void init_new_thread_stack(void *sig_stack, void (*usr1_handler)(int)); extern void init_new_thread_signals(int altstack); extern int run_kernel_thread(int (*fn)(void *), void *arg, void **jmp_ptr); diff --git a/arch/um/os-Linux/process.c b/arch/um/os-Linux/process.c index 7f5e2dac2a3..d261888f39c 100644 --- a/arch/um/os-Linux/process.c +++ b/arch/um/os-Linux/process.c @@ -19,6 +19,7 @@ #include "irq_user.h" #include "kern_util.h" #include "longjmp.h" +#include "skas_ptrace.h" #define ARBITRARY_ADDR -1 #define FAILURE_PID -1 @@ -100,6 +101,21 @@ void os_kill_process(int pid, int reap_child) } +/* This is here uniquely to have access to the userspace errno, i.e. the one + * used by ptrace in case of error. + */ + +long os_ptrace_ldt(long pid, long addr, long data) +{ + int ret; + + ret = ptrace(PTRACE_LDT, pid, addr, data); + + if (ret < 0) + return -errno; + return ret; +} + /* Kill off a ptraced child by all means available. kill it normally first, * then PTRACE_KILL it, then PTRACE_CONT it in case it's in a run state from * which it can't exit directly. diff --git a/arch/um/sys-i386/ldt.c b/arch/um/sys-i386/ldt.c index 1fa09a79a10..fe0877b3509 100644 --- a/arch/um/sys-i386/ldt.c +++ b/arch/um/sys-i386/ldt.c @@ -107,7 +107,7 @@ long write_ldt_entry(struct mm_id * mm_idp, int func, struct user_desc * desc, * So we need to switch child's mm into our userspace, then * later switch back. * - * Note: I'm unshure: should interrupts be disabled here? + * Note: I'm unsure: should interrupts be disabled here? */ if(!current->active_mm || current->active_mm == &init_mm || mm_idp != ¤t->active_mm->context.skas.id) @@ -129,9 +129,7 @@ long write_ldt_entry(struct mm_id * mm_idp, int func, struct user_desc * desc, pid = userspace_pid[cpu]; } - res = ptrace(PTRACE_LDT, pid, 0, (unsigned long) &ldt_op); - if(res) - res = errno; + res = os_ptrace_ldt(pid, 0, (unsigned long) &ldt_op); if(proc_mm) put_cpu(); @@ -181,8 +179,7 @@ static long read_ldt_from_host(void __user * ptr, unsigned long bytecount) */ cpu = get_cpu(); - res = ptrace(PTRACE_LDT, userspace_pid[cpu], 0, - (unsigned long) &ptrace_ldt); + res = os_ptrace_ldt(userspace_pid[cpu], 0, (unsigned long) &ptrace_ldt); put_cpu(); if(res < 0) goto out; -- cgit v1.2.3 From 635dd50b7dc69b698e8808ff2802a6cfc31385a8 Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Fri, 24 Feb 2006 13:03:55 -0800 Subject: [PATCH] uml: fix ((unused)) attribute Use __attribute_used__ instead of __attribute__ ((unused)). This will help with GCC > 3.2. Signed-off-by: Paolo 'Blaisorblade' Giarrusso Acked-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/um/include/init.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/um/include/init.h b/arch/um/include/init.h index cbd79a8d213..d4de7c0120c 100644 --- a/arch/um/include/init.h +++ b/arch/um/include/init.h @@ -122,7 +122,7 @@ extern struct uml_param __uml_setup_start, __uml_setup_end; #define __exitcall(fn) static exitcall_t __exitcall_##fn __exit_call = fn -#define __init_call __attribute__ ((unused,__section__ (".initcall.init"))) +#define __init_call __attribute_used__ __attribute__ ((__section__ (".initcall.init"))) #endif -- cgit v1.2.3 From dc1561ac019ff7b6f75c5175abd2ec65c8dbd581 Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Fri, 24 Feb 2006 13:03:56 -0800 Subject: [PATCH] uml: os_connect_socket error path fixup Fix an fd leak and a return of -1 instead of -errno in the error path - this showed up in intensive testing of HPPFS, the os_connect_socket user. Signed-off-by: Paolo 'Blaisorblade' Giarrusso Acked-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/um/os-Linux/file.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) (limited to 'arch') diff --git a/arch/um/os-Linux/file.c b/arch/um/os-Linux/file.c index f55773c819e..3bd10deea28 100644 --- a/arch/um/os-Linux/file.c +++ b/arch/um/os-Linux/file.c @@ -272,14 +272,23 @@ int os_connect_socket(char *name) snprintf(sock.sun_path, sizeof(sock.sun_path), "%s", name); fd = socket(AF_UNIX, SOCK_STREAM, 0); - if(fd < 0) - return(fd); + if(fd < 0) { + err = -errno; + goto out; + } err = connect(fd, (struct sockaddr *) &sock, sizeof(sock)); - if(err) - return(-errno); + if(err) { + err = -errno; + goto out_close; + } - return(fd); + return fd; + +out_close: + close(fd); +out: + return err; } void os_close_file(int fd) -- cgit v1.2.3 From f462e8f913bdc7a28ce55508d0c045a0c445b157 Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Fri, 24 Feb 2006 13:03:57 -0800 Subject: [PATCH] uml: better error reporting for read_output Do precise error handling: print precise error messages, distinguishing short reads and read errors. This functions fails frequently enough for me so I bothered doing this fix. Signed-off-by: Paolo 'Blaisorblade' Giarrusso Acked-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/um/drivers/net_user.c | 34 ++++++++++++++++++++++------------ 1 file changed, 22 insertions(+), 12 deletions(-) (limited to 'arch') diff --git a/arch/um/drivers/net_user.c b/arch/um/drivers/net_user.c index 098fa65981a..0e2f06187ea 100644 --- a/arch/um/drivers/net_user.c +++ b/arch/um/drivers/net_user.c @@ -47,10 +47,12 @@ void tap_check_ips(char *gate_addr, unsigned char *eth_addr) } } +/* Do reliable error handling as this fails frequently enough. */ void read_output(int fd, char *output, int len) { - int remain, n, actual; + int remain, ret, expected; char c; + char *str; if(output == NULL){ output = &c; @@ -58,23 +60,31 @@ void read_output(int fd, char *output, int len) } *output = '\0'; - n = os_read_file(fd, &remain, sizeof(remain)); - if(n != sizeof(remain)){ - printk("read_output - read of length failed, err = %d\n", -n); - return; + ret = os_read_file(fd, &remain, sizeof(remain)); + + if (ret != sizeof(remain)) { + expected = sizeof(remain); + str = "length"; + goto err; } while(remain != 0){ - n = (remain < len) ? remain : len; - actual = os_read_file(fd, output, n); - if(actual != n){ - printk("read_output - read of data failed, " - "err = %d\n", -actual); - return; + expected = (remain < len) ? remain : len; + ret = os_read_file(fd, output, expected); + if (ret != expected) { + str = "data"; + goto err; } - remain -= actual; + remain -= ret; } + return; + +err: + if (ret < 0) + printk("read_output - read of %s failed, errno = %d\n", str, -ret); + else + printk("read_output - read of %s failed, read only %d of %d bytes\n", str, ret, expected); } int net_read(int fd, void *buf, int len) -- cgit v1.2.3 From fe1db50c7222c67466e41241bc7ef17b469bcf1d Mon Sep 17 00:00:00 2001 From: Paolo 'Blaisorblade' Giarrusso Date: Fri, 24 Feb 2006 13:03:58 -0800 Subject: [PATCH] uml: tidying COW code Improve (especially for coherence) some prototypes, and return code of init_cow_file in error case - for a short write return -EINVAL, otherwise return the error we got! Signed-off-by: Paolo 'Blaisorblade' Giarrusso Acked-by: Jeff Dike Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/um/drivers/cow.h | 2 +- arch/um/drivers/cow_sys.h | 4 ++-- arch/um/drivers/cow_user.c | 3 ++- arch/um/drivers/ubd_kern.c | 2 +- 4 files changed, 6 insertions(+), 5 deletions(-) (limited to 'arch') diff --git a/arch/um/drivers/cow.h b/arch/um/drivers/cow.h index dc36b222100..04e3958266e 100644 --- a/arch/um/drivers/cow.h +++ b/arch/um/drivers/cow.h @@ -46,7 +46,7 @@ extern int file_reader(__u64 offset, char *buf, int len, void *arg); extern int read_cow_header(int (*reader)(__u64, char *, int, void *), void *arg, __u32 *version_out, char **backing_file_out, time_t *mtime_out, - unsigned long long *size_out, int *sectorsize_out, + __u64 *size_out, int *sectorsize_out, __u32 *align_out, int *bitmap_offset_out); extern int write_cow_header(char *cow_file, int fd, char *backing_file, diff --git a/arch/um/drivers/cow_sys.h b/arch/um/drivers/cow_sys.h index df25263d2ad..94de4ead4f7 100644 --- a/arch/um/drivers/cow_sys.h +++ b/arch/um/drivers/cow_sys.h @@ -23,12 +23,12 @@ static inline char *cow_strdup(char *str) return(uml_strdup(str)); } -static inline int cow_seek_file(int fd, unsigned long long offset) +static inline int cow_seek_file(int fd, __u64 offset) { return(os_seek_file(fd, offset)); } -static inline int cow_file_size(char *file, unsigned long long *size_out) +static inline int cow_file_size(char *file, __u64 *size_out) { return(os_file_size(file, size_out)); } diff --git a/arch/um/drivers/cow_user.c b/arch/um/drivers/cow_user.c index d1c86bc377b..61951b72126 100644 --- a/arch/um/drivers/cow_user.c +++ b/arch/um/drivers/cow_user.c @@ -362,7 +362,8 @@ int init_cow_file(int fd, char *cow_file, char *backing_file, int sectorsize, if(err != sizeof(zero)){ cow_printf("Write of bitmap to new COW file '%s' failed, " "err = %d\n", cow_file, -err); - err = -EINVAL; + if (err >= 0) + err = -EINVAL; goto out; } diff --git a/arch/um/drivers/ubd_kern.c b/arch/um/drivers/ubd_kern.c index 101efd26d46..fa617e0719a 100644 --- a/arch/um/drivers/ubd_kern.c +++ b/arch/um/drivers/ubd_kern.c @@ -1135,7 +1135,7 @@ static int path_requires_switch(char *from_cmdline, char *from_cow, char *cow) static int backing_file_mismatch(char *file, __u64 size, time_t mtime) { unsigned long modtime; - long long actual; + unsigned long long actual; int err; err = os_file_modtime(file, &modtime); -- cgit v1.2.3 From 124d90be62343f71bbb7a6b4a907b5584181e6d5 Mon Sep 17 00:00:00 2001 From: Prasanna S Panchamukhi Date: Fri, 24 Feb 2006 13:04:08 -0800 Subject: [PATCH] Kprobes causes NX protection fault on i686 SMP Fix a problem seen on i686 machine with NX support where the instruction could not be single stepped because of NX bit set on the memory pages allocated by kprobes module. This patch provides allocation of instruction solt so that the processor can execute the instruction from that location similar to x86_64 architecture. Thanks to Bibo and Masami for testing this patch. Signed-off-by: Prasanna S Panchamukhi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/kprobes.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/i386/kernel/kprobes.c b/arch/i386/kernel/kprobes.c index 6483eeb1a4e..694a1399763 100644 --- a/arch/i386/kernel/kprobes.c +++ b/arch/i386/kernel/kprobes.c @@ -58,6 +58,11 @@ static inline int is_IF_modifier(kprobe_opcode_t opcode) int __kprobes arch_prepare_kprobe(struct kprobe *p) { + /* insn: must be on special executable page on i386. */ + p->ainsn.insn = get_insn_slot(); + if (!p->ainsn.insn) + return -ENOMEM; + memcpy(p->ainsn.insn, p->addr, MAX_INSN_SIZE * sizeof(kprobe_opcode_t)); p->opcode = *p->addr; return 0; @@ -77,6 +82,13 @@ void __kprobes arch_disarm_kprobe(struct kprobe *p) (unsigned long) p->addr + sizeof(kprobe_opcode_t)); } +void __kprobes arch_remove_kprobe(struct kprobe *p) +{ + down(&kprobe_mutex); + free_insn_slot(p->ainsn.insn); + up(&kprobe_mutex); +} + static inline void save_previous_kprobe(struct kprobe_ctlblk *kcb) { kcb->prev_kprobe.kp = kprobe_running(); @@ -111,7 +123,7 @@ static inline void prepare_singlestep(struct kprobe *p, struct pt_regs *regs) if (p->opcode == BREAKPOINT_INSTRUCTION) regs->eip = (unsigned long)p->addr; else - regs->eip = (unsigned long)&p->ainsn.insn; + regs->eip = (unsigned long)p->ainsn.insn; } /* Called with kretprobe_lock held */ @@ -351,7 +363,7 @@ static void __kprobes resume_execution(struct kprobe *p, { unsigned long *tos = (unsigned long *)®s->esp; unsigned long next_eip = 0; - unsigned long copy_eip = (unsigned long)&p->ainsn.insn; + unsigned long copy_eip = (unsigned long)p->ainsn.insn; unsigned long orig_eip = (unsigned long)p->addr; switch (p->ainsn.insn[0]) { -- cgit v1.2.3 From 8d5c822b2920be9016806f61fd552d2301cfa2fc Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Fri, 24 Feb 2006 13:04:10 -0800 Subject: [PATCH] voyager: fix boot panic by adding topology export It looks like I can't get away without exporting topology functions from voyager any longer, so add them to the voyager subarchitecture. Signed-off-by: James Bottomley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/mach-voyager/voyager_basic.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) (limited to 'arch') diff --git a/arch/i386/mach-voyager/voyager_basic.c b/arch/i386/mach-voyager/voyager_basic.c index aa49a33a572..6761d294f26 100644 --- a/arch/i386/mach-voyager/voyager_basic.c +++ b/arch/i386/mach-voyager/voyager_basic.c @@ -23,6 +23,9 @@ #include #include #include +#include +#include +#include #include #include #include @@ -329,3 +332,15 @@ void machine_power_off(void) pm_power_off(); } +static struct i386_cpu cpu_devices[NR_CPUS]; + +static int __init topology_init(void) +{ + int i; + + for_each_present_cpu(i) + register_cpu(&cpu_devices[i].cpu, i, NULL); + return 0; +} + +subsys_initcall(topology_init); -- cgit v1.2.3 From f68a106f224c21148c5264a429fac149dc7ad0ac Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Fri, 24 Feb 2006 13:04:11 -0800 Subject: [PATCH] voyager: fix the cpu_possible_map to make voyager boot again Right at the moment (thanks to a patch from Andrew), cpu_possible_map on voyager is CPU_MASK_NONE, which means the machine always thinks it has no CPUs. Fix that by doing an early initialisation of the cpu_possible_map from the cpu_phys_present_map. (akpm: we aim to please) Signed-off-by: James Bottomley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/mach-voyager/voyager_smp.c | 1 + 1 file changed, 1 insertion(+) (limited to 'arch') diff --git a/arch/i386/mach-voyager/voyager_smp.c b/arch/i386/mach-voyager/voyager_smp.c index 6e4c3baef6c..8165626a5c3 100644 --- a/arch/i386/mach-voyager/voyager_smp.c +++ b/arch/i386/mach-voyager/voyager_smp.c @@ -402,6 +402,7 @@ find_smp_config(void) cpus_addr(phys_cpu_present_map)[0] |= voyager_extended_cmos_read(VOYAGER_PROCESSOR_PRESENT_MASK + 1) << 8; cpus_addr(phys_cpu_present_map)[0] |= voyager_extended_cmos_read(VOYAGER_PROCESSOR_PRESENT_MASK + 2) << 16; cpus_addr(phys_cpu_present_map)[0] |= voyager_extended_cmos_read(VOYAGER_PROCESSOR_PRESENT_MASK + 3) << 24; + cpu_possible_map = phys_cpu_present_map; printk("VOYAGER SMP: phys_cpu_present_map = 0x%lx\n", cpus_addr(phys_cpu_present_map)[0]); /* Here we set up the VIC to enable SMP */ /* enable the CPIs by writing the base vector to their register */ -- cgit v1.2.3 From 2b932f6cf052920fb3a6281499e08209b08f5086 Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Fri, 24 Feb 2006 13:04:14 -0800 Subject: [PATCH] x86: fix broken SMP boot sequence Recent GDT changes broke the SMP boot sequence if the booting CPU is numbered anything other than zero. There's also a subtle source of error in that the boot time CPU now uses cpu_gdt_table (which is actually the GDT for booting CPUs in head.S). This patch fixes both problems by making GDT descriptors themselves allocated from a per_cpu area and switching to them in cpu_init(), which now means that cpu_gdt_table is exclusively used for booting CPUs again. Signed-off-by: James Bottomley Cc: Zachary Amsden Cc: Matt Tolentino Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/cpu/common.c | 32 ++++++++++++++++++++++++++++---- arch/i386/kernel/efi.c | 12 +++++++----- arch/i386/kernel/head.S | 2 -- arch/i386/kernel/i386_ksyms.c | 2 -- arch/i386/kernel/smpboot.c | 6 ------ 5 files changed, 35 insertions(+), 19 deletions(-) (limited to 'arch') diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c index 7eb9213734a..4ecd4b326de 100644 --- a/arch/i386/kernel/cpu/common.c +++ b/arch/i386/kernel/cpu/common.c @@ -4,6 +4,7 @@ #include #include #include +#include #include #include #include @@ -18,6 +19,9 @@ #include "cpu.h" +DEFINE_PER_CPU(struct Xgt_desc_struct, cpu_gdt_descr); +EXPORT_PER_CPU_SYMBOL(cpu_gdt_descr); + DEFINE_PER_CPU(unsigned char, cpu_16bit_stack[CPU_16BIT_STACK_SIZE]); EXPORT_PER_CPU_SYMBOL(cpu_16bit_stack); @@ -571,8 +575,9 @@ void __devinit cpu_init(void) int cpu = smp_processor_id(); struct tss_struct * t = &per_cpu(init_tss, cpu); struct thread_struct *thread = ¤t->thread; - struct desc_struct *gdt = get_cpu_gdt_table(cpu); + struct desc_struct *gdt; __u32 stk16_off = (__u32)&per_cpu(cpu_16bit_stack, cpu); + struct Xgt_desc_struct *cpu_gdt_descr = &per_cpu(cpu_gdt_descr, cpu); if (cpu_test_and_set(cpu, cpu_initialized)) { printk(KERN_WARNING "CPU#%d already initialized!\n", cpu); @@ -589,6 +594,25 @@ void __devinit cpu_init(void) set_in_cr4(X86_CR4_TSD); } + /* + * This is a horrible hack to allocate the GDT. The problem + * is that cpu_init() is called really early for the boot CPU + * (and hence needs bootmem) but much later for the secondary + * CPUs, when bootmem will have gone away + */ + if (NODE_DATA(0)->bdata->node_bootmem_map) { + gdt = (struct desc_struct *)alloc_bootmem_pages(PAGE_SIZE); + /* alloc_bootmem_pages panics on failure, so no check */ + memset(gdt, 0, PAGE_SIZE); + } else { + gdt = (struct desc_struct *)get_zeroed_page(GFP_KERNEL); + if (unlikely(!gdt)) { + printk(KERN_CRIT "CPU%d failed to allocate GDT\n", cpu); + for (;;) + local_irq_enable(); + } + } + /* * Initialize the per-CPU GDT with the boot GDT, * and set up the GDT descriptor: @@ -601,10 +625,10 @@ void __devinit cpu_init(void) ((((__u64)stk16_off) << 32) & 0xff00000000000000ULL) | (CPU_16BIT_STACK_SIZE - 1); - cpu_gdt_descr[cpu].size = GDT_SIZE - 1; - cpu_gdt_descr[cpu].address = (unsigned long)gdt; + cpu_gdt_descr->size = GDT_SIZE - 1; + cpu_gdt_descr->address = (unsigned long)gdt; - load_gdt(&cpu_gdt_descr[cpu]); + load_gdt(cpu_gdt_descr); load_idt(&idt_descr); /* diff --git a/arch/i386/kernel/efi.c b/arch/i386/kernel/efi.c index ecad519fd39..e3e42fd6240 100644 --- a/arch/i386/kernel/efi.c +++ b/arch/i386/kernel/efi.c @@ -103,17 +103,19 @@ static void efi_call_phys_prelog(void) */ local_flush_tlb(); - cpu_gdt_descr[0].address = __pa(cpu_gdt_descr[0].address); - load_gdt((struct Xgt_desc_struct *) __pa(&cpu_gdt_descr[0])); + per_cpu(cpu_gdt_descr, 0).address = + __pa(per_cpu(cpu_gdt_descr, 0).address); + load_gdt((struct Xgt_desc_struct *)__pa(&per_cpu(cpu_gdt_descr, 0))); } static void efi_call_phys_epilog(void) { unsigned long cr4; - cpu_gdt_descr[0].address = - (unsigned long) __va(cpu_gdt_descr[0].address); - load_gdt(&cpu_gdt_descr[0]); + per_cpu(cpu_gdt_descr, 0).address = + (unsigned long)__va(per_cpu(cpu_gdt_descr, 0).address); + load_gdt((struct Xgt_desc_struct *)__va(&per_cpu(cpu_gdt_descr, 0))); + cr4 = read_cr4(); if (cr4 & X86_CR4_PSE) { diff --git a/arch/i386/kernel/head.S b/arch/i386/kernel/head.S index 2bee6499edd..e0b7c632efb 100644 --- a/arch/i386/kernel/head.S +++ b/arch/i386/kernel/head.S @@ -534,5 +534,3 @@ ENTRY(cpu_gdt_table) .quad 0x0000000000000000 /* 0xf0 - unused */ .quad 0x0000000000000000 /* 0xf8 - GDT entry 31: double-fault TSS */ - /* Be sure this is zeroed to avoid false validations in Xen */ - .fill PAGE_SIZE_asm / 8 - GDT_ENTRIES,8,0 diff --git a/arch/i386/kernel/i386_ksyms.c b/arch/i386/kernel/i386_ksyms.c index 3999bec50c3..055325056a7 100644 --- a/arch/i386/kernel/i386_ksyms.c +++ b/arch/i386/kernel/i386_ksyms.c @@ -3,8 +3,6 @@ #include #include -EXPORT_SYMBOL_GPL(cpu_gdt_descr); - EXPORT_SYMBOL(__down_failed); EXPORT_SYMBOL(__down_failed_interruptible); EXPORT_SYMBOL(__down_failed_trylock); diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c index fb00ab7b761..eba7f53f8b4 100644 --- a/arch/i386/kernel/smpboot.c +++ b/arch/i386/kernel/smpboot.c @@ -898,12 +898,6 @@ static int __devinit do_boot_cpu(int apicid, int cpu) unsigned long start_eip; unsigned short nmi_high = 0, nmi_low = 0; - if (!cpu_gdt_descr[cpu].address && - !(cpu_gdt_descr[cpu].address = get_zeroed_page(GFP_KERNEL))) { - printk("Failed to allocate GDT for CPU %d\n", cpu); - return 1; - } - ++cpucount; /* -- cgit v1.2.3 From c04030e16dbea2f7581f82cc6688695927f6ac5b Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Fri, 24 Feb 2006 13:04:21 -0800 Subject: [PATCH] flags parameter for linkat I'm currently at the POSIX meeting and one thing covered was the incompatibility of Linux's link() with the POSIX definition. The name. Linux does not follow symlinks, POSIX requires it does. Even if somebody thinks this is a good default behavior we cannot change this because it would break the ABI. But the fact remains that some application might want this behavior. We have one chance to help implementing this without breaking the behavior. For this we could use the new linkat interface which would need a new flags parameter. If the new parameter is AT_SYMLINK_FOLLOW the new behavior could be invoked. I do not want to introduce such a patch now. But we could add the parameter now, just don't use it. The patch below would do this. Can we get this late patch applied before the release more or less fixes the syscall API? Signed-off-by: Ulrich Drepper Signed-off-by: Ralf Baechle Cc: Heiko Carstens Cc: Martin Schwidefsky Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/mips/kernel/scall32-o32.S | 2 +- arch/s390/kernel/compat_wrapper.S | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall32-o32.S index d83e033dbc8..2f2dc54b2e2 100644 --- a/arch/mips/kernel/scall32-o32.S +++ b/arch/mips/kernel/scall32-o32.S @@ -626,7 +626,7 @@ einval: li v0, -EINVAL sys sys_fstatat64 4 sys sys_unlinkat 3 sys sys_renameat 4 /* 4295 */ - sys sys_linkat 4 + sys sys_linkat 5 sys sys_symlinkat 3 sys sys_readlinkat 4 sys sys_fchmodat 3 diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S index 615964cca15..50e80138e7a 100644 --- a/arch/s390/kernel/compat_wrapper.S +++ b/arch/s390/kernel/compat_wrapper.S @@ -1552,6 +1552,7 @@ sys_linkat_wrapper: llgtr %r3,%r3 # const char * lgfr %r4,%r4 # int llgtr %r5,%r5 # const char * + lgfr %r6,%r6 # int jg sys_linkat .globl sys_symlinkat_wrapper -- cgit v1.2.3 From 9c869edac591977314323a4eaad5f7633fca684f Mon Sep 17 00:00:00 2001 From: Zachary Amsden Date: Fri, 24 Feb 2006 13:04:27 -0800 Subject: [PATCH] Fix topology.c location When compiling a non-default subarch, topology.c is missing from the kernel build. This causes builds with CONFIG_HOTPLUG_CPU to fail. In addition, on Intel processors with cpuid level > 4, it causes intel_cacheinfo.c to reference uninitialized data that should have been set up by the initcall in topology.c which calls register_cpu. This causes a kernel panic on boot on newer Intel processors. Moving topology.c to arch/i386/kernel fixes both of these problems. Thanks to Dan Hecht for finding and fixing this problem. Signed-off-by: Zachary Amsden Signed-off-by: Dan Hecht Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/i386/kernel/Makefile | 2 +- arch/i386/kernel/topology.c | 97 +++++++++++++++++++++++++++++++++++++++ arch/i386/mach-default/Makefile | 2 +- arch/i386/mach-default/topology.c | 97 --------------------------------------- 4 files changed, 99 insertions(+), 99 deletions(-) create mode 100644 arch/i386/kernel/topology.c delete mode 100644 arch/i386/mach-default/topology.c (limited to 'arch') diff --git a/arch/i386/kernel/Makefile b/arch/i386/kernel/Makefile index 60c3f76dfca..53bb9a79e27 100644 --- a/arch/i386/kernel/Makefile +++ b/arch/i386/kernel/Makefile @@ -7,7 +7,7 @@ extra-y := head.o init_task.o vmlinux.lds obj-y := process.o semaphore.o signal.o entry.o traps.o irq.o \ ptrace.o time.o ioport.o ldt.o setup.o i8259.o sys_i386.o \ pci-dma.o i386_ksyms.o i387.o dmi_scan.o bootflag.o \ - quirks.o i8237.o + quirks.o i8237.o topology.o obj-y += cpu/ obj-y += timers/ diff --git a/arch/i386/kernel/topology.c b/arch/i386/kernel/topology.c new file mode 100644 index 00000000000..67a0e1baa28 --- /dev/null +++ b/arch/i386/kernel/topology.c @@ -0,0 +1,97 @@ +/* + * arch/i386/kernel/topology.c - Populate driverfs with topology information + * + * Written by: Matthew Dobson, IBM Corporation + * Original Code: Paul Dorwin, IBM Corporation, Patrick Mochel, OSDL + * + * Copyright (C) 2002, IBM Corp. + * + * All rights reserved. + * + * 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. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + * + * Send feedback to + */ +#include +#include +#include +#include + +static struct i386_cpu cpu_devices[NR_CPUS]; + +int arch_register_cpu(int num){ + struct node *parent = NULL; + +#ifdef CONFIG_NUMA + int node = cpu_to_node(num); + if (node_online(node)) + parent = &node_devices[node].node; +#endif /* CONFIG_NUMA */ + + return register_cpu(&cpu_devices[num].cpu, num, parent); +} + +#ifdef CONFIG_HOTPLUG_CPU + +void arch_unregister_cpu(int num) { + struct node *parent = NULL; + +#ifdef CONFIG_NUMA + int node = cpu_to_node(num); + if (node_online(node)) + parent = &node_devices[node].node; +#endif /* CONFIG_NUMA */ + + return unregister_cpu(&cpu_devices[num].cpu, parent); +} +EXPORT_SYMBOL(arch_register_cpu); +EXPORT_SYMBOL(arch_unregister_cpu); +#endif /*CONFIG_HOTPLUG_CPU*/ + + + +#ifdef CONFIG_NUMA +#include +#include + +struct i386_node node_devices[MAX_NUMNODES]; + +static int __init topology_init(void) +{ + int i; + + for_each_online_node(i) + arch_register_node(i); + + for_each_present_cpu(i) + arch_register_cpu(i); + return 0; +} + +#else /* !CONFIG_NUMA */ + +static int __init topology_init(void) +{ + int i; + + for_each_present_cpu(i) + arch_register_cpu(i); + return 0; +} + +#endif /* CONFIG_NUMA */ + +subsys_initcall(topology_init); diff --git a/arch/i386/mach-default/Makefile b/arch/i386/mach-default/Makefile index e95bb023792..012fe34459e 100644 --- a/arch/i386/mach-default/Makefile +++ b/arch/i386/mach-default/Makefile @@ -2,4 +2,4 @@ # Makefile for the linux kernel. # -obj-y := setup.o topology.o +obj-y := setup.o diff --git a/arch/i386/mach-default/topology.c b/arch/i386/mach-default/topology.c deleted file mode 100644 index b64314069e7..00000000000 --- a/arch/i386/mach-default/topology.c +++ /dev/null @@ -1,97 +0,0 @@ -/* - * arch/i386/mach-generic/topology.c - Populate driverfs with topology information - * - * Written by: Matthew Dobson, IBM Corporation - * Original Code: Paul Dorwin, IBM Corporation, Patrick Mochel, OSDL - * - * Copyright (C) 2002, IBM Corp. - * - * All rights reserved. - * - * 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. - * - * This program is distributed in the hope that it will be useful, but - * WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or - * NON INFRINGEMENT. See the GNU General Public License for more - * details. - * - * You should have received a copy of the GNU General Public License - * along with this program; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. - * - * Send feedback to - */ -#include -#include -#include -#include - -static struct i386_cpu cpu_devices[NR_CPUS]; - -int arch_register_cpu(int num){ - struct node *parent = NULL; - -#ifdef CONFIG_NUMA - int node = cpu_to_node(num); - if (node_online(node)) - parent = &node_devices[node].node; -#endif /* CONFIG_NUMA */ - - return register_cpu(&cpu_devices[num].cpu, num, parent); -} - -#ifdef CONFIG_HOTPLUG_CPU - -void arch_unregister_cpu(int num) { - struct node *parent = NULL; - -#ifdef CONFIG_NUMA - int node = cpu_to_node(num); - if (node_online(node)) - parent = &node_devices[node].node; -#endif /* CONFIG_NUMA */ - - return unregister_cpu(&cpu_devices[num].cpu, parent); -} -EXPORT_SYMBOL(arch_register_cpu); -EXPORT_SYMBOL(arch_unregister_cpu); -#endif /*CONFIG_HOTPLUG_CPU*/ - - - -#ifdef CONFIG_NUMA -#include -#include - -struct i386_node node_devices[MAX_NUMNODES]; - -static int __init topology_init(void) -{ - int i; - - for_each_online_node(i) - arch_register_node(i); - - for_each_present_cpu(i) - arch_register_cpu(i); - return 0; -} - -#else /* !CONFIG_NUMA */ - -static int __init topology_init(void) -{ - int i; - - for_each_present_cpu(i) - arch_register_cpu(i); - return 0; -} - -#endif /* CONFIG_NUMA */ - -subsys_initcall(topology_init); -- cgit v1.2.3 From 329dda083e496bc5ffbb4b1973243bd8a9420e24 Mon Sep 17 00:00:00 2001 From: Kumar Gala Date: Fri, 24 Feb 2006 10:54:52 -0600 Subject: [PATCH] powerpc: Fix mem= cmdline handling on arch/powerpc for !MULTIPLATFORM mem= command line option was being ignored in arch/powerpc if we were not a CONFIG_MULTIPLATFORM (which is handled via prom_init stub). The initial command line extraction and parsing needed to be moved earlier in the boot process and have code to actual parse mem= and do something about it. Also, fixed a compile warning in the file. Signed-off-by: Kumar Gala Acked-by: Segher Boessenkool Signed-off-by: Linus Torvalds --- arch/powerpc/kernel/prom.c | 54 ++++++++++++++++++++++++++++++++-------------- 1 file changed, 38 insertions(+), 16 deletions(-) (limited to 'arch') diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 294832a7e0a..6dbd2172677 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c @@ -816,8 +816,6 @@ void __init unflatten_device_tree(void) { unsigned long start, mem, size; struct device_node **allnextp = &allnodes; - char *p = NULL; - int l = 0; DBG(" -> unflatten_device_tree()\n"); @@ -857,19 +855,6 @@ void __init unflatten_device_tree(void) if (of_chosen == NULL) of_chosen = of_find_node_by_path("/chosen@0"); - /* Retreive command line */ - if (of_chosen != NULL) { - p = (char *)get_property(of_chosen, "bootargs", &l); - if (p != NULL && l > 0) - strlcpy(cmd_line, p, min(l, COMMAND_LINE_SIZE)); - } -#ifdef CONFIG_CMDLINE - if (l == 0 || (l == 1 && (*p) == 0)) - strlcpy(cmd_line, CONFIG_CMDLINE, COMMAND_LINE_SIZE); -#endif /* CONFIG_CMDLINE */ - - DBG("Command line is: %s\n", cmd_line); - DBG(" <- unflatten_device_tree()\n"); } @@ -940,6 +925,8 @@ static int __init early_init_dt_scan_chosen(unsigned long node, { u32 *prop; unsigned long *lprop; + unsigned long l; + char *p; DBG("search \"chosen\", depth: %d, uname: %s\n", depth, uname); @@ -1004,6 +991,41 @@ static int __init early_init_dt_scan_chosen(unsigned long node, crashk_res.end = crashk_res.start + *lprop - 1; #endif + /* Retreive command line */ + p = of_get_flat_dt_prop(node, "bootargs", &l); + if (p != NULL && l > 0) + strlcpy(cmd_line, p, min((int)l, COMMAND_LINE_SIZE)); + +#ifdef CONFIG_CMDLINE + if (l == 0 || (l == 1 && (*p) == 0)) + strlcpy(cmd_line, CONFIG_CMDLINE, COMMAND_LINE_SIZE); +#endif /* CONFIG_CMDLINE */ + + DBG("Command line is: %s\n", cmd_line); + + if (strstr(cmd_line, "mem=")) { + char *p, *q; + unsigned long maxmem = 0; + + for (q = cmd_line; (p = strstr(q, "mem=")) != 0; ) { + q = p + 4; + if (p > cmd_line && p[-1] != ' ') + continue; + maxmem = simple_strtoul(q, &q, 0); + if (*q == 'k' || *q == 'K') { + maxmem <<= 10; + ++q; + } else if (*q == 'm' || *q == 'M') { + maxmem <<= 20; + ++q; + } else if (*q == 'g' || *q == 'G') { + maxmem <<= 30; + ++q; + } + } + memory_limit = maxmem; + } + /* break now */ return 1; } @@ -1124,7 +1146,7 @@ static void __init early_reserve_mem(void) size_32 = *(reserve_map_32++); if (size_32 == 0) break; - DBG("reserving: %lx -> %lx\n", base_32, size_32); + DBG("reserving: %x -> %x\n", base_32, size_32); lmb_reserve(base_32, size_32); } return; -- cgit v1.2.3 From a0124d780d06db711e8a92135d774940588a27da Mon Sep 17 00:00:00 2001 From: Dave Jones Date: Sat, 25 Feb 2006 03:55:38 -0500 Subject: [PATCH] x86-64: react to new topology.c location Commit 9c869edac591977314323a4eaad5f7633fca684f moved the i386 topology.c file. That change broke x86-64 compiles, as it uses the same file. Signed-off-by: Dave Jones Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86_64/kernel/Makefile b/arch/x86_64/kernel/Makefile index 72fe60c20d3..a098a11e775 100644 --- a/arch/x86_64/kernel/Makefile +++ b/arch/x86_64/kernel/Makefile @@ -43,7 +43,7 @@ CFLAGS_vsyscall.o := $(PROFILING) -g0 bootflag-y += ../../i386/kernel/bootflag.o cpuid-$(subst m,y,$(CONFIG_X86_CPUID)) += ../../i386/kernel/cpuid.o -topology-y += ../../i386/mach-default/topology.o +topology-y += ../../i386/kernel/topology.o microcode-$(subst m,y,$(CONFIG_MICROCODE)) += ../../i386/kernel/microcode.o intel_cacheinfo-y += ../../i386/kernel/cpu/intel_cacheinfo.o quirks-y += ../../i386/kernel/quirks.o -- cgit v1.2.3 From 60b08c67220cf6faef7410ac6adba23a8a743bf7 Mon Sep 17 00:00:00 2001 From: Jon Mason Date: Sun, 26 Feb 2006 04:18:22 +0100 Subject: [PATCH] x86_64: no_iommu removal in pci-gart.c In previous versions of pci-gart.c, no_iommu was used to determine if IOMMU was disabled in the GART DMA mapping functions. This changed in 2.6.16 and now gart_xxx() functions are only called if gart is enabled. Therefore, uses of no_iommu in the GART code are no longer necessary and can be removed. Also, it removes double deceleration of no_iommu and force_iommu in pci.h and proto.h, by removing the deceleration in pci.h. Lastly, end_pfn off by one error. Tested (along with patch 1/2) on dual opteron with gart enabled, iommu=soft, and iommu=off. Signed-off-by: Jon Mason Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/aperture.c | 2 +- arch/x86_64/kernel/pci-gart.c | 22 +++++----------------- 2 files changed, 6 insertions(+), 18 deletions(-) (limited to 'arch') diff --git a/arch/x86_64/kernel/aperture.c b/arch/x86_64/kernel/aperture.c index e4e2b7d01f8..a0f955b9995 100644 --- a/arch/x86_64/kernel/aperture.c +++ b/arch/x86_64/kernel/aperture.c @@ -248,7 +248,7 @@ void __init iommu_hole_init(void) /* Got the aperture from the AGP bridge */ } else if (swiotlb && !valid_agp) { /* Do nothing */ - } else if ((!no_iommu && end_pfn >= MAX_DMA32_PFN) || + } else if ((!no_iommu && end_pfn > MAX_DMA32_PFN) || force_iommu || valid_agp || fallback_aper_force) { diff --git a/arch/x86_64/kernel/pci-gart.c b/arch/x86_64/kernel/pci-gart.c index dd0718dc178..0c3f052ba6c 100644 --- a/arch/x86_64/kernel/pci-gart.c +++ b/arch/x86_64/kernel/pci-gart.c @@ -228,11 +228,6 @@ static inline int need_iommu(struct device *dev, unsigned long addr, size_t size int mmu = high; if (force_iommu) mmu = 1; - if (no_iommu) { - if (high) - panic("PCI-DMA: high address but no IOMMU.\n"); - mmu = 0; - } return mmu; } @@ -241,11 +236,6 @@ static inline int nonforced_iommu(struct device *dev, unsigned long addr, size_t u64 mask = *dev->dma_mask; int high = addr + size >= mask; int mmu = high; - if (no_iommu) { - if (high) - panic("PCI-DMA: high address but no IOMMU.\n"); - mmu = 0; - } return mmu; } @@ -310,7 +300,7 @@ void gart_unmap_sg(struct device *dev, struct scatterlist *sg, int nents, int di for (i = 0; i < nents; i++) { struct scatterlist *s = &sg[i]; - if (!s->dma_length) + if (!s->dma_length || !s->length) break; dma_unmap_single(dev, s->dma_address, s->dma_length, dir); } @@ -364,6 +354,7 @@ static int __dma_map_cont(struct scatterlist *sg, int start, int stopat, BUG_ON(i > start && s->offset); if (i == start) { + *sout = *s; sout->dma_address = iommu_bus_base; sout->dma_address += iommu_page*PAGE_SIZE + s->offset; sout->dma_length = s->length; @@ -390,6 +381,7 @@ static inline int dma_map_cont(struct scatterlist *sg, int start, int stopat, { if (!need) { BUG_ON(stopat - start != 1); + *sout = sg[start]; sout->dma_length = sg[start].length; return 0; } @@ -632,17 +624,13 @@ static int __init pci_iommu_init(void) (agp_copy_info(agp_bridge, &info) < 0); #endif - if (swiotlb) { - no_iommu = 1; + if (swiotlb) return -1; - } - + if (no_iommu || (!force_iommu && end_pfn <= MAX_DMA32_PFN) || !iommu_aperture || (no_agp && init_k8_gatt(&info) < 0)) { - no_iommu = 1; - no_iommu_init(); printk(KERN_INFO "PCI-DMA: Disabling IOMMU.\n"); if (end_pfn > MAX_DMA32_PFN) { printk(KERN_ERR "WARNING more than 4GB of memory " -- cgit v1.2.3 From 13a229abc25640813f1480c0478dfc6bdbc1c19e Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:31 +0100 Subject: [PATCH] x86_64: Only do the clustered systems have unsynchronized TSC assumption on IBM systems Big Unisys systems have multiple clusters too, but they have an synchronized TSC. I'm using the SMBIOS to check for vendor == IBM. Cc: Chris McDermott Cc: "Protasevich, Natalie" Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/i386/kernel/acpi/boot.c | 3 --- arch/x86_64/kernel/apic.c | 9 ++++++++- 2 files changed, 8 insertions(+), 4 deletions(-) (limited to 'arch') diff --git a/arch/i386/kernel/acpi/boot.c b/arch/i386/kernel/acpi/boot.c index 79577f0ace9..8309a7b2cd6 100644 --- a/arch/i386/kernel/acpi/boot.c +++ b/arch/i386/kernel/acpi/boot.c @@ -44,9 +44,6 @@ extern void __init clustered_apic_check(void); extern int gsi_irq_sharing(int gsi); #include -static inline int acpi_madt_oem_check(char *oem_id, char *oem_table_id) { return 0; } - - #else /* X86 */ #ifdef CONFIG_X86_LOCAL_APIC diff --git a/arch/x86_64/kernel/apic.c b/arch/x86_64/kernel/apic.c index e5b14c57eaa..d70605eda33 100644 --- a/arch/x86_64/kernel/apic.c +++ b/arch/x86_64/kernel/apic.c @@ -962,12 +962,14 @@ void smp_apic_timer_interrupt(struct pt_regs *regs) irq_exit(); } +int __initdata unsync_tsc_on_multicluster; + /* * oem_force_hpet_timer -- force HPET mode for some boxes. * * Thus far, the major user of this is IBM's Summit2 series: * - * Clustered boxes may have unsynced TSC problems if they are + * Some clustered boxes may have unsynced TSC problems if they are * multi-chassis. Use available data to take a good guess. * If in doubt, go HPET. */ @@ -977,6 +979,11 @@ __cpuinit int oem_force_hpet_timer(void) unsigned id; DECLARE_BITMAP(clustermap, NUM_APIC_CLUSTERS); + /* Only do this check on IBM machines - big Unisys systems + use multiple clusters too, but have synchronized TSC */ + if (!unsync_tsc_on_multicluster) + return 0; + bitmap_zero(clustermap, NUM_APIC_CLUSTERS); for (i = 0; i < NR_CPUS; i++) { -- cgit v1.2.3 From 6070f9ec6b03cc46cd0242523326f7a296f47c1c Mon Sep 17 00:00:00 2001 From: Andreas Deresch Date: Sun, 26 Feb 2006 04:18:34 +0100 Subject: [PATCH] i386: Handle non existing APICs without panicing [description from AK] This fixes booting in APIC mode on some ACER laptops. x86-64 did a similar change some time ago. See http://bugzilla.kernel.org/show_bug.cgi?id=4700 for details Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/i386/kernel/io_apic.c | 6 ++++-- arch/i386/kernel/mpparse.c | 10 ++++++++-- 2 files changed, 12 insertions(+), 4 deletions(-) (limited to 'arch') diff --git a/arch/i386/kernel/io_apic.c b/arch/i386/kernel/io_apic.c index f2dd218d88c..235822b3f41 100644 --- a/arch/i386/kernel/io_apic.c +++ b/arch/i386/kernel/io_apic.c @@ -2566,8 +2566,10 @@ int __init io_apic_get_unique_id (int ioapic, int apic_id) spin_unlock_irqrestore(&ioapic_lock, flags); /* Sanity check */ - if (reg_00.bits.ID != apic_id) - panic("IOAPIC[%d]: Unable change apic_id!\n", ioapic); + if (reg_00.bits.ID != apic_id) { + printk("IOAPIC[%d]: Unable to change apic_id!\n", ioapic); + return -1; + } } apic_printk(APIC_VERBOSE, KERN_INFO diff --git a/arch/i386/kernel/mpparse.c b/arch/i386/kernel/mpparse.c index e7609abf379..e6e2f43db85 100644 --- a/arch/i386/kernel/mpparse.c +++ b/arch/i386/kernel/mpparse.c @@ -915,6 +915,7 @@ void __init mp_register_ioapic ( u32 gsi_base) { int idx = 0; + int tmpid; if (nr_ioapics >= MAX_IO_APICS) { printk(KERN_ERR "ERROR: Max # of I/O APICs (%d) exceeded " @@ -935,9 +936,14 @@ void __init mp_register_ioapic ( set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address); if ((boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) && (boot_cpu_data.x86 < 15)) - mp_ioapics[idx].mpc_apicid = io_apic_get_unique_id(idx, id); + tmpid = io_apic_get_unique_id(idx, id); else - mp_ioapics[idx].mpc_apicid = id; + tmpid = id; + if (tmpid == -1) { + nr_ioapics--; + return; + } + mp_ioapics[idx].mpc_apicid = tmpid; mp_ioapics[idx].mpc_apicver = io_apic_get_version(idx); /* -- cgit v1.2.3 From e78256b8f3e2850ad55c2d69e1429e6c2607afd3 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:37 +0100 Subject: [PATCH] x86-64/i386: Use common X86_PM_TIMER option and make it EMBEDDED This makes x86-64 use the common X86_PM_TIMER Kconfig entry in drivers/acpi And since PM timer is needed for correct timing on a lot of systems now (e.g. AMD dual cores) and we often get bug reports from people who forgot to set it make it depend on CONFIG_EMBEDDED. x86-64 had this change before and it's a good thing. I also fixed the description slightly to make this more clear. Cc: len.brown@intel.com Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/Kconfig | 15 --------------- 1 file changed, 15 deletions(-) (limited to 'arch') diff --git a/arch/x86_64/Kconfig b/arch/x86_64/Kconfig index 2f9deca31cc..babc31b3ef1 100644 --- a/arch/x86_64/Kconfig +++ b/arch/x86_64/Kconfig @@ -354,21 +354,6 @@ config HPET_TIMER as it is off-chip. You can find the HPET spec at . -config X86_PM_TIMER - bool "PM timer" if EMBEDDED - depends on ACPI - default y - help - Support the ACPI PM timer for time keeping. This is slow, - but is useful on some chipsets without HPET on systems with more - than one CPU. On a single processor or single socket multi core - system it is normally not required. - When the PM timer is active 64bit vsyscalls are disabled - and should not be enabled (/proc/sys/kernel/vsyscall64 should - not be changed). - The kernel selects the PM timer only as a last resort, so it is - useful to enable just in case. - config HPET_EMULATE_RTC bool "Provide RTC interrupt" depends on HPET_TIMER && RTC=y -- cgit v1.2.3 From 1f9921539208f6d88f600a801e333d718e4a13ff Mon Sep 17 00:00:00 2001 From: Chris McDermott Date: Sun, 26 Feb 2006 04:18:40 +0100 Subject: [PATCH] x86_64: Fix NMI watchdog on x460 [description from AK] Old check for the IO-APIC watchdog during the timer check was wrong - it obviously should only drop into this if the IO-APIC watchdog is used. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/io_apic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86_64/kernel/io_apic.c b/arch/x86_64/kernel/io_apic.c index 2585c1d92b2..ffab8a75666 100644 --- a/arch/x86_64/kernel/io_apic.c +++ b/arch/x86_64/kernel/io_apic.c @@ -1850,7 +1850,7 @@ static inline void check_timer(void) } printk(" failed.\n"); - if (nmi_watchdog) { + if (nmi_watchdog == NMI_IO_APIC) { printk(KERN_WARNING "timer doesn't work through the IO-APIC - disabling NMI Watchdog!\n"); nmi_watchdog = 0; } -- cgit v1.2.3 From e2c0388866dc12bef56b178b958f9b778fe6c687 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:46 +0100 Subject: [PATCH] x86_64: Fix the additional_cpus=.. option It didn't set up the CPU possible map early enough, so the option didn't actually work. Noticed by Heiko Carstens Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/setup.c | 6 ++++++ arch/x86_64/kernel/smpboot.c | 2 +- 2 files changed, 7 insertions(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index 9435ab7d6fb..5de7eaf5d97 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -423,6 +423,12 @@ static __init void parse_cmdline_early (char ** cmdline_p) else if(!memcmp(from, "elfcorehdr=", 11)) elfcorehdr_addr = memparse(from+11, &from); #endif + +#ifdef CONFIG_SMP + else if (!memcmp(from, "additional_cpus=", 16)) + setup_additional_cpus(from+16); +#endif + next_char: c = *(from++); if (!c) diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c index 67e4e28f4df..b82eb86e4f5 100644 --- a/arch/x86_64/kernel/smpboot.c +++ b/arch/x86_64/kernel/smpboot.c @@ -1244,7 +1244,7 @@ void __cpu_die(unsigned int cpu) printk(KERN_ERR "CPU %u didn't die...\n", cpu); } -static __init int setup_additional_cpus(char *s) +__init int setup_additional_cpus(char *s) { return get_option(&s, &additional_cpus); } -- cgit v1.2.3 From e8b917775b572bc27de105f1317c2de4335db5b3 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:49 +0100 Subject: [PATCH] x86_64: Move the SMP time selection earlier SMP time selection originally ran after all CPUs were brought up because it needed to know the number of CPUs to decide if it needs an MP safe timer or not. This is not needed anymore because we know present CPUs early. This fixes a couple of problems: - apicmaintimer didn't always work because it relied on state that was set up time_init_gtod too late. - The output for the used timer in early kernel log was misleading because time_init_gtod could actually change it later. Now always print the final timer choice Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/smpboot.c | 2 -- arch/x86_64/kernel/time.c | 22 +++++++++++----------- 2 files changed, 11 insertions(+), 13 deletions(-) (limited to 'arch') diff --git a/arch/x86_64/kernel/smpboot.c b/arch/x86_64/kernel/smpboot.c index b82eb86e4f5..66e98659d07 100644 --- a/arch/x86_64/kernel/smpboot.c +++ b/arch/x86_64/kernel/smpboot.c @@ -1152,8 +1152,6 @@ void __init smp_cpus_done(unsigned int max_cpus) setup_ioapic_dest(); #endif - time_init_gtod(); - check_nmi_watchdog(); } diff --git a/arch/x86_64/kernel/time.c b/arch/x86_64/kernel/time.c index 67841d11ed1..3080f84bf7b 100644 --- a/arch/x86_64/kernel/time.c +++ b/arch/x86_64/kernel/time.c @@ -48,6 +48,8 @@ static void cpufreq_delayed_get(void); extern void i8254_timer_resume(void); extern int using_apic_timer; +static char *time_init_gtod(void); + DEFINE_SPINLOCK(rtc_lock); DEFINE_SPINLOCK(i8253_lock); @@ -901,6 +903,7 @@ static struct irqaction irq0 = { void __init time_init(void) { char *timename; + char *gtod; #ifdef HPET_HACK_ENABLE_DANGEROUS if (!vxtime.hpet_address) { @@ -945,21 +948,19 @@ void __init time_init(void) timename = "PIT"; } - printk(KERN_INFO "time.c: Using %ld.%06ld MHz %s timer.\n", - vxtime_hz / 1000000, vxtime_hz % 1000000, timename); + vxtime.mode = VXTIME_TSC; + gtod = time_init_gtod(); + + printk(KERN_INFO "time.c: Using %ld.%06ld MHz WALL %s GTOD %s timer.\n", + vxtime_hz / 1000000, vxtime_hz % 1000000, timename, gtod); printk(KERN_INFO "time.c: Detected %d.%03d MHz processor.\n", cpu_khz / 1000, cpu_khz % 1000); - vxtime.mode = VXTIME_TSC; vxtime.quot = (1000000L << 32) / vxtime_hz; vxtime.tsc_quot = (1000L << 32) / cpu_khz; vxtime.last_tsc = get_cycles_sync(); setup_irq(0, &irq0); set_cyc2ns_scale(cpu_khz); - -#ifndef CONFIG_SMP - time_init_gtod(); -#endif } /* @@ -981,9 +982,9 @@ __cpuinit int unsynchronized_tsc(void) } /* - * Decide after all CPUs are booted what mode gettimeofday should use. + * Decide what mode gettimeofday should use. */ -void __init time_init_gtod(void) +__init static char *time_init_gtod(void) { char *timetype; @@ -1011,8 +1012,7 @@ void __init time_init_gtod(void) timetype = hpet_use_timer ? "HPET/TSC" : "PIT/TSC"; vxtime.mode = VXTIME_TSC; } - - printk(KERN_INFO "time.c: Using %s based timekeeping.\n", timetype); + return timetype; } __setup("report_lost_ticks", time_setup); -- cgit v1.2.3 From ab9b32ee626e9b6df4ce2560a70ae15e62423cf4 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 26 Feb 2006 04:18:52 +0100 Subject: [PATCH] x86_64: Better ATI timer fix The previous experiment for using apicmaintimer on ATI systems didn't work out very well. In particular laptops with C2/C3 support often don't let it tick during idle, which makes it useless. There were also some other bugs that made the apicmaintimer often not used at all. I tried some other experiments - running timer over RTC and some other things but they didn't really work well neither. I rechecked the specs now and it turns out this simple change is actually enough to avoid the double ticks on the ATI systems. We just turn off IRQ 0 in the 8254 and only route it directly using the IO-APIC. I tested it on a few ATI systems and it worked there. In fact it worked on all chipsets (NVidia, Intel, AMD, ATI) I tried it on. According to the ACPI spec routing should always work through the IO-APIC so I think it's the correct thing to do anyways (and most of the old gunk in check_timer should be thrown away for x86-64). But for 2.6.16 it's best to do a fairly minimal change: - Use the known to be working everywhere-but-ATI IRQ0 both over 8254 and IO-APIC setup everywhere - Except on ATI disable IRQ0 in the 8254 - Remove the code to select apicmaintimer on ATI chipsets - Add some boot options to allow to override this (just paranoia) In 2.6.17 I hope to switch the default over to this for everybody. Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/io_apic.c | 46 ++++++++++++++++++++++++++++---------------- 1 file changed, 29 insertions(+), 17 deletions(-) (limited to 'arch') diff --git a/arch/x86_64/kernel/io_apic.c b/arch/x86_64/kernel/io_apic.c index ffab8a75666..ffed464e6b1 100644 --- a/arch/x86_64/kernel/io_apic.c +++ b/arch/x86_64/kernel/io_apic.c @@ -50,6 +50,8 @@ static int no_timer_check; int disable_timer_pin_1 __initdata; +int timer_over_8254 __initdata = 1; + /* Where if anywhere is the i8259 connect in external int mode */ static struct { int pin, apic; } ioapic_i8259 = { -1, -1 }; @@ -251,6 +253,20 @@ static int __init enable_ioapic_setup(char *str) __setup("noapic", disable_ioapic_setup); __setup("apic", enable_ioapic_setup); +static int __init setup_disable_8254_timer(char *s) +{ + timer_over_8254 = -1; + return 1; +} +static int __init setup_enable_8254_timer(char *s) +{ + timer_over_8254 = 2; + return 1; +} + +__setup("disable_8254_timer", setup_disable_8254_timer); +__setup("enable_8254_timer", setup_enable_8254_timer); + #include #include #include @@ -309,27 +325,20 @@ void __init check_ioapic(void) #endif /* RED-PEN skip them on mptables too? */ return; + + /* This should be actually default, but + for 2.6.16 let's do it for ATI only where + it's really needed. */ case PCI_VENDOR_ID_ATI: - if (apic_runs_main_timer != 0) - break; -#ifdef CONFIG_ACPI - /* Don't do this for laptops right - right now because their timer - doesn't necessarily tick in C2/3 */ - if (acpi_fadt.revision >= 3 && - (acpi_fadt.plvl2_lat + acpi_fadt.plvl3_lat) < 1100) { - printk(KERN_INFO -"ATI board detected, but seems to be a laptop. Timer might be shakey, sorry\n"); - break; - } -#endif + if (timer_over_8254 == 1) { + timer_over_8254 = 0; printk(KERN_INFO - "ATI board detected. Using APIC/PM timer.\n"); - apic_runs_main_timer = 1; - nohpet = 1; + "ATI board detected. Disabling timer routing over 8254.\n"); + } return; } + /* No multi-function device? */ type = read_pci_config_byte(num,slot,func, PCI_HEADER_TYPE); @@ -1773,6 +1782,8 @@ static inline void unlock_ExtINT_logic(void) * a wide range of boards and BIOS bugs. Fortunately only the timer IRQ * is so screwy. Thanks to Brian Perkins for testing/hacking this beast * fanatically on his truly buggy board. + * + * FIXME: really need to revamp this for modern platforms only. */ static inline void check_timer(void) { @@ -1795,7 +1806,8 @@ static inline void check_timer(void) */ apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT); init_8259A(1); - enable_8259A_irq(0); + if (timer_over_8254 > 0) + enable_8259A_irq(0); pin1 = find_isa_irq_pin(0, mp_INT); apic1 = find_isa_irq_apic(0, mp_INT); -- cgit v1.2.3 From d51761233d9e3be4cdf10f7482a50463bbd78c78 Mon Sep 17 00:00:00 2001 From: Brian Magnuson Date: Mon, 27 Feb 2006 04:02:04 +0100 Subject: [PATCH] fix build on x86_64 with !CONFIG_HOTPLUG_CPU The commit e2c0388866dc12bef56b178b958f9b778fe6c687 added setup_additional_cpus to setup.c but this is only defined if CONFIG_HOTPLUG_CPU is set. This patch changes the #ifdef to reflect that. Signed-off-by: Brian Magnuson Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds --- arch/x86_64/kernel/setup.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/x86_64/kernel/setup.c b/arch/x86_64/kernel/setup.c index 5de7eaf5d97..aa55e3cec66 100644 --- a/arch/x86_64/kernel/setup.c +++ b/arch/x86_64/kernel/setup.c @@ -424,7 +424,7 @@ static __init void parse_cmdline_early (char ** cmdline_p) elfcorehdr_addr = memparse(from+11, &from); #endif -#ifdef CONFIG_SMP +#ifdef CONFIG_HOTPLUG_CPU else if (!memcmp(from, "additional_cpus=", 16)) setup_additional_cpus(from+16); #endif -- cgit v1.2.3 From e18f9b4be430189d79a01a75734bf7cfdc22cc3f Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Sun, 26 Feb 2006 17:07:45 -0600 Subject: [PATCH] fix voyager after topology.c move Commit 9c869edac591977314323a4eaad5f7633fca684f broke voyager again rather subtly because it already had its own topology exporting functions, so now each CPU gets registered twice. I think we can actually use the generic ones, so I don't propose reverting it. The attached should eliminate the voyager topology functions in favour of the generic ones. I also added a define to ensure voyager is never hotplug CPU (we don't have the support in the SMP harness). Signed-off-by: James Bottomley Signed-off-by: Linus Torvalds --- arch/i386/Kconfig | 2 +- arch/i386/mach-voyager/voyager_basic.c | 14 -------------- 2 files changed, 1 insertion(+), 15 deletions(-) (limited to 'arch') diff --git a/arch/i386/Kconfig b/arch/i386/Kconfig index 0afec8566e7..af411596a31 100644 --- a/arch/i386/Kconfig +++ b/arch/i386/Kconfig @@ -733,7 +733,7 @@ config PHYSICAL_START config HOTPLUG_CPU bool "Support for hot-pluggable CPUs (EXPERIMENTAL)" - depends on SMP && HOTPLUG && EXPERIMENTAL + depends on SMP && HOTPLUG && EXPERIMENTAL && !X86_VOYAGER ---help--- Say Y here to experiment with turning CPUs off and on. CPUs can be controlled through /sys/devices/system/cpu. diff --git a/arch/i386/mach-voyager/voyager_basic.c b/arch/i386/mach-voyager/voyager_basic.c index 6761d294f26..b584060ec00 100644 --- a/arch/i386/mach-voyager/voyager_basic.c +++ b/arch/i386/mach-voyager/voyager_basic.c @@ -25,7 +25,6 @@ #include #include #include -#include #include #include #include @@ -331,16 +330,3 @@ void machine_power_off(void) if (pm_power_off) pm_power_off(); } - -static struct i386_cpu cpu_devices[NR_CPUS]; - -static int __init topology_init(void) -{ - int i; - - for_each_present_cpu(i) - register_cpu(&cpu_devices[i].cpu, i, NULL); - return 0; -} - -subsys_initcall(topology_init); -- cgit v1.2.3 From 7abea9214585823f7f19d91872d7c6f8874bef9a Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sat, 25 Feb 2006 13:39:56 -0800 Subject: [SPARC64]: Make cpu_present_map available earlier. The change to kernel/sched.c's init code to use for_each_cpu() requires that the cpu_possible_map be setup much earlier. Set it up via setup_arch(), constrained to NR_CPUS, and later constrain it to max_cpus in smp_prepare_cpus(). This fixes SMP booting on sparc64. Signed-off-by: David S. Miller --- arch/sparc64/kernel/setup.c | 2 ++ arch/sparc64/kernel/smp.c | 28 +++++++++++++++++++--------- 2 files changed, 21 insertions(+), 9 deletions(-) (limited to 'arch') diff --git a/arch/sparc64/kernel/setup.c b/arch/sparc64/kernel/setup.c index 054461e6946..158bd31e15b 100644 --- a/arch/sparc64/kernel/setup.c +++ b/arch/sparc64/kernel/setup.c @@ -542,6 +542,8 @@ void __init setup_arch(char **cmdline_p) } #endif + smp_setup_cpu_possible_map(); + paging_init(); } diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c index 1fb6323e65a..1f7ad8a6905 100644 --- a/arch/sparc64/kernel/smp.c +++ b/arch/sparc64/kernel/smp.c @@ -1079,18 +1079,12 @@ int setup_profiling_timer(unsigned int multiplier) return 0; } +/* Constrain the number of cpus to max_cpus. */ void __init smp_prepare_cpus(unsigned int max_cpus) { - int instance, mid; - - instance = 0; - while (!cpu_find_by_instance(instance, NULL, &mid)) { - if (mid < max_cpus) - cpu_set(mid, phys_cpu_present_map); - instance++; - } - if (num_possible_cpus() > max_cpus) { + int instance, mid; + instance = 0; while (!cpu_find_by_instance(instance, NULL, &mid)) { if (mid != boot_cpu_id) { @@ -1105,6 +1099,22 @@ void __init smp_prepare_cpus(unsigned int max_cpus) smp_store_cpu_info(boot_cpu_id); } +/* Set this up early so that things like the scheduler can init + * properly. We use the same cpu mask for both the present and + * possible cpu map. + */ +void __init smp_setup_cpu_possible_map(void) +{ + int instance, mid; + + instance = 0; + while (!cpu_find_by_instance(instance, NULL, &mid)) { + if (mid < NR_CPUS) + cpu_set(mid, phys_cpu_present_map); + instance++; + } +} + void __devinit smp_prepare_boot_cpu(void) { if (hard_smp_processor_id() >= NR_CPUS) { -- cgit v1.2.3 From add2b6fdae9d7fc251c229e76252f731917094c4 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 26 Feb 2006 20:24:40 -0800 Subject: Make Kprobes depend on modules Commit 9ec4b1f356b3bad928ae8e2aa9caebfa737d52df made kprobes not compile without module support, so just make that clear in the Kconfig file. Also, since it's marked EXPERIMENTAL, make that dependency explicit too. Signed-off-by: Linus Torvalds --- arch/i386/Kconfig | 1 + arch/ia64/Kconfig | 1 + arch/powerpc/Kconfig | 2 +- arch/sparc64/Kconfig | 1 + arch/x86_64/Kconfig | 1 + 5 files changed, 5 insertions(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/i386/Kconfig b/arch/i386/Kconfig index af411596a31..5b1a7d46d1d 100644 --- a/arch/i386/Kconfig +++ b/arch/i386/Kconfig @@ -1060,6 +1060,7 @@ source "arch/i386/oprofile/Kconfig" config KPROBES bool "Kprobes (EXPERIMENTAL)" + depends on EXPERIMENTAL && MODULES help Kprobes allows you to trap at almost any kernel address and execute a callback function. register_kprobe() establishes diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index 845cd0902a5..a85ea9d37f0 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -453,6 +453,7 @@ source "arch/ia64/oprofile/Kconfig" config KPROBES bool "Kprobes (EXPERIMENTAL)" + depends on EXPERIMENTAL && MODULES help Kprobes allows you to trap at almost any kernel address and execute a callback function. register_kprobe() establishes diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 80d114a3a83..a834f9e0bbb 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -963,7 +963,7 @@ source "arch/powerpc/oprofile/Kconfig" config KPROBES bool "Kprobes (EXPERIMENTAL)" - depends on PPC64 + depends on PPC64 && EXPERIMENTAL && MODULES help Kprobes allows you to trap at almost any kernel address and execute a callback function. register_kprobe() establishes diff --git a/arch/sparc64/Kconfig b/arch/sparc64/Kconfig index ab733be9af0..4c0a50a7655 100644 --- a/arch/sparc64/Kconfig +++ b/arch/sparc64/Kconfig @@ -383,6 +383,7 @@ source "arch/sparc64/oprofile/Kconfig" config KPROBES bool "Kprobes (EXPERIMENTAL)" + depends on EXPERIMENTAL && MODULES help Kprobes allows you to trap at almost any kernel address and execute a callback function. register_kprobe() establishes diff --git a/arch/x86_64/Kconfig b/arch/x86_64/Kconfig index babc31b3ef1..e18eb79bf85 100644 --- a/arch/x86_64/Kconfig +++ b/arch/x86_64/Kconfig @@ -577,6 +577,7 @@ source "arch/x86_64/oprofile/Kconfig" config KPROBES bool "Kprobes (EXPERIMENTAL)" + depends on EXPERIMENTAL && MODULES help Kprobes allows you to trap at almost any kernel address and execute a callback function. register_kprobe() establishes -- cgit v1.2.3 From 9b6695a8adfe0916e81ddd810a5b9db3eb8b0e46 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Thu, 23 Feb 2006 12:23:27 +0000 Subject: [MIPS] SMP: Fix initialization order bug. A recent change requires cpu_possible_map to be initialized before smp_sched_init() but most MIPS platforms were initializing their processors in the prom_prepare_cpus callback of smp_prepare_cpus. The simple fix of calling prom_prepare_cpus from one of the earlier SMP initialization hooks doesn't work well either since IPIs may require init_IRQ() to have completed, so bit the bullet and split prom_prepare_cpus into two initialization functions, plat_smp_setup which is called early from setup_arch and plat_prepare_cpus called where prom_prepare_cpus used to be called. Signed-off-by: Ralf Baechle --- arch/mips/kernel/setup.c | 3 +++ arch/mips/kernel/smp.c | 2 +- arch/mips/kernel/smp_mt.c | 13 +++++++------ arch/mips/pmc-sierra/yosemite/smp.c | 24 ++++++------------------ arch/mips/sgi-ip27/ip27-smp.c | 7 ++++++- arch/mips/sibyte/cfe/smp.c | 10 +++++++--- 6 files changed, 30 insertions(+), 29 deletions(-) (limited to 'arch') diff --git a/arch/mips/kernel/setup.c b/arch/mips/kernel/setup.c index d86affa2127..d9293c558e4 100644 --- a/arch/mips/kernel/setup.c +++ b/arch/mips/kernel/setup.c @@ -540,6 +540,9 @@ void __init setup_arch(char **cmdline_p) sparse_init(); paging_init(); resource_init(); +#ifdef CONFIG_SMP + plat_smp_setup(); +#endif } int __init fpu_disable(char *s) diff --git a/arch/mips/kernel/smp.c b/arch/mips/kernel/smp.c index 5e189862e52..06ed9075242 100644 --- a/arch/mips/kernel/smp.c +++ b/arch/mips/kernel/smp.c @@ -236,7 +236,7 @@ void __init smp_prepare_cpus(unsigned int max_cpus) init_new_context(current, &init_mm); current_thread_info()->cpu = 0; smp_tune_scheduling(); - prom_prepare_cpus(max_cpus); + plat_prepare_cpus(max_cpus); } /* preload SMP state for boot cpu */ diff --git a/arch/mips/kernel/smp_mt.c b/arch/mips/kernel/smp_mt.c index c930364830d..993b8bf56aa 100644 --- a/arch/mips/kernel/smp_mt.c +++ b/arch/mips/kernel/smp_mt.c @@ -143,7 +143,7 @@ static struct irqaction irq_call = { * Make sure all CPU's are in a sensible state before we boot any of the * secondarys */ -void prom_prepare_cpus(unsigned int max_cpus) +void plat_smp_setup(void) { unsigned long val; int i, num; @@ -179,11 +179,9 @@ void prom_prepare_cpus(unsigned int max_cpus) write_vpe_c0_vpeconf0(tmp); /* Record this as available CPU */ - if (i < max_cpus) { - cpu_set(i, phys_cpu_present_map); - __cpu_number_map[i] = ++num; - __cpu_logical_map[num] = i; - } + cpu_set(i, phys_cpu_present_map); + __cpu_number_map[i] = ++num; + __cpu_logical_map[num] = i; } /* disable multi-threading with TC's */ @@ -241,7 +239,10 @@ void prom_prepare_cpus(unsigned int max_cpus) set_vi_handler (MIPS_CPU_IPI_RESCHED_IRQ, ipi_resched_dispatch); set_vi_handler (MIPS_CPU_IPI_CALL_IRQ, ipi_call_dispatch); } +} +void __init plat_prepare_cpus(unsigned int max_cpus) +{ cpu_ipi_resched_irq = MIPSCPU_INT_BASE + MIPS_CPU_IPI_RESCHED_IRQ; cpu_ipi_call_irq = MIPSCPU_INT_BASE + MIPS_CPU_IPI_CALL_IRQ; diff --git a/arch/mips/pmc-sierra/yosemite/smp.c b/arch/mips/pmc-sierra/yosemite/smp.c index 7f8fda96219..c197311e15d 100644 --- a/arch/mips/pmc-sierra/yosemite/smp.c +++ b/arch/mips/pmc-sierra/yosemite/smp.c @@ -50,37 +50,25 @@ void __init prom_grab_secondary(void) * We don't want to start the secondary CPU yet nor do we have a nice probing * feature in PMON so we just assume presence of the secondary core. */ -static char maxcpus_string[] __initdata = - KERN_WARNING "max_cpus set to 0; using 1 instead\n"; - -void __init prom_prepare_cpus(unsigned int max_cpus) +void __init plat_smp_setup(void) { - int enabled = 0, i; - - if (max_cpus == 0) { - printk(maxcpus_string); - max_cpus = 1; - } + int i; cpus_clear(phys_cpu_present_map); for (i = 0; i < 2; i++) { - if (i == max_cpus) - break; - - /* - * The boot CPU - */ cpu_set(i, phys_cpu_present_map); __cpu_number_map[i] = i; __cpu_logical_map[i] = i; - enabled++; } +} +void __init plat_prepare_cpus(unsigned int max_cpus) +{ /* * Be paranoid. Enable the IPI only if we're really about to go SMP. */ - if (enabled > 1) + if (cpus_weight(cpu_possible_map)) set_c0_status(STATUSF_IP5); } diff --git a/arch/mips/sgi-ip27/ip27-smp.c b/arch/mips/sgi-ip27/ip27-smp.c index dbef3f6b565..09fa7f5216f 100644 --- a/arch/mips/sgi-ip27/ip27-smp.c +++ b/arch/mips/sgi-ip27/ip27-smp.c @@ -140,7 +140,7 @@ static __init void intr_clear_all(nasid_t nasid) REMOTE_HUB_CLR_INTR(nasid, i); } -void __init prom_prepare_cpus(unsigned int max_cpus) +void __init plat_smp_setup(void) { cnodeid_t cnode; @@ -161,6 +161,11 @@ void __init prom_prepare_cpus(unsigned int max_cpus) alloc_cpupda(0, 0); } +void __init plat_prepare_cpus(unsigned int max_cpus) +{ + /* We already did everything necessary earlier */ +} + /* * Launch a slave into smp_bootstrap(). It doesn't take an argument, and we * set sp to the kernel stack of the newly created idle process, gp to the proc diff --git a/arch/mips/sibyte/cfe/smp.c b/arch/mips/sibyte/cfe/smp.c index 4477af3d807..eab20e2db32 100644 --- a/arch/mips/sibyte/cfe/smp.c +++ b/arch/mips/sibyte/cfe/smp.c @@ -31,7 +31,7 @@ * * Common setup before any secondaries are started */ -void __init prom_prepare_cpus(unsigned int max_cpus) +void __init plat_smp_setup(void) { int i, num; @@ -40,14 +40,18 @@ void __init prom_prepare_cpus(unsigned int max_cpus) __cpu_number_map[0] = 0; __cpu_logical_map[0] = 0; - for (i=1, num=0; i Date: Tue, 21 Feb 2006 15:59:00 +0900 Subject: [MIPS] Use generic compat routines for readdir, getdents Not just cleanup but also fixes O32 readdir(2) emulation. Signed-off-by: Atsushi Nemoto Signed-off-by: Ralf Baechle --- arch/mips/kernel/linux32.c | 54 ------------------------------------------ arch/mips/kernel/scall64-n32.S | 2 +- arch/mips/kernel/scall64-o32.S | 4 ++-- 3 files changed, 3 insertions(+), 57 deletions(-) (limited to 'arch') diff --git a/arch/mips/kernel/linux32.c b/arch/mips/kernel/linux32.c index 5f68b220c26..e00e5f6e7fd 100644 --- a/arch/mips/kernel/linux32.c +++ b/arch/mips/kernel/linux32.c @@ -161,60 +161,6 @@ out: return error; } -struct dirent32 { - unsigned int d_ino; - unsigned int d_off; - unsigned short d_reclen; - char d_name[NAME_MAX + 1]; -}; - -static void -xlate_dirent(void *dirent64, void *dirent32, long n) -{ - long off; - struct dirent *dirp; - struct dirent32 *dirp32; - - off = 0; - while (off < n) { - dirp = (struct dirent *)(dirent64 + off); - dirp32 = (struct dirent32 *)(dirent32 + off); - off += dirp->d_reclen; - dirp32->d_ino = dirp->d_ino; - dirp32->d_off = (unsigned int)dirp->d_off; - dirp32->d_reclen = dirp->d_reclen; - strncpy(dirp32->d_name, dirp->d_name, dirp->d_reclen - ((3 * 4) + 2)); - } - return; -} - -asmlinkage long -sys32_getdents(unsigned int fd, void * dirent32, unsigned int count) -{ - long n; - void *dirent64; - - dirent64 = (void *)((unsigned long)(dirent32 + (sizeof(long) - 1)) & ~(sizeof(long) - 1)); - if ((n = sys_getdents(fd, dirent64, count - (dirent64 - dirent32))) < 0) - return(n); - xlate_dirent(dirent64, dirent32, n); - return(n); -} - -asmlinkage int old_readdir(unsigned int fd, void * dirent, unsigned int count); - -asmlinkage int -sys32_readdir(unsigned int fd, void * dirent32, unsigned int count) -{ - int n; - struct dirent dirent64; - - if ((n = old_readdir(fd, &dirent64, count)) < 0) - return(n); - xlate_dirent(&dirent64, dirent32, dirent64.d_reclen); - return(n); -} - asmlinkage int sys32_waitpid(compat_pid_t pid, unsigned int *stat_addr, int options) { diff --git a/arch/mips/kernel/scall64-n32.S b/arch/mips/kernel/scall64-n32.S index d87b5446fa1..02c8267e45e 100644 --- a/arch/mips/kernel/scall64-n32.S +++ b/arch/mips/kernel/scall64-n32.S @@ -195,7 +195,7 @@ EXPORT(sysn32_call_table) PTR sys_fdatasync PTR sys_truncate PTR sys_ftruncate /* 6075 */ - PTR sys32_getdents + PTR compat_sys_getdents PTR sys_getcwd PTR sys_chdir PTR sys_fchdir diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall64-o32.S index 5b0414018c9..797e0d87488 100644 --- a/arch/mips/kernel/scall64-o32.S +++ b/arch/mips/kernel/scall64-o32.S @@ -293,7 +293,7 @@ sys_call_table: PTR sys_uselib PTR sys_swapon PTR sys_reboot - PTR sys32_readdir + PTR compat_sys_old_readdir PTR old_mmap /* 4090 */ PTR sys_munmap PTR sys_truncate @@ -345,7 +345,7 @@ sys_call_table: PTR sys_setfsuid PTR sys_setfsgid PTR sys32_llseek /* 4140 */ - PTR sys32_getdents + PTR compat_sys_getdents PTR compat_sys_select PTR sys_flock PTR sys_msync -- cgit v1.2.3 From 637029c6cb5efcbaa3d5831af4c1972bdd629779 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Mon, 27 Feb 2006 20:41:56 -0800 Subject: Revert "[PATCH] x86_64: Only do the clustered systems have unsynchronized TSC assumption on IBM systems" This reverts commit 13a229abc25640813f1480c0478dfc6bdbc1c19e. Quoth Andi: "After some consideration and feedback from various people it turns out this wasn't that good an idea. It has some problems and needs more work. Since it was only an optimization anyways it's best to just back it out again for now." Signed-off-by: Linus Torvalds --- arch/i386/kernel/acpi/boot.c | 3 +++ arch/x86_64/kernel/apic.c | 9 +-------- 2 files changed, 4 insertions(+), 8 deletions(-) (limited to 'arch') diff --git a/arch/i386/kernel/acpi/boot.c b/arch/i386/kernel/acpi/boot.c index 8309a7b2cd6..79577f0ace9 100644 --- a/arch/i386/kernel/acpi/boot.c +++ b/arch/i386/kernel/acpi/boot.c @@ -44,6 +44,9 @@ extern void __init clustered_apic_check(void); extern int gsi_irq_sharing(int gsi); #include +static inline int acpi_madt_oem_check(char *oem_id, char *oem_table_id) { return 0; } + + #else /* X86 */ #ifdef CONFIG_X86_LOCAL_APIC diff --git a/arch/x86_64/kernel/apic.c b/arch/x86_64/kernel/apic.c index d70605eda33..e5b14c57eaa 100644 --- a/arch/x86_64/kernel/apic.c +++ b/arch/x86_64/kernel/apic.c @@ -962,14 +962,12 @@ void smp_apic_timer_interrupt(struct pt_regs *regs) irq_exit(); } -int __initdata unsync_tsc_on_multicluster; - /* * oem_force_hpet_timer -- force HPET mode for some boxes. * * Thus far, the major user of this is IBM's Summit2 series: * - * Some clustered boxes may have unsynced TSC problems if they are + * Clustered boxes may have unsynced TSC problems if they are * multi-chassis. Use available data to take a good guess. * If in doubt, go HPET. */ @@ -979,11 +977,6 @@ __cpuinit int oem_force_hpet_timer(void) unsigned id; DECLARE_BITMAP(clustermap, NUM_APIC_CLUSTERS); - /* Only do this check on IBM machines - big Unisys systems - use multiple clusters too, but have synchronized TSC */ - if (!unsync_tsc_on_multicluster) - return 0; - bitmap_zero(clustermap, NUM_APIC_CLUSTERS); for (i = 0; i < NR_CPUS; i++) { -- cgit v1.2.3 From 827c1a6c1a5dcb2902fecfb648f9af6a532934eb Mon Sep 17 00:00:00 2001 From: John Rose Date: Fri, 24 Feb 2006 11:34:23 -0600 Subject: [PATCH] powerpc: fix dynamic PCI probe regression Some hotplug driver functions were migrated to the kernel for use by EEH in commit 2bf6a8fa21570f37fd1789610da30f70a05ac5e3. Previously, the PCI Hotplug module had been changed to use the new OFDT-based PCI probe when appropriate: 5fa80fcdca9d20d30c9ecec30d4dbff4ed93a5c6 When rpaphp_pci_config_slot() was moved from the rpaphp driver to the new kernel function pcibios_add_pci_devices(), the OFDT-based probe stuff was dropped. This patch restores it. Signed-off-by: John Rose Signed-off-by: Paul Mackerras --- arch/powerpc/platforms/pseries/eeh.c | 14 ++++++++++++ arch/powerpc/platforms/pseries/pci_dlpar.c | 36 +++++++++++++++++------------- 2 files changed, 34 insertions(+), 16 deletions(-) (limited to 'arch') diff --git a/arch/powerpc/platforms/pseries/eeh.c b/arch/powerpc/platforms/pseries/eeh.c index 83578313ee7..2ab9dcdfb41 100644 --- a/arch/powerpc/platforms/pseries/eeh.c +++ b/arch/powerpc/platforms/pseries/eeh.c @@ -893,6 +893,20 @@ void eeh_add_device_tree_early(struct device_node *dn) } EXPORT_SYMBOL_GPL(eeh_add_device_tree_early); +void eeh_add_device_tree_late(struct pci_bus *bus) +{ + struct pci_dev *dev; + + list_for_each_entry(dev, &bus->devices, bus_list) { + eeh_add_device_late(dev); + if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE) { + struct pci_bus *subbus = dev->subordinate; + if (subbus) + eeh_add_device_tree_late(subbus); + } + } +} + /** * eeh_add_device_late - perform EEH initialization for the indicated pci device * @dev: pci device for which to set up EEH diff --git a/arch/powerpc/platforms/pseries/pci_dlpar.c b/arch/powerpc/platforms/pseries/pci_dlpar.c index bdaa8aabdaa..f3bad900bbc 100644 --- a/arch/powerpc/platforms/pseries/pci_dlpar.c +++ b/arch/powerpc/platforms/pseries/pci_dlpar.c @@ -106,6 +106,8 @@ pcibios_fixup_new_pci_devices(struct pci_bus *bus, int fix_bus) } } } + + eeh_add_device_tree_late(bus); } EXPORT_SYMBOL_GPL(pcibios_fixup_new_pci_devices); @@ -114,7 +116,6 @@ pcibios_pci_config_bridge(struct pci_dev *dev) { u8 sec_busno; struct pci_bus *child_bus; - struct pci_dev *child_dev; /* Get busno of downstream bus */ pci_read_config_byte(dev, PCI_SECONDARY_BUS, &sec_busno); @@ -129,10 +130,6 @@ pcibios_pci_config_bridge(struct pci_dev *dev) pci_scan_child_bus(child_bus); - list_for_each_entry(child_dev, &child_bus->devices, bus_list) { - eeh_add_device_late(child_dev); - } - /* Fixup new pci devices without touching bus struct */ pcibios_fixup_new_pci_devices(child_bus, 0); @@ -160,18 +157,25 @@ pcibios_add_pci_devices(struct pci_bus * bus) eeh_add_device_tree_early(dn); - /* pci_scan_slot should find all children */ - slotno = PCI_SLOT(PCI_DN(dn->child)->devfn); - num = pci_scan_slot(bus, PCI_DEVFN(slotno, 0)); - if (num) { - pcibios_fixup_new_pci_devices(bus, 1); - pci_bus_add_devices(bus); - } + if (_machine == PLATFORM_PSERIES_LPAR) { + /* use ofdt-based probe */ + of_scan_bus(dn, bus); + if (!list_empty(&bus->devices)) { + pcibios_fixup_new_pci_devices(bus, 0); + pci_bus_add_devices(bus); + } + } else { + /* use legacy probe */ + slotno = PCI_SLOT(PCI_DN(dn->child)->devfn); + num = pci_scan_slot(bus, PCI_DEVFN(slotno, 0)); + if (num) { + pcibios_fixup_new_pci_devices(bus, 1); + pci_bus_add_devices(bus); + } - list_for_each_entry(dev, &bus->devices, bus_list) { - eeh_add_device_late (dev); - if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE) - pcibios_pci_config_bridge(dev); + list_for_each_entry(dev, &bus->devices, bus_list) + if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE) + pcibios_pci_config_bridge(dev); } } EXPORT_SYMBOL_GPL(pcibios_add_pci_devices); -- cgit v1.2.3 From 634473db86502b6444c3cebd279a06e0b8737527 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Sun, 26 Feb 2006 08:09:00 +1100 Subject: [PATCH] powerpc: vdso 64bits gettimeofday bug A bug in the assembly code of the vdso can cause gettimeofday() to hang or to return incorrect results. The wrong register was used to test for pending updates of the calibration variables and to create a dependency for subsequent loads. This fixes it. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Paul Mackerras --- arch/powerpc/kernel/vdso64/gettimeofday.S | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch') diff --git a/arch/powerpc/kernel/vdso64/gettimeofday.S b/arch/powerpc/kernel/vdso64/gettimeofday.S index ccaeda5136d..4ee871f1cad 100644 --- a/arch/powerpc/kernel/vdso64/gettimeofday.S +++ b/arch/powerpc/kernel/vdso64/gettimeofday.S @@ -225,9 +225,9 @@ V_FUNCTION_BEGIN(__do_get_xsec) .cfi_startproc /* check for update count & load values */ 1: ld r8,CFG_TB_UPDATE_COUNT(r3) - andi. r0,r4,1 /* pending update ? loop */ + andi. r0,r8,1 /* pending update ? loop */ bne- 1b - xor r0,r4,r4 /* create dependency */ + xor r0,r8,r8 /* create dependency */ add r3,r3,r0 /* Get TB & offset it */ -- cgit v1.2.3 From 273d2803817c9e050e8d6c3c271db7d61f2fb259 Mon Sep 17 00:00:00 2001 From: Olaf Hering Date: Mon, 27 Feb 2006 15:52:59 +0100 Subject: [PATCH] powerpc: fix NULL pointer in handle_eeh_events This patch fixes a crash in handle_eeh_events, but ethtool -t still doesnt work right. ... pepino:~ # cpu 0x3: Vector: 300 (Data Access) at [c00000005192bbe0] pc: c00000000004a380: .handle_eeh_events+0xe0/0x23c lr: c00000000004a374: .handle_eeh_events+0xd4/0x23c sp: c00000005192be60 msr: 9000000000009032 dar: 268 dsisr: 40000000 current = 0xc0000001fe7bf1a0 paca = 0xc00000000048b280 pid = 16322, comm = eehd enter ? for help [c00000005192bf00] c00000000004a808 .eeh_event_handler+0xcc/0x130 [c00000005192bf90] c000000000025e00 .kernel_thread+0x4c/0x68 ... (none):/# /usr/sbin/ethtool -i eth0 driver: e100 version: 3.5.10-k2-NAPI firmware-version: N/A bus-info: 0000:21:01.0 (none):/# /usr/sbin/ethtool -t eth0 Call Trace: [C00000000F8DEFF0] [C00000000000F270] .show_stack+0x74/0x1b4 (unreliable) [C00000000F8DF0A0] [C000000000049D04] .eeh_dn_check_failure+0x290/0x2d8 [C00000000F8DF150] [C000000000049E58] .eeh_check_failure+0x10c/0x138 [C00000000F8DF1E0] [C0000000002DFDB0] .e100_hw_reset+0x70/0xf4 [C00000000F8DF270] [C0000000002E1BBC] .e100_hw_init+0x2c/0x260 [C00000000F8DF310] [C0000000002E2464] .e100_loopback_test+0x8c/0x220 [C00000000F8DF3C0] [C0000000002E28DC] .e100_diag_test+0xdc/0x16c [C00000000F8DF490] [C000000000420BE0] .dev_ethtool+0xf24/0x14f8 [C00000000F8DF8F0] [C00000000041F4A8] .dev_ioctl+0x5cc/0x740 [C00000000F8DFA20] [C00000000040FEFC] .sock_ioctl+0x3d0/0x404 [C00000000F8DFAC0] [C0000000000D513C] .do_ioctl+0x68/0x108 [C00000000F8DFB50] [C0000000000D56B0] .vfs_ioctl+0x4d4/0x510 [C00000000F8DFC10] [C0000000000D5740] .sys_ioctl+0x54/0x94 [C00000000F8DFCC0] [C0000000000FB6EC] .ethtool_ioctl+0x11c/0x150 [C00000000F8DFD60] [C0000000000F7E40] .compat_sys_ioctl+0x338/0x3bc [C00000000F8DFE30] [C00000000000871C] syscall_exit+0x0/0x40 EEH: Detected PCI bus error on device 0000:21:01.0 EEH: This PCI device has failed 1 times since last reboot: - modprobe: FATAL: Could not load /lib/modules/2.6.16-rc4-git7/modules.dep: No such file or directory Cannot get strings: No such device (none):/# (none):/# EEH: Unable to configure device bridge (-3) for /pci@400000000110/pci@2,2 (none):/# Call Trace: [C00000000FA17940] [C00000000000F270] .show_stack+0x74/0x1b4 (unreliable) [C00000000FA179F0] [C000000000049D04] .eeh_dn_check_failure+0x290/0x2d8 [C00000000FA17AA0] [C00000000001E114] .rtas_read_config+0x120/0x154 [C00000000FA17B40] [C000000000049664] .early_enable_eeh+0x274/0x2bc [C00000000FA17C00] [C000000000049708] .eeh_add_device_early+0x5c/0x6c [C00000000FA17C90] [C000000000049748] .eeh_add_device_tree_early+0x30/0x5c [C00000000FA17D20] [C000000000046568] .pcibios_add_pci_devices+0x8c/0x1f8 [C00000000FA17DD0] [C00000000004A528] .eeh_reset_device+0xe0/0x110 [C00000000FA17E60] [C00000000004A698] .handle_eeh_events+0x140/0x250 [C00000000FA17F00] [C00000000004AC7C] .eeh_event_handler+0xe8/0x140 [C00000000FA17F90] [C000000000025784] .kernel_thread+0x4c/0x68 EEH: Detected PCI bus error on device EEH: This PCI device has failed 1 times since last reboot: - EEH: Unable to configure device bridge (-3) for /pci@400000000110/pci@2,2 Call Trace: [C00000000FA17940] [C00000000000F270] .show_stack+0x74/0x1b4 (unreliable) [C00000000FA179F0] [C000000000049D04] .eeh_dn_check_failure+0x290/0x2d8 [C00000000FA17AA0] [C00000000001E114] .rtas_read_config+0x120/0x154 [C00000000FA17B40] [C000000000049664] .early_enable_eeh+0x274/0x2bc [C00000000FA17C00] [C000000000049708] .eeh_add_device_early+0x5c/0x6c [C00000000FA17C90] [C000000000049748] .eeh_add_device_tree_early+0x30/0x5c [C00000000FA17D20] [C000000000046568] .pcibios_add_pci_devices+0x8c/0x1f8 [C00000000FA17DD0] [C00000000004A528] .eeh_reset_device+0xe0/0x110 [C00000000FA17E60] [C00000000004A698] .handle_eeh_events+0x140/0x250 [C00000000FA17F00] [C00000000004AC7C] .eeh_event_handler+0xe8/0x140 [C00000000FA17F90] [C000000000025784] .kernel_thread+0x4c/0x68 EEH: Detected PCI bus error on device EEH: This PCI device has failed 1 times since last reboot: - EEH: Unable to configure device bridge (-3) for /pci@400000000110/pci@2,2 Call Trace: [C00000000FA17940] [C00000000000F270] .show_stack+0x74/0x1b4 (unreliable) [C00000000FA179F0] [C000000000049D04] .eeh_dn_check_failure+0x290/0x2d8 [C00000000FA17AA0] [C00000000001E114] .rtas_read_config+0x120/0x154 [C00000000FA17B40] [C000000000049664] .early_enable_eeh+0x274/0x2bc [C00000000FA17C00] [C000000000049708] .eeh_add_device_early+0x5c/0x6c [C00000000FA17C90] [C000000000049748] .eeh_add_device_tree_early+0x30/0x5c [C00000000FA17D20] [C000000000046568] .pcibios_add_pci_devices+0x8c/0x1f8 [C00000000FA17DD0] [C00000000004A528] .eeh_reset_device+0xe0/0x110 [C00000000FA17E60] [C00000000004A698] .handle_eeh_events+0x140/0x250 [C00000000FA17F00] [C00000000004AC7C] .eeh_event_handler+0xe8/0x140 [C00000000FA17F90] [C000000000025784] .kernel_thread+0x4c/0x68 EEH: Detected PCI bus error on device and so on Signed-off-by: Olaf Hering Acked-by: Linas Vepstas Signed-off-by: Paul Mackerras --- arch/powerpc/platforms/pseries/eeh_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/powerpc/platforms/pseries/eeh_driver.c b/arch/powerpc/platforms/pseries/eeh_driver.c index e3cbba49fd6..b811d5ff92f 100644 --- a/arch/powerpc/platforms/pseries/eeh_driver.c +++ b/arch/powerpc/platforms/pseries/eeh_driver.c @@ -37,7 +37,7 @@ static inline const char * pcid_name (struct pci_dev *pdev) { - if (pdev->dev.driver) + if (pdev && pdev->dev.driver) return pdev->dev.driver->name; return ""; } -- cgit v1.2.3 From 56ec6462af9cba56a04439154e5768672d6f390f Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Tue, 28 Feb 2006 14:54:26 +1100 Subject: [PATCH] powerpc/iseries: Fix double phys_to_abs bug in htab_bolt_mapping Before the merge I updated create_pte_mapping() to work for iSeries, by calling iSeries_hpte_bolt_or_insert. (4c55130b2aa93370f1bf52d2304394e91cf8ee39) Later we changed iSeries_hpte_insert to cope with the bolting case, and called that instead from create_pte_mapping() (which was renamed to htab_bolt_mapping) (3c726f8dee6f55e96475574e9f645327e461884c). Unfortunately that change introduced a subtle bug, where we pass an absolute address to iSeries_hpte_insert() where it expects a physical address. This leads to us calling phys_to_abs() twice on the physical address, which is seriously bogus. This only causes a problem if the absolute address from the first translation can be looked up again in the chunk_map, which depends on the size and layout of memory. I've seen it fail on one box, but not others. The minimal fix is to pass the physical address to iSeries_hpte_insert(). For 2.6.17 we should make phys_to_abs() BUG if we try to double-translate an address. Signed-off-by: Michael Ellerman Signed-off-by: Paul Mackerras --- arch/powerpc/mm/hash_utils_64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch') diff --git a/arch/powerpc/mm/hash_utils_64.c b/arch/powerpc/mm/hash_utils_64.c index b1f614c612d..e9d589eefc1 100644 --- a/arch/powerpc/mm/hash_utils_64.c +++ b/arch/powerpc/mm/hash_utils_64.c @@ -169,7 +169,7 @@ int htab_bolt_mapping(unsigned long vstart, unsigned long vend, #ifdef CONFIG_PPC_ISERIES if (_machine == PLATFORM_ISERIES_LPAR) ret = iSeries_hpte_insert(hpteg, va, - virt_to_abs(paddr), + __pa(vaddr), tmp_mode, HPTE_V_BOLTED, psize); -- cgit v1.2.3