From acff181d3574244e651913df77332e897b88bff4 Mon Sep 17 00:00:00 2001 From: roel kluin Date: Tue, 21 Oct 2008 19:49:09 -0400 Subject: [PATCH 01/10] printk: remove unused code from kernel/printk.c both log_buf_copy() and log_buf_len are unused. Signed-off-by: Ingo Molnar --- kernel/printk.c | 39 --------------------------------------- 1 file changed, 39 deletions(-) diff --git a/kernel/printk.c b/kernel/printk.c index 6341af77eb65..f492f1583d77 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -232,45 +232,6 @@ static inline void boot_delay_msec(void) } #endif -/* - * Return the number of unread characters in the log buffer. - */ -static int log_buf_get_len(void) -{ - return logged_chars; -} - -/* - * Copy a range of characters from the log buffer. - */ -int log_buf_copy(char *dest, int idx, int len) -{ - int ret, max; - bool took_lock = false; - - if (!oops_in_progress) { - spin_lock_irq(&logbuf_lock); - took_lock = true; - } - - max = log_buf_get_len(); - if (idx < 0 || idx >= max) { - ret = -1; - } else { - if (len > max) - len = max; - ret = len; - idx += (log_end - max); - while (len-- > 0) - dest[len] = LOG_BUF(idx + len); - } - - if (took_lock) - spin_unlock_irq(&logbuf_lock); - - return ret; -} - /* * Commands to do_syslog: * From bea92112415635ecb7e681355834413c7c048f67 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Wed, 22 Oct 2008 19:31:11 +0900 Subject: [PATCH 02/10] kernel/resource: fix reserve_region_with_split() section mismatch Impact: cleanup, small kernel text size reduction, no functionality changed reserve_region_with_split() calls in to __reserve_region_with_split(), which is an __init function. The only caller of reserve_region_with_split() is an __init function, so make it __init too. Signed-off-by: Paul Mundt Signed-off-by: Ingo Molnar --- kernel/resource.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/resource.c b/kernel/resource.c index 4089d12af6e0..7fec0e427234 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -571,7 +571,7 @@ static void __init __reserve_region_with_split(struct resource *root, } -void reserve_region_with_split(struct resource *root, +void __init reserve_region_with_split(struct resource *root, resource_size_t start, resource_size_t end, const char *name) { From ed6e5e507e4752c3fb1090d0601f46e7a78c860e Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Tue, 14 Oct 2008 17:50:40 -0700 Subject: [PATCH 03/10] xen: don't reload cr3 on suspend It isn't necessary, and it makes the code needlessly non-portable. Signed-off-by: Jeremy Fitzhardinge Cc: Isaku Yamahata Signed-off-by: Ingo Molnar --- drivers/xen/manage.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/xen/manage.c b/drivers/xen/manage.c index d0e87cbe157c..9b91617b9582 100644 --- a/drivers/xen/manage.c +++ b/drivers/xen/manage.c @@ -39,8 +39,6 @@ static int xen_suspend(void *data) BUG_ON(!irqs_disabled()); - load_cr3(swapper_pg_dir); - err = device_power_down(PMSG_SUSPEND); if (err) { printk(KERN_ERR "xen_suspend: device_power_down failed: %d\n", From 75909fd619d15400e7c6d0fc3af09838ee8b166e Mon Sep 17 00:00:00 2001 From: Isaku Yamahata Date: Tue, 14 Oct 2008 17:50:41 -0700 Subject: [PATCH 04/10] xen: portability clean up and some minor clean up for xencomm.c clean up of xencomm.c. is_phys_contiguous() is arch dependent function that depends on how virtual memory are laid out. So split out the function into arch specific code. Signed-off-by: Isaku Yamahata Signed-off-by: Jeremy Fitzhardinge Cc: Isaku Yamahata Signed-off-by: Ingo Molnar Cc: "Luck, Tony" --- drivers/xen/xencomm.c | 23 ++++------------------- 1 file changed, 4 insertions(+), 19 deletions(-) diff --git a/drivers/xen/xencomm.c b/drivers/xen/xencomm.c index 797cb4e31f07..a240b2c20b99 100644 --- a/drivers/xen/xencomm.c +++ b/drivers/xen/xencomm.c @@ -23,13 +23,7 @@ #include #include #include -#ifdef __ia64__ -#include /* for is_kern_addr() */ -#endif - -#ifdef HAVE_XEN_PLATFORM_COMPAT_H -#include -#endif +#include /* for xencomm_is_phys_contiguous() */ static int xencomm_init(struct xencomm_desc *desc, void *buffer, unsigned long bytes) @@ -157,20 +151,11 @@ static int xencomm_create(void *buffer, unsigned long bytes, return 0; } -/* check if memory address is within VMALLOC region */ -static int is_phys_contiguous(unsigned long addr) -{ - if (!is_kernel_addr(addr)) - return 0; - - return (addr < VMALLOC_START) || (addr >= VMALLOC_END); -} - static struct xencomm_handle *xencomm_create_inline(void *ptr) { unsigned long paddr; - BUG_ON(!is_phys_contiguous((unsigned long)ptr)); + BUG_ON(!xencomm_is_phys_contiguous((unsigned long)ptr)); paddr = (unsigned long)xencomm_pa(ptr); BUG_ON(paddr & XENCOMM_INLINE_FLAG); @@ -202,7 +187,7 @@ struct xencomm_handle *xencomm_map(void *ptr, unsigned long bytes) int rc; struct xencomm_desc *desc; - if (is_phys_contiguous((unsigned long)ptr)) + if (xencomm_is_phys_contiguous((unsigned long)ptr)) return xencomm_create_inline(ptr); rc = xencomm_create(ptr, bytes, &desc, GFP_KERNEL); @@ -219,7 +204,7 @@ struct xencomm_handle *__xencomm_map_no_alloc(void *ptr, unsigned long bytes, int rc; struct xencomm_desc *desc = NULL; - if (is_phys_contiguous((unsigned long)ptr)) + if (xencomm_is_phys_contiguous((unsigned long)ptr)) return xencomm_create_inline(ptr); rc = xencomm_create_mini(ptr, bytes, xc_desc, From ff3c536291ce96ef6f45704cd37eaed71127dd42 Mon Sep 17 00:00:00 2001 From: Isaku Yamahata Date: Tue, 14 Oct 2008 17:50:44 -0700 Subject: [PATCH 05/10] xen: compilation fix of drivers/xen/events.c on IA64 use set_xen_guest_handle() instead of direct assigning. > linux-2.6/drivers/xen/events.c: In function 'xen_poll_irq': > linux-2.6/drivers/xen/events.c:757: error: incompatible types in assignment > make[4]: *** [drivers/xen/events.o] Error 1 Signed-off-by: Isaku Yamahata Signed-off-by: Jeremy Fitzhardinge Cc: Isaku Yamahata Signed-off-by: Ingo Molnar --- drivers/xen/events.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 9ce1ab6c268d..1e3b934a4cf7 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -774,7 +774,7 @@ void xen_poll_irq(int irq) poll.nr_ports = 1; poll.timeout = 0; - poll.ports = &evtchn; + set_xen_guest_handle(poll.ports, &evtchn); if (HYPERVISOR_sched_op(SCHEDOP_poll, &poll) != 0) BUG(); From a2b89b596c5a0b288adac84b17bdda6bde8d144e Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Thu, 23 Oct 2008 18:42:03 +0900 Subject: [PATCH 06/10] swiotlb: remove panic for alloc_coherent failure swiotlb_alloc_coherent calls panic() when allocated swiotlb pages is not fit for a device's dma mask. However, alloc_coherent failure is not a disaster at all. AFAIK, none of other x86 and IA64 IOMMU implementations don't crash in case of alloc_coherent failure. There are some drivers that don't check alloc_coherent failure but not many (about ten and I've already started to fix some of them). alloc_coherent returns NULL in case of failure so it's likely that these guilty drivers crash immediately. So swiotlb doesn't need to call panic() just for them. Reported-by: Takashi Iwai Signed-off-by: FUJITA Tomonori Tested-by: Takashi Iwai Signed-off-by: Ingo Molnar --- lib/swiotlb.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/lib/swiotlb.c b/lib/swiotlb.c index f8eebd489149..78330c37a61b 100644 --- a/lib/swiotlb.c +++ b/lib/swiotlb.c @@ -497,8 +497,10 @@ swiotlb_alloc_coherent(struct device *hwdev, size_t size, printk("hwdev DMA mask = 0x%016Lx, dev_addr = 0x%016Lx\n", (unsigned long long)*hwdev->dma_mask, (unsigned long long)dev_addr); - panic("swiotlb_alloc_coherent: allocated memory is out of " - "range for device"); + + /* DMA_TO_DEVICE to avoid memcpy in unmap_single */ + unmap_single(hwdev, ret, size, DMA_TO_DEVICE); + return NULL; } *dma_handle = dev_addr; return ret; From 75bebb7f0c2a709812cccb4d3151a21b012c5cad Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Thu, 23 Oct 2008 20:46:55 +0900 Subject: [PATCH 07/10] x86: use GFP_DMA for 24bit coherent_dma_mask dma_alloc_coherent (include/asm-x86/dma-mapping.h) avoids GFP_DMA allocation first and if the allocated address is not fit for the device's coherent_dma_mask, then dma_alloc_coherent does GFP_DMA allocation. This is because dma_alloc_coherent avoids precious GFP_DMA zone if possible. This is also how the old dma_alloc_coherent (arch/x86/kernel/pci-dma.c) works. However, if the coherent_dma_mask of a device is 24bit, there is no point to go into the above GFP_DMA retry mechanism. We had better use GFP_DMA in the first place. Signed-off-by: FUJITA Tomonori Tested-by: Takashi Iwai Signed-off-by: Ingo Molnar --- arch/x86/include/asm/dma-mapping.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/arch/x86/include/asm/dma-mapping.h b/arch/x86/include/asm/dma-mapping.h index 4a5397bfce27..7f225a4b2a26 100644 --- a/arch/x86/include/asm/dma-mapping.h +++ b/arch/x86/include/asm/dma-mapping.h @@ -255,9 +255,11 @@ static inline unsigned long dma_alloc_coherent_mask(struct device *dev, static inline gfp_t dma_alloc_coherent_gfp_flags(struct device *dev, gfp_t gfp) { -#ifdef CONFIG_X86_64 unsigned long dma_mask = dma_alloc_coherent_mask(dev, gfp); + if (dma_mask <= DMA_24BIT_MASK) + gfp |= GFP_DMA; +#ifdef CONFIG_X86_64 if (dma_mask <= DMA_32BIT_MASK && !(gfp & GFP_DMA)) gfp |= GFP_DMA32; #endif From 03967c5267b0e7312d1d55dc814d94cf190ca573 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Thu, 23 Oct 2008 23:14:29 +0900 Subject: [PATCH 08/10] x86: restore the old swiotlb alloc_coherent behavior This restores the old swiotlb alloc_coherent behavior (before the alloc_coherent rewrite): http://lkml.org/lkml/2008/8/12/200 The old alloc_coherent avoids GFP_DMA allocation first and if the allocated address is not fit for the device's coherent_dma_mask, then dma_alloc_coherent does GFP_DMA allocation. If it fails, alloc_coherent calls swiotlb_alloc_coherent (in short, we rarely used swiotlb_alloc_coherent). After the alloc_coherent rewrite, dma_alloc_coherent (include/asm-x86/dma-mapping.h) directly calls swiotlb_alloc_coherent. It means that we possibly can't handle a device having dma_masks > 24bit < 32bits since swiotlb_alloc_coherent doesn't have the above GFP_DMA retry mechanism. This patch fixes x86's swiotlb alloc_coherent to use the GFP_DMA retry mechanism, which dma_generic_alloc_coherent() provides now (pci-nommu.c and GART IOMMU driver also use dma_generic_alloc_coherent). Signed-off-by: FUJITA Tomonori Signed-off-by: Ingo Molnar --- arch/x86/kernel/pci-swiotlb_64.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/pci-swiotlb_64.c b/arch/x86/kernel/pci-swiotlb_64.c index c4ce0332759e..3c539d111abb 100644 --- a/arch/x86/kernel/pci-swiotlb_64.c +++ b/arch/x86/kernel/pci-swiotlb_64.c @@ -18,9 +18,21 @@ swiotlb_map_single_phys(struct device *hwdev, phys_addr_t paddr, size_t size, return swiotlb_map_single(hwdev, phys_to_virt(paddr), size, direction); } +static void *x86_swiotlb_alloc_coherent(struct device *hwdev, size_t size, + dma_addr_t *dma_handle, gfp_t flags) +{ + void *vaddr; + + vaddr = dma_generic_alloc_coherent(hwdev, size, dma_handle, flags); + if (vaddr) + return vaddr; + + return swiotlb_alloc_coherent(hwdev, size, dma_handle, flags); +} + struct dma_mapping_ops swiotlb_dma_ops = { .mapping_error = swiotlb_dma_mapping_error, - .alloc_coherent = swiotlb_alloc_coherent, + .alloc_coherent = x86_swiotlb_alloc_coherent, .free_coherent = swiotlb_free_coherent, .map_single = swiotlb_map_single_phys, .unmap_single = swiotlb_unmap_single, From 46fec7ac40e452a2ea5e63648d98b6bb2b5898f9 Mon Sep 17 00:00:00 2001 From: qinghuang feng Date: Tue, 28 Oct 2008 17:24:28 +0800 Subject: [PATCH 09/10] lockdep: minor fix for debug_show_all_locks() When we failed to get tasklist_lock eventually (count equals 0), we should only print " ignoring it.\n", and not print " locked it.\n" needlessly. Signed-off-by: Qinghuang Feng Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/kernel/lockdep.c b/kernel/lockdep.c index dbda475b13bd..11832acdde77 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -3417,9 +3417,10 @@ void debug_show_all_locks(void) } printk(" ignoring it.\n"); unlock = 0; + } else { + if (count != 10) + printk(KERN_CONT " locked it.\n"); } - if (count != 10) - printk(" locked it.\n"); do_each_thread(g, p) { /* From 6afe40b4dace385d7ba2faf24b352f066f3b71bf Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 28 Oct 2008 11:14:58 +0100 Subject: [PATCH 10/10] lockdep: fix irqs on/off ip tracing Impact: fix lockdep lock-api-caller output when irqsoff tracing is enabled 81d68a96 "ftrace: trace irq disabled critical timings" added wrappers around trace_hardirqs_on/off_caller. However these functions use __builtin_return_address(0) to figure out which function actually disabled or enabled irqs. The result is that we save the ips of trace_hardirqs_on/off instead of the real caller. Not very helpful. However since the patch from Steven the ip already gets passed. So use that and get rid of __builtin_return_address(0) in these two functions. Signed-off-by: Heiko Carstens Signed-off-by: Ingo Molnar --- kernel/lockdep.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 11832acdde77..06e157119d2b 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -2169,12 +2169,11 @@ void early_boot_irqs_on(void) /* * Hardirqs will be enabled: */ -void trace_hardirqs_on_caller(unsigned long a0) +void trace_hardirqs_on_caller(unsigned long ip) { struct task_struct *curr = current; - unsigned long ip; - time_hardirqs_on(CALLER_ADDR0, a0); + time_hardirqs_on(CALLER_ADDR0, ip); if (unlikely(!debug_locks || current->lockdep_recursion)) return; @@ -2188,7 +2187,6 @@ void trace_hardirqs_on_caller(unsigned long a0) } /* we'll do an OFF -> ON transition: */ curr->hardirqs_enabled = 1; - ip = (unsigned long) __builtin_return_address(0); if (DEBUG_LOCKS_WARN_ON(!irqs_disabled())) return; @@ -2224,11 +2222,11 @@ EXPORT_SYMBOL(trace_hardirqs_on); /* * Hardirqs were disabled: */ -void trace_hardirqs_off_caller(unsigned long a0) +void trace_hardirqs_off_caller(unsigned long ip) { struct task_struct *curr = current; - time_hardirqs_off(CALLER_ADDR0, a0); + time_hardirqs_off(CALLER_ADDR0, ip); if (unlikely(!debug_locks || current->lockdep_recursion)) return; @@ -2241,7 +2239,7 @@ void trace_hardirqs_off_caller(unsigned long a0) * We have done an ON -> OFF transition: */ curr->hardirqs_enabled = 0; - curr->hardirq_disable_ip = _RET_IP_; + curr->hardirq_disable_ip = ip; curr->hardirq_disable_event = ++curr->irq_events; debug_atomic_inc(&hardirqs_off_events); } else