diff options
Diffstat (limited to 'drivers/char/drm')
-rw-r--r-- | drivers/char/drm/ati_pcigart.c | 102 | ||||
-rw-r--r-- | drivers/char/drm/drmP.h | 6 | ||||
-rw-r--r-- | drivers/char/drm/drm_fops.c | 7 | ||||
-rw-r--r-- | drivers/char/drm/drm_lock.c | 35 | ||||
-rw-r--r-- | drivers/char/drm/drm_pciids.h | 10 | ||||
-rw-r--r-- | drivers/char/drm/drm_scatter.c | 11 | ||||
-rw-r--r-- | drivers/char/drm/drm_sysfs.c | 2 | ||||
-rw-r--r-- | drivers/char/drm/drm_vm.c | 145 | ||||
-rw-r--r-- | drivers/char/drm/i915_dma.c | 8 | ||||
-rw-r--r-- | drivers/char/drm/i915_drv.c | 52 | ||||
-rw-r--r-- | drivers/char/drm/i915_drv.h | 21 | ||||
-rw-r--r-- | drivers/char/drm/r128_cce.c | 1 | ||||
-rw-r--r-- | drivers/char/drm/r300_cmdbuf.c | 54 | ||||
-rw-r--r-- | drivers/char/drm/radeon_cp.c | 82 | ||||
-rw-r--r-- | drivers/char/drm/radeon_drm.h | 12 | ||||
-rw-r--r-- | drivers/char/drm/radeon_drv.h | 38 | ||||
-rw-r--r-- | drivers/char/drm/radeon_mem.c | 8 | ||||
-rw-r--r-- | drivers/char/drm/via_dma.c | 59 | ||||
-rw-r--r-- | drivers/char/drm/via_dmablit.c | 2 |
19 files changed, 448 insertions, 207 deletions
diff --git a/drivers/char/drm/ati_pcigart.c b/drivers/char/drm/ati_pcigart.c index d352dbb4ccf..141f4dfa0a1 100644 --- a/drivers/char/drm/ati_pcigart.c +++ b/drivers/char/drm/ati_pcigart.c @@ -35,42 +35,23 @@ # define ATI_PCIGART_PAGE_SIZE 4096 /**< PCI GART page size */ -static void *drm_ati_alloc_pcigart_table(int order) +static int drm_ati_alloc_pcigart_table(struct drm_device *dev, + struct drm_ati_pcigart_info *gart_info) { - unsigned long address; - struct page *page; - int i; - - DRM_DEBUG("%d order\n", order); - - address = __get_free_pages(GFP_KERNEL | __GFP_COMP, - order); - if (address == 0UL) { - return NULL; - } - - page = virt_to_page(address); - - for (i = 0; i < order; i++, page++) - SetPageReserved(page); + gart_info->table_handle = drm_pci_alloc(dev, gart_info->table_size, + PAGE_SIZE, + gart_info->table_mask); + if (gart_info->table_handle == NULL) + return -ENOMEM; - DRM_DEBUG("returning 0x%08lx\n", address); - return (void *)address; + return 0; } -static void drm_ati_free_pcigart_table(void *address, int order) +static void drm_ati_free_pcigart_table(struct drm_device *dev, + struct drm_ati_pcigart_info *gart_info) { - struct page *page; - int i; - int num_pages = 1 << order; - DRM_DEBUG("\n"); - - page = virt_to_page((unsigned long)address); - - for (i = 0; i < num_pages; i++, page++) - ClearPageReserved(page); - - free_pages((unsigned long)address, order); + drm_pci_free(dev, gart_info->table_handle); + gart_info->table_handle = NULL; } int drm_ati_pcigart_cleanup(struct drm_device *dev, struct drm_ati_pcigart_info *gart_info) @@ -78,8 +59,7 @@ int drm_ati_pcigart_cleanup(struct drm_device *dev, struct drm_ati_pcigart_info struct drm_sg_mem *entry = dev->sg; unsigned long pages; int i; - int order; - int num_pages, max_pages; + int max_pages; /* we need to support large memory configurations */ if (!entry) { @@ -87,15 +67,7 @@ int drm_ati_pcigart_cleanup(struct drm_device *dev, struct drm_ati_pcigart_info return 0; } - order = drm_order((gart_info->table_size + (PAGE_SIZE-1)) / PAGE_SIZE); - num_pages = 1 << order; - if (gart_info->bus_addr) { - if (gart_info->gart_table_location == DRM_ATI_GART_MAIN) { - pci_unmap_single(dev->pdev, gart_info->bus_addr, - num_pages * PAGE_SIZE, - PCI_DMA_TODEVICE); - } max_pages = (gart_info->table_size / sizeof(u32)); pages = (entry->pages <= max_pages) @@ -112,10 +84,9 @@ int drm_ati_pcigart_cleanup(struct drm_device *dev, struct drm_ati_pcigart_info gart_info->bus_addr = 0; } - if (gart_info->gart_table_location == DRM_ATI_GART_MAIN - && gart_info->addr) { - drm_ati_free_pcigart_table(gart_info->addr, order); - gart_info->addr = NULL; + if (gart_info->gart_table_location == DRM_ATI_GART_MAIN && + gart_info->table_handle) { + drm_ati_free_pcigart_table(dev, gart_info); } return 1; @@ -127,11 +98,10 @@ int drm_ati_pcigart_init(struct drm_device *dev, struct drm_ati_pcigart_info *ga struct drm_sg_mem *entry = dev->sg; void *address = NULL; unsigned long pages; - u32 *pci_gart, page_base, bus_address = 0; + u32 *pci_gart, page_base; + dma_addr_t bus_address = 0; int i, j, ret = 0; - int order; int max_pages; - int num_pages; if (!entry) { DRM_ERROR("no scatter/gather memory!\n"); @@ -141,36 +111,20 @@ int drm_ati_pcigart_init(struct drm_device *dev, struct drm_ati_pcigart_info *ga if (gart_info->gart_table_location == DRM_ATI_GART_MAIN) { DRM_DEBUG("PCI: no table in VRAM: using normal RAM\n"); - order = drm_order((gart_info->table_size + - (PAGE_SIZE-1)) / PAGE_SIZE); - num_pages = 1 << order; - address = drm_ati_alloc_pcigart_table(order); - if (!address) { + ret = drm_ati_alloc_pcigart_table(dev, gart_info); + if (ret) { DRM_ERROR("cannot allocate PCI GART page!\n"); goto done; } - if (!dev->pdev) { - DRM_ERROR("PCI device unknown!\n"); - goto done; - } - - bus_address = pci_map_single(dev->pdev, address, - num_pages * PAGE_SIZE, - PCI_DMA_TODEVICE); - if (bus_address == 0) { - DRM_ERROR("unable to map PCIGART pages!\n"); - order = drm_order((gart_info->table_size + - (PAGE_SIZE-1)) / PAGE_SIZE); - drm_ati_free_pcigart_table(address, order); - address = NULL; - goto done; - } + address = gart_info->table_handle->vaddr; + bus_address = gart_info->table_handle->busaddr; } else { address = gart_info->addr; bus_address = gart_info->bus_addr; - DRM_DEBUG("PCI: Gart Table: VRAM %08X mapped at %08lX\n", - bus_address, (unsigned long)address); + DRM_DEBUG("PCI: Gart Table: VRAM %08LX mapped at %08lX\n", + (unsigned long long)bus_address, + (unsigned long)address); } pci_gart = (u32 *) address; @@ -214,6 +168,12 @@ int drm_ati_pcigart_init(struct drm_device *dev, struct drm_ati_pcigart_info *ga } } + if (gart_info->gart_table_location == DRM_ATI_GART_MAIN) + dma_sync_single_for_device(&dev->pdev->dev, + bus_address, + max_pages * sizeof(u32), + PCI_DMA_TODEVICE); + ret = 1; #if defined(__i386__) || defined(__x86_64__) diff --git a/drivers/char/drm/drmP.h b/drivers/char/drm/drmP.h index 19d3be5c4b2..6540948d517 100644 --- a/drivers/char/drm/drmP.h +++ b/drivers/char/drm/drmP.h @@ -54,6 +54,7 @@ #include <linux/pci.h> #include <linux/jiffies.h> #include <linux/smp_lock.h> /* For (un)lock_kernel */ +#include <linux/dma-mapping.h> #include <linux/mm.h> #include <linux/cdev.h> #include <linux/mutex.h> @@ -551,6 +552,8 @@ struct drm_ati_pcigart_info { int gart_reg_if; void *addr; dma_addr_t bus_addr; + dma_addr_t table_mask; + struct drm_dma_handle *table_handle; drm_local_map_t mapping; int table_size; }; @@ -568,7 +571,7 @@ struct drm_driver { void (*postclose) (struct drm_device *, struct drm_file *); void (*lastclose) (struct drm_device *); int (*unload) (struct drm_device *); - int (*suspend) (struct drm_device *); + int (*suspend) (struct drm_device *, pm_message_t state); int (*resume) (struct drm_device *); int (*dma_ioctl) (struct drm_device *dev, void *data, struct drm_file *file_priv); void (*dma_ready) (struct drm_device *); @@ -637,7 +640,6 @@ struct drm_head { struct drm_device *dev; struct proc_dir_entry *dev_root; /**< proc directory entry */ dev_t device; /**< Device number for mknod */ - struct class_device *dev_class; }; /** diff --git a/drivers/char/drm/drm_fops.c b/drivers/char/drm/drm_fops.c index 3992f73299c..f09d4b5002b 100644 --- a/drivers/char/drm/drm_fops.c +++ b/drivers/char/drm/drm_fops.c @@ -326,6 +326,7 @@ int drm_release(struct inode *inode, struct file *filp) struct drm_file *file_priv = filp->private_data; struct drm_device *dev = file_priv->head->dev; int retcode = 0; + unsigned long irqflags; lock_kernel(); @@ -357,9 +358,11 @@ int drm_release(struct inode *inode, struct file *filp) */ do{ - spin_lock(&dev->lock.spinlock); + spin_lock_irqsave(&dev->lock.spinlock, + irqflags); locked = dev->lock.idle_has_lock; - spin_unlock(&dev->lock.spinlock); + spin_unlock_irqrestore(&dev->lock.spinlock, + irqflags); if (locked) break; schedule(); diff --git a/drivers/char/drm/drm_lock.c b/drivers/char/drm/drm_lock.c index bea2a7d5b2b..12dcdd1832f 100644 --- a/drivers/char/drm/drm_lock.c +++ b/drivers/char/drm/drm_lock.c @@ -53,6 +53,7 @@ int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv) DECLARE_WAITQUEUE(entry, current); struct drm_lock *lock = data; int ret = 0; + unsigned long irqflags; ++file_priv->lock_count; @@ -71,9 +72,9 @@ int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv) return -EINVAL; add_wait_queue(&dev->lock.lock_queue, &entry); - spin_lock(&dev->lock.spinlock); + spin_lock_irqsave(&dev->lock.spinlock, irqflags); dev->lock.user_waiters++; - spin_unlock(&dev->lock.spinlock); + spin_unlock_irqrestore(&dev->lock.spinlock, irqflags); for (;;) { __set_current_state(TASK_INTERRUPTIBLE); if (!dev->lock.hw_lock) { @@ -95,9 +96,9 @@ int drm_lock(struct drm_device *dev, void *data, struct drm_file *file_priv) break; } } - spin_lock(&dev->lock.spinlock); + spin_lock_irqsave(&dev->lock.spinlock, irqflags); dev->lock.user_waiters--; - spin_unlock(&dev->lock.spinlock); + spin_unlock_irqrestore(&dev->lock.spinlock, irqflags); __set_current_state(TASK_RUNNING); remove_wait_queue(&dev->lock.lock_queue, &entry); @@ -198,8 +199,9 @@ int drm_lock_take(struct drm_lock_data *lock_data, { unsigned int old, new, prev; volatile unsigned int *lock = &lock_data->hw_lock->lock; + unsigned long irqflags; - spin_lock(&lock_data->spinlock); + spin_lock_irqsave(&lock_data->spinlock, irqflags); do { old = *lock; if (old & _DRM_LOCK_HELD) @@ -211,7 +213,7 @@ int drm_lock_take(struct drm_lock_data *lock_data, } prev = cmpxchg(lock, old, new); } while (prev != old); - spin_unlock(&lock_data->spinlock); + spin_unlock_irqrestore(&lock_data->spinlock, irqflags); if (_DRM_LOCKING_CONTEXT(old) == context) { if (old & _DRM_LOCK_HELD) { @@ -272,15 +274,16 @@ int drm_lock_free(struct drm_lock_data *lock_data, unsigned int context) { unsigned int old, new, prev; volatile unsigned int *lock = &lock_data->hw_lock->lock; + unsigned long irqflags; - spin_lock(&lock_data->spinlock); + spin_lock_irqsave(&lock_data->spinlock, irqflags); if (lock_data->kernel_waiters != 0) { drm_lock_transfer(lock_data, 0); lock_data->idle_has_lock = 1; - spin_unlock(&lock_data->spinlock); + spin_unlock_irqrestore(&lock_data->spinlock, irqflags); return 1; } - spin_unlock(&lock_data->spinlock); + spin_unlock_irqrestore(&lock_data->spinlock, irqflags); do { old = *lock; @@ -344,19 +347,20 @@ static int drm_notifier(void *priv) void drm_idlelock_take(struct drm_lock_data *lock_data) { int ret = 0; + unsigned long irqflags; - spin_lock(&lock_data->spinlock); + spin_lock_irqsave(&lock_data->spinlock, irqflags); lock_data->kernel_waiters++; if (!lock_data->idle_has_lock) { - spin_unlock(&lock_data->spinlock); + spin_unlock_irqrestore(&lock_data->spinlock, irqflags); ret = drm_lock_take(lock_data, DRM_KERNEL_CONTEXT); - spin_lock(&lock_data->spinlock); + spin_lock_irqsave(&lock_data->spinlock, irqflags); if (ret == 1) lock_data->idle_has_lock = 1; } - spin_unlock(&lock_data->spinlock); + spin_unlock_irqrestore(&lock_data->spinlock, irqflags); } EXPORT_SYMBOL(drm_idlelock_take); @@ -364,8 +368,9 @@ void drm_idlelock_release(struct drm_lock_data *lock_data) { unsigned int old, prev; volatile unsigned int *lock = &lock_data->hw_lock->lock; + unsigned long irqflags; - spin_lock(&lock_data->spinlock); + spin_lock_irqsave(&lock_data->spinlock, irqflags); if (--lock_data->kernel_waiters == 0) { if (lock_data->idle_has_lock) { do { @@ -376,7 +381,7 @@ void drm_idlelock_release(struct drm_lock_data *lock_data) lock_data->idle_has_lock = 0; } } - spin_unlock(&lock_data->spinlock); + spin_unlock_irqrestore(&lock_data->spinlock, irqflags); } EXPORT_SYMBOL(drm_idlelock_release); diff --git a/drivers/char/drm/drm_pciids.h b/drivers/char/drm/drm_pciids.h index f5246884367..a6a499f97e2 100644 --- a/drivers/char/drm/drm_pciids.h +++ b/drivers/char/drm/drm_pciids.h @@ -83,6 +83,7 @@ {0x1002, 0x5460, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV380|RADEON_IS_MOBILITY}, \ {0x1002, 0x5462, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV380|RADEON_IS_MOBILITY}, \ {0x1002, 0x5464, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV380|RADEON_IS_MOBILITY}, \ + {0x1002, 0x5657, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV380|RADEON_NEW_MEMMAP}, \ {0x1002, 0x5548, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_R420|RADEON_NEW_MEMMAP}, \ {0x1002, 0x5549, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_R420|RADEON_NEW_MEMMAP}, \ {0x1002, 0x554A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_R420|RADEON_NEW_MEMMAP}, \ @@ -204,9 +205,9 @@ {0x1002, 0x71D6, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV530|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x71DA, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV530|RADEON_NEW_MEMMAP}, \ {0x1002, 0x71DE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV530|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ - {0x1002, 0x7200, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV530|RADEON_NEW_MEMMAP}, \ - {0x1002, 0x7210, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV530|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ - {0x1002, 0x7211, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV530|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x7200, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV515|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x7210, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV515|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x7211, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV515|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x7240, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_R580|RADEON_NEW_MEMMAP}, \ {0x1002, 0x7243, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_R580|RADEON_NEW_MEMMAP}, \ {0x1002, 0x7244, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_R580|RADEON_NEW_MEMMAP}, \ @@ -236,6 +237,8 @@ {0x1002, 0x7297, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RV560|RADEON_NEW_MEMMAP}, \ {0x1002, 0x7834, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RS300|RADEON_IS_IGP|RADEON_NEW_MEMMAP}, \ {0x1002, 0x7835, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RS300|RADEON_IS_IGP|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x791e, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RS690|RADEON_IS_IGP|RADEON_NEW_MEMMAP|RADEON_IS_IGPGART}, \ + {0x1002, 0x791f, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_RS690|RADEON_IS_IGP|RADEON_NEW_MEMMAP|RADEON_IS_IGPGART}, \ {0, 0, 0} #define r128_PCI_IDS \ @@ -313,6 +316,7 @@ {0x1039, 0x5300, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, \ {0x1039, 0x6300, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, \ {0x1039, 0x6330, PCI_ANY_ID, PCI_ANY_ID, 0, 0, SIS_CHIP_315}, \ + {0x1039, 0x6351, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, \ {0x1039, 0x7300, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0}, \ {0x18CA, 0x0040, PCI_ANY_ID, PCI_ANY_ID, 0, 0, SIS_CHIP_315}, \ {0x18CA, 0x0042, PCI_ANY_ID, PCI_ANY_ID, 0, 0, SIS_CHIP_315}, \ diff --git a/drivers/char/drm/drm_scatter.c b/drivers/char/drm/drm_scatter.c index 26d8f675ed5..b2b0f3d4171 100644 --- a/drivers/char/drm/drm_scatter.c +++ b/drivers/char/drm/drm_scatter.c @@ -36,6 +36,15 @@ #define DEBUG_SCATTER 0 +static inline void *drm_vmalloc_dma(unsigned long size) +{ +#if defined(__powerpc__) && defined(CONFIG_NOT_COHERENT_CACHE) + return __vmalloc(size, GFP_KERNEL, PAGE_KERNEL | _PAGE_NO_CACHE); +#else + return vmalloc_32(size); +#endif +} + void drm_sg_cleanup(struct drm_sg_mem * entry) { struct page *page; @@ -104,7 +113,7 @@ int drm_sg_alloc(struct drm_device *dev, struct drm_scatter_gather * request) } memset((void *)entry->busaddr, 0, pages * sizeof(*entry->busaddr)); - entry->virtual = vmalloc_32(pages << PAGE_SHIFT); + entry->virtual = drm_vmalloc_dma(pages << PAGE_SHIFT); if (!entry->virtual) { drm_free(entry->busaddr, entry->pages * sizeof(*entry->busaddr), DRM_MEM_PAGES); diff --git a/drivers/char/drm/drm_sysfs.c b/drivers/char/drm/drm_sysfs.c index fa36153619e..05ed5043254 100644 --- a/drivers/char/drm/drm_sysfs.c +++ b/drivers/char/drm/drm_sysfs.c @@ -36,7 +36,7 @@ static int drm_sysfs_suspend(struct device *dev, pm_message_t state) printk(KERN_ERR "%s\n", __FUNCTION__); if (drm_dev->driver->suspend) - return drm_dev->driver->suspend(drm_dev); + return drm_dev->driver->suspend(drm_dev, state); return 0; } diff --git a/drivers/char/drm/drm_vm.c b/drivers/char/drm/drm_vm.c index cea4105374b..945df72a51a 100644 --- a/drivers/char/drm/drm_vm.c +++ b/drivers/char/drm/drm_vm.c @@ -54,19 +54,30 @@ static pgprot_t drm_io_prot(uint32_t map_type, struct vm_area_struct *vma) pgprot_val(tmp) |= _PAGE_NO_CACHE; if (map_type == _DRM_REGISTERS) pgprot_val(tmp) |= _PAGE_GUARDED; -#endif -#if defined(__ia64__) +#elif defined(__ia64__) if (efi_range_is_wc(vma->vm_start, vma->vm_end - vma->vm_start)) tmp = pgprot_writecombine(tmp); else tmp = pgprot_noncached(tmp); +#elif defined(__sparc__) + tmp = pgprot_noncached(tmp); +#endif + return tmp; +} + +static pgprot_t drm_dma_prot(uint32_t map_type, struct vm_area_struct *vma) +{ + pgprot_t tmp = vm_get_page_prot(vma->vm_flags); + +#if defined(__powerpc__) && defined(CONFIG_NOT_COHERENT_CACHE) + tmp |= _PAGE_NO_CACHE; #endif return tmp; } /** - * \c nopage method for AGP virtual memory. + * \c fault method for AGP virtual memory. * * \param vma virtual memory area. * \param address access address. @@ -76,8 +87,7 @@ static pgprot_t drm_io_prot(uint32_t map_type, struct vm_area_struct *vma) * map, get the page, increment the use count and return it. */ #if __OS_HAS_AGP -static __inline__ struct page *drm_do_vm_nopage(struct vm_area_struct *vma, - unsigned long address) +static int drm_do_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { struct drm_file *priv = vma->vm_file->private_data; struct drm_device *dev = priv->head->dev; @@ -89,19 +99,24 @@ static __inline__ struct page *drm_do_vm_nopage(struct vm_area_struct *vma, * Find the right map */ if (!drm_core_has_AGP(dev)) - goto vm_nopage_error; + goto vm_fault_error; if (!dev->agp || !dev->agp->cant_use_aperture) - goto vm_nopage_error; + goto vm_fault_error; if (drm_ht_find_item(&dev->map_hash, vma->vm_pgoff, &hash)) - goto vm_nopage_error; + goto vm_fault_error; r_list = drm_hash_entry(hash, struct drm_map_list, hash); map = r_list->map; if (map && map->type == _DRM_AGP) { - unsigned long offset = address - vma->vm_start; + /* + * Using vm_pgoff as a selector forces us to use this unusual + * addressing scheme. + */ + unsigned long offset = (unsigned long)vmf->virtual_address - + vma->vm_start; unsigned long baddr = map->offset + offset; struct drm_agp_mem *agpmem; struct page *page; @@ -123,7 +138,7 @@ static __inline__ struct page *drm_do_vm_nopage(struct vm_area_struct *vma, } if (!agpmem) - goto vm_nopage_error; + goto vm_fault_error; /* * Get the page, inc the use count, and return it @@ -131,22 +146,21 @@ static __inline__ struct page *drm_do_vm_nopage(struct vm_area_struct *vma, offset = (baddr - agpmem->bound) >> PAGE_SHIFT; page = virt_to_page(__va(agpmem->memory->memory[offset])); get_page(page); + vmf->page = page; DRM_DEBUG ("baddr = 0x%lx page = 0x%p, offset = 0x%lx, count=%d\n", baddr, __va(agpmem->memory->memory[offset]), offset, page_count(page)); - - return page; + return 0; } - vm_nopage_error: - return NOPAGE_SIGBUS; /* Disallow mremap */ +vm_fault_error: + return VM_FAULT_SIGBUS; /* Disallow mremap */ } #else /* __OS_HAS_AGP */ -static __inline__ struct page *drm_do_vm_nopage(struct vm_area_struct *vma, - unsigned long address) +static int drm_do_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - return NOPAGE_SIGBUS; + return VM_FAULT_SIGBUS; } #endif /* __OS_HAS_AGP */ @@ -160,28 +174,26 @@ static __inline__ struct page *drm_do_vm_nopage(struct vm_area_struct *vma, * Get the mapping, find the real physical page to map, get the page, and * return it. */ -static __inline__ struct page *drm_do_vm_shm_nopage(struct vm_area_struct *vma, - unsigned long address) +static int drm_do_vm_shm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { struct drm_map *map = (struct drm_map *) vma->vm_private_data; unsigned long offset; unsigned long i; struct page *page; - if (address > vma->vm_end) - return NOPAGE_SIGBUS; /* Disallow mremap */ if (!map) - return NOPAGE_SIGBUS; /* Nothing allocated */ + return VM_FAULT_SIGBUS; /* Nothing allocated */ - offset = address - vma->vm_start; + offset = (unsigned long)vmf->virtual_address - vma->vm_start; i = (unsigned long)map->handle + offset; page = vmalloc_to_page((void *)i); if (!page) - return NOPAGE_SIGBUS; + return VM_FAULT_SIGBUS; get_page(page); + vmf->page = page; - DRM_DEBUG("0x%lx\n", address); - return page; + DRM_DEBUG("shm_fault 0x%lx\n", offset); + return 0; } /** @@ -263,7 +275,7 @@ static void drm_vm_shm_close(struct vm_area_struct *vma) } /** - * \c nopage method for DMA virtual memory. + * \c fault method for DMA virtual memory. * * \param vma virtual memory area. * \param address access address. @@ -271,8 +283,7 @@ static void drm_vm_shm_close(struct vm_area_struct *vma) * * Determine the page number from the page offset and get it from drm_device_dma::pagelist. */ -static __inline__ struct page *drm_do_vm_dma_nopage(struct vm_area_struct *vma, - unsigned long address) +static int drm_do_vm_dma_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { struct drm_file *priv = vma->vm_file->private_data; struct drm_device *dev = priv->head->dev; @@ -282,24 +293,23 @@ static __inline__ struct page *drm_do_vm_dma_nopage(struct vm_area_struct *vma, struct page *page; if (!dma) - return NOPAGE_SIGBUS; /* Error */ - if (address > vma->vm_end) - return NOPAGE_SIGBUS; /* Disallow mremap */ + return VM_FAULT_SIGBUS; /* Error */ if (!dma->pagelist) - return NOPAGE_SIGBUS; /* Nothing allocated */ + return VM_FAULT_SIGBUS; /* Nothing allocated */ - offset = address - vma->vm_start; /* vm_[pg]off[set] should be 0 */ - page_nr = offset >> PAGE_SHIFT; + offset = (unsigned long)vmf->virtual_address - vma->vm_start; /* vm_[pg]off[set] should be 0 */ + page_nr = offset >> PAGE_SHIFT; /* page_nr could just be vmf->pgoff */ page = virt_to_page((dma->pagelist[page_nr] + (offset & (~PAGE_MASK)))); get_page(page); + vmf->page = page; - DRM_DEBUG("0x%lx (page %lu)\n", address, page_nr); - return page; + DRM_DEBUG("dma_fault 0x%lx (page %lu)\n", offset, page_nr); + return 0; } /** - * \c nopage method for scatter-gather virtual memory. + * \c fault method for scatter-gather virtual memory. * * \param vma virtual memory area. * \param address access address. @@ -307,8 +317,7 @@ static __inline__ struct page *drm_do_vm_dma_nopage(struct vm_area_struct *vma, * * Determine the map offset from the page offset and get it from drm_sg_mem::pagelist. */ -static __inline__ struct page *drm_do_vm_sg_nopage(struct vm_area_struct *vma, - unsigned long address) +static int drm_do_vm_sg_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { struct drm_map *map = (struct drm_map *) vma->vm_private_data; struct drm_file *priv = vma->vm_file->private_data; @@ -320,77 +329,64 @@ static __inline__ struct page *drm_do_vm_sg_nopage(struct vm_area_struct *vma, struct page *page; if (!entry) - return NOPAGE_SIGBUS; /* Error */ - if (address > vma->vm_end) - return NOPAGE_SIGBUS; /* Disallow mremap */ + return VM_FAULT_SIGBUS; /* Error */ if (!entry->pagelist) - return NOPAGE_SIGBUS; /* Nothing allocated */ + return VM_FAULT_SIGBUS; /* Nothing allocated */ - offset = address - vma->vm_start; + offset = (unsigned long)vmf->virtual_address - vma->vm_start; map_offset = map->offset - (unsigned long)dev->sg->virtual; page_offset = (offset >> PAGE_SHIFT) + (map_offset >> PAGE_SHIFT); page = entry->pagelist[page_offset]; get_page(page); + vmf->page = page; - return page; + return 0; } -static struct page *drm_vm_nopage(struct vm_area_struct *vma, - unsigned long address, int *type) +static int drm_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - if (type) - *type = VM_FAULT_MINOR; - return drm_do_vm_nopage(vma, address); + return drm_do_vm_fault(vma, vmf); } -static struct page *drm_vm_shm_nopage(struct vm_area_struct *vma, - unsigned long address, int *type) +static int drm_vm_shm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - if (type) - *type = VM_FAULT_MINOR; - return drm_do_vm_shm_nopage(vma, address); + return drm_do_vm_shm_fault(vma, vmf); } -static struct page *drm_vm_dma_nopage(struct vm_area_struct *vma, - unsigned long address, int *type) +static int drm_vm_dma_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - if (type) - *type = VM_FAULT_MINOR; - return drm_do_vm_dma_nopage(vma, address); + return drm_do_vm_dma_fault(vma, vmf); } -static struct page *drm_vm_sg_nopage(struct vm_area_struct *vma, - unsigned long address, int *type) +static int drm_vm_sg_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { - if (type) - *type = VM_FAULT_MINOR; - return drm_do_vm_sg_nopage(vma, address); + return drm_do_vm_sg_fault(vma, vmf); } /** AGP virtual memory operations */ static struct vm_operations_struct drm_vm_ops = { - .nopage = drm_vm_nopage, + .fault = drm_vm_fault, .open = drm_vm_open, .close = drm_vm_close, }; /** Shared virtual memory operations */ static struct vm_operations_struct drm_vm_shm_ops = { - .nopage = drm_vm_shm_nopage, + .fault = drm_vm_shm_fault, .open = drm_vm_open, .close = drm_vm_shm_close, }; /** DMA virtual memory operations */ static struct vm_operations_struct drm_vm_dma_ops = { - .nopage = drm_vm_dma_nopage, + .fault = drm_vm_dma_fault, .open = drm_vm_open, .close = drm_vm_close, }; /** Scatter-gather virtual memory operations */ static struct vm_operations_struct drm_vm_sg_ops = { - .nopage = drm_vm_sg_nopage, + .fault = drm_vm_sg_fault, .open = drm_vm_open, .close = drm_vm_close, }; @@ -604,7 +600,7 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma) /* * On some platforms we can't talk to bus dma address from the CPU, so for * memory of type DRM_AGP, we'll deal with sorting out the real physical - * pages and mappings in nopage() + * pages and mappings in fault() */ #if defined(__powerpc__) pgprot_val(vma->vm_page_prot) |= _PAGE_NO_CACHE; @@ -618,9 +614,6 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma) offset = dev->driver->get_reg_ofs(dev); vma->vm_flags |= VM_IO; /* not in core dump */ vma->vm_page_prot = drm_io_prot(map->type, vma); -#ifdef __sparc__ - vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); -#endif if (io_remap_pfn_range(vma, vma->vm_start, (map->offset + offset) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, @@ -634,11 +627,12 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma) break; case _DRM_CONSISTENT: /* Consistent memory is really like shared memory. But - * it's allocated in a different way, so avoid nopage */ + * it's allocated in a different way, so avoid fault */ if (remap_pfn_range(vma, vma->vm_start, page_to_pfn(virt_to_page(map->handle)), vma->vm_end - vma->vm_start, vma->vm_page_prot)) return -EAGAIN; + vma->vm_page_prot = drm_dma_prot(map->type, vma); /* fall through to _DRM_SHM */ case _DRM_SHM: vma->vm_ops = &drm_vm_shm_ops; @@ -646,6 +640,7 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma) /* Don't let this area swap. Change when DRM_KERNEL advisory is supported. */ vma->vm_flags |= VM_RESERVED; + vma->vm_page_prot = drm_dma_prot(map->type, vma); break; case _DRM_SCATTER_GATHER: vma->vm_ops = &drm_vm_sg_ops; diff --git a/drivers/char/drm/i915_dma.c b/drivers/char/drm/i915_dma.c index 43986d81ae3..a043bb12301 100644 --- a/drivers/char/drm/i915_dma.c +++ b/drivers/char/drm/i915_dma.c @@ -171,7 +171,7 @@ static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init) dev_priv->allow_batchbuffer = 1; /* Program Hardware Status Page */ - if (!IS_G33(dev)) { + if (!I915_NEED_GFX_HWS(dev)) { dev_priv->status_page_dmah = drm_pci_alloc(dev, PAGE_SIZE, PAGE_SIZE, 0xffffffff); @@ -720,6 +720,9 @@ static int i915_set_status_page(struct drm_device *dev, void *data, drm_i915_private_t *dev_priv = dev->dev_private; drm_i915_hws_addr_t *hws = data; + if (!I915_NEED_GFX_HWS(dev)) + return -EINVAL; + if (!dev_priv) { DRM_ERROR("called with no initialization\n"); return -EINVAL; @@ -801,6 +804,9 @@ void i915_driver_lastclose(struct drm_device * dev) { drm_i915_private_t *dev_priv = dev->dev_private; + if (!dev_priv) + return; + if (dev_priv->agp_heap) i915_mem_takedown(&(dev_priv->agp_heap)); diff --git a/drivers/char/drm/i915_drv.c b/drivers/char/drm/i915_drv.c index 52e51033d32..b2b451dc446 100644 --- a/drivers/char/drm/i915_drv.c +++ b/drivers/char/drm/i915_drv.c @@ -160,6 +160,7 @@ static void i915_save_vga(struct drm_device *dev) dev_priv->saveAR[i] = i915_read_ar(st01, i, 0); inb(st01); outb(dev_priv->saveAR_INDEX, VGA_AR_INDEX); + inb(st01); /* Graphics controller registers */ for (i = 0; i < 9; i++) @@ -221,10 +222,12 @@ static void i915_restore_vga(struct drm_device *dev) dev_priv->saveGR[0x18]); /* Attribute controller registers */ + inb(st01); for (i = 0; i < 20; i++) i915_write_ar(st01, i, dev_priv->saveAR[i], 0); inb(st01); /* switch back to index mode */ outb(dev_priv->saveAR_INDEX | 0x20, VGA_AR_INDEX); + inb(st01); /* VGA color palette registers */ outb(dev_priv->saveDACMASK, VGA_DACMASK); @@ -236,7 +239,7 @@ static void i915_restore_vga(struct drm_device *dev) } -static int i915_suspend(struct drm_device *dev) +static int i915_suspend(struct drm_device *dev, pm_message_t state) { struct drm_i915_private *dev_priv = dev->dev_private; int i; @@ -247,6 +250,9 @@ static int i915_suspend(struct drm_device *dev) return -ENODEV; } + if (state.event == PM_EVENT_PRETHAW) + return 0; + pci_save_state(dev->pdev); pci_read_config_byte(dev->pdev, LBB, &dev_priv->saveLBB); @@ -276,6 +282,7 @@ static int i915_suspend(struct drm_device *dev) dev_priv->saveDSPATILEOFF = I915_READ(DSPATILEOFF); } i915_save_palette(dev, PIPE_A); + dev_priv->savePIPEASTAT = I915_READ(I915REG_PIPEASTAT); /* Pipe & plane B info */ dev_priv->savePIPEBCONF = I915_READ(PIPEBCONF); @@ -303,6 +310,7 @@ static int i915_suspend(struct drm_device *dev) dev_priv->saveDSPBTILEOFF = I915_READ(DSPBTILEOFF); } i915_save_palette(dev, PIPE_B); + dev_priv->savePIPEBSTAT = I915_READ(I915REG_PIPEBSTAT); /* CRT state */ dev_priv->saveADPA = I915_READ(ADPA); @@ -329,12 +337,26 @@ static int i915_suspend(struct drm_device *dev) dev_priv->saveFBC_CONTROL2 = I915_READ(FBC_CONTROL2); dev_priv->saveFBC_CONTROL = I915_READ(FBC_CONTROL); + /* Interrupt state */ + dev_priv->saveIIR = I915_READ(I915REG_INT_IDENTITY_R); + dev_priv->saveIER = I915_READ(I915REG_INT_ENABLE_R); + dev_priv->saveIMR = I915_READ(I915REG_INT_MASK_R); + /* VGA state */ dev_priv->saveVCLK_DIVISOR_VGA0 = I915_READ(VCLK_DIVISOR_VGA0); dev_priv->saveVCLK_DIVISOR_VGA1 = I915_READ(VCLK_DIVISOR_VGA1); dev_priv->saveVCLK_POST_DIV = I915_READ(VCLK_POST_DIV); dev_priv->saveVGACNTRL = I915_READ(VGACNTRL); + /* Clock gating state */ + dev_priv->saveDSPCLK_GATE_D = I915_READ(DSPCLK_GATE_D); + + /* Cache mode state */ + dev_priv->saveCACHE_MODE_0 = I915_READ(CACHE_MODE_0); + + /* Memory Arbitration state */ + dev_priv->saveMI_ARB_STATE = I915_READ(MI_ARB_STATE); + /* Scratch space */ for (i = 0; i < 16; i++) { dev_priv->saveSWF0[i] = I915_READ(SWF0 + (i << 2)); @@ -345,9 +367,11 @@ static int i915_suspend(struct drm_device *dev) i915_save_vga(dev); - /* Shut down the device */ - pci_disable_device(dev->pdev); - pci_set_power_state(dev->pdev, PCI_D3hot); + if (state.event == PM_EVENT_SUSPEND) { + /* Shut down the device */ + pci_disable_device(dev->pdev); + pci_set_power_state(dev->pdev, PCI_D3hot); + } return 0; } @@ -400,9 +424,7 @@ static int i915_resume(struct drm_device *dev) I915_WRITE(DSPATILEOFF, dev_priv->saveDSPATILEOFF); } - if ((dev_priv->saveDPLL_A & DPLL_VCO_ENABLE) && - (dev_priv->saveDPLL_A & DPLL_VGA_MODE_DIS)) - I915_WRITE(PIPEACONF, dev_priv->savePIPEACONF); + I915_WRITE(PIPEACONF, dev_priv->savePIPEACONF); i915_restore_palette(dev, PIPE_A); /* Enable the plane */ @@ -444,10 +466,9 @@ static int i915_resume(struct drm_device *dev) I915_WRITE(DSPBTILEOFF, dev_priv->saveDSPBTILEOFF); } - if ((dev_priv->saveDPLL_B & DPLL_VCO_ENABLE) && - (dev_priv->saveDPLL_B & DPLL_VGA_MODE_DIS)) - I915_WRITE(PIPEBCONF, dev_priv->savePIPEBCONF); - i915_restore_palette(dev, PIPE_A); + I915_WRITE(PIPEBCONF, dev_priv->savePIPEBCONF); + + i915_restore_palette(dev, PIPE_B); /* Enable the plane */ I915_WRITE(DSPBCNTR, dev_priv->saveDSPBCNTR); I915_WRITE(DSPBBASE, I915_READ(DSPBBASE)); @@ -485,6 +506,15 @@ static int i915_resume(struct drm_device *dev) I915_WRITE(VCLK_POST_DIV, dev_priv->saveVCLK_POST_DIV); udelay(150); + /* Clock gating state */ + I915_WRITE (DSPCLK_GATE_D, dev_priv->saveDSPCLK_GATE_D); + + /* Cache mode state */ + I915_WRITE (CACHE_MODE_0, dev_priv->saveCACHE_MODE_0 | 0xffff0000); + + /* Memory arbitration state */ + I915_WRITE (MI_ARB_STATE, dev_priv->saveMI_ARB_STATE | 0xffff0000); + for (i = 0; i < 16; i++) { I915_WRITE(SWF0 + (i << 2), dev_priv->saveSWF0[i]); I915_WRITE(SWF10 + (i << 2), dev_priv->saveSWF1[i+7]); diff --git a/drivers/char/drm/i915_drv.h b/drivers/char/drm/i915_drv.h index f8308bfb261..675d88bda06 100644 --- a/drivers/char/drm/i915_drv.h +++ b/drivers/char/drm/i915_drv.h @@ -134,6 +134,7 @@ typedef struct drm_i915_private { u32 saveVBLANK_A; u32 saveVSYNC_A; u32 saveBCLRPAT_A; + u32 savePIPEASTAT; u32 saveDSPASTRIDE; u32 saveDSPASIZE; u32 saveDSPAPOS; @@ -154,6 +155,7 @@ typedef struct drm_i915_private { u32 saveVBLANK_B; u32 saveVSYNC_B; u32 saveBCLRPAT_B; + u32 savePIPEBSTAT; u32 saveDSPBSTRIDE; u32 saveDSPBSIZE; u32 saveDSPBPOS; @@ -182,6 +184,12 @@ typedef struct drm_i915_private { u32 saveFBC_LL_BASE; u32 saveFBC_CONTROL; u32 saveFBC_CONTROL2; + u32 saveIER; + u32 saveIIR; + u32 saveIMR; + u32 saveCACHE_MODE_0; + u32 saveDSPCLK_GATE_D; + u32 saveMI_ARB_STATE; u32 saveSWF0[16]; u32 saveSWF1[16]; u32 saveSWF2[3]; @@ -450,6 +458,10 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); */ #define DMA_FADD_S 0x20d4 +/* Memory Interface Arbitration State + */ +#define MI_ARB_STATE 0x20e4 + /* Cache mode 0 reg. * - Manipulating render cache behaviour is central * to the concept of zone rendering, tuning this reg can help avoid @@ -460,6 +472,7 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); * bit of interest either set or cleared. EG: (BIT<<16) | BIT to set. */ #define Cache_Mode_0 0x2120 +#define CACHE_MODE_0 0x2120 #define CM0_MASK_SHIFT 16 #define CM0_IZ_OPT_DISABLE (1<<6) #define CM0_ZR_OPT_DISABLE (1<<5) @@ -655,6 +668,8 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); /** P1 value is 2 greater than this field */ # define VGA0_PD_P1_MASK (0x1f << 0) +#define DSPCLK_GATE_D 0x6200 + /* I830 CRTC registers */ #define HTOTAL_A 0x60000 #define HBLANK_A 0x60004 @@ -1077,8 +1092,8 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); #define IS_I915G(dev) ((dev)->pci_device == 0x2582 || (dev)->pci_device == 0x258a) #define IS_I915GM(dev) ((dev)->pci_device == 0x2592) #define IS_I945G(dev) ((dev)->pci_device == 0x2772) -#define IS_I945GM(dev) ((dev)->pci_device == 0x27A2) - +#define IS_I945GM(dev) ((dev)->pci_device == 0x27A2 ||\ + (dev)->pci_device == 0x27AE) #define IS_I965G(dev) ((dev)->pci_device == 0x2972 || \ (dev)->pci_device == 0x2982 || \ (dev)->pci_device == 0x2992 || \ @@ -1101,6 +1116,8 @@ extern int i915_wait_ring(struct drm_device * dev, int n, const char *caller); #define IS_MOBILE(dev) (IS_I830(dev) || IS_I85X(dev) || IS_I915GM(dev) || \ IS_I945GM(dev) || IS_I965GM(dev) || IS_IGD_GM(dev)) +#define I915_NEED_GFX_HWS(dev) (IS_G33(dev) || IS_IGD_GM(dev)) + #define PRIMARY_RINGBUFFER_SIZE (128*1024) #endif diff --git a/drivers/char/drm/r128_cce.c b/drivers/char/drm/r128_cce.c index 892e0a58984..f36adbd3aaf 100644 --- a/drivers/char/drm/r128_cce.c +++ b/drivers/char/drm/r128_cce.c @@ -558,6 +558,7 @@ static int r128_do_init_cce(struct drm_device * dev, drm_r128_init_t * init) #if __OS_HAS_AGP if (dev_priv->is_pci) { #endif + dev_priv->gart_info.table_mask = DMA_BIT_MASK(32); dev_priv->gart_info.gart_table_location = DRM_ATI_GART_MAIN; dev_priv->gart_info.table_size = R128_PCIGART_TABLE_SIZE; dev_priv->gart_info.addr = NULL; diff --git a/drivers/char/drm/r300_cmdbuf.c b/drivers/char/drm/r300_cmdbuf.c index 0f4afc44245..f535812e405 100644 --- a/drivers/char/drm/r300_cmdbuf.c +++ b/drivers/char/drm/r300_cmdbuf.c @@ -729,6 +729,47 @@ static void r300_discard_buffer(struct drm_device * dev, struct drm_buf * buf) buf->used = 0; } +static void r300_cmd_wait(drm_radeon_private_t * dev_priv, + drm_r300_cmd_header_t header) +{ + u32 wait_until; + RING_LOCALS; + + if (!header.wait.flags) + return; + + wait_until = 0; + + switch(header.wait.flags) { + case R300_WAIT_2D: + wait_until = RADEON_WAIT_2D_IDLE; + break; + case R300_WAIT_3D: + wait_until = RADEON_WAIT_3D_IDLE; + break; + case R300_NEW_WAIT_2D_3D: + wait_until = RADEON_WAIT_2D_IDLE|RADEON_WAIT_3D_IDLE; + break; + case R300_NEW_WAIT_2D_2D_CLEAN: + wait_until = RADEON_WAIT_2D_IDLE|RADEON_WAIT_2D_IDLECLEAN; + break; + case R300_NEW_WAIT_3D_3D_CLEAN: + wait_until = RADEON_WAIT_3D_IDLE|RADEON_WAIT_3D_IDLECLEAN; + break; + case R300_NEW_WAIT_2D_2D_CLEAN_3D_3D_CLEAN: + wait_until = RADEON_WAIT_2D_IDLE|RADEON_WAIT_2D_IDLECLEAN; + wait_until |= RADEON_WAIT_3D_IDLE|RADEON_WAIT_3D_IDLECLEAN; + break; + default: + return; + } + + BEGIN_RING(2); + OUT_RING(CP_PACKET0(RADEON_WAIT_UNTIL, 0)); + OUT_RING(wait_until); + ADVANCE_RING(); +} + static int r300_scratch(drm_radeon_private_t *dev_priv, drm_radeon_kcmd_buffer_t *cmdbuf, drm_r300_cmd_header_t header) @@ -909,19 +950,8 @@ int r300_do_cp_cmdbuf(struct drm_device *dev, break; case R300_CMD_WAIT: - /* simple enough, we can do it here */ DRM_DEBUG("R300_CMD_WAIT\n"); - if (header.wait.flags == 0) - break; /* nothing to do */ - - { - RING_LOCALS; - - BEGIN_RING(2); - OUT_RING(CP_PACKET0(RADEON_WAIT_UNTIL, 0)); - OUT_RING((header.wait.flags & 0xf) << 14); - ADVANCE_RING(); - } + r300_cmd_wait(dev_priv, header); break; case R300_CMD_SCRATCH: diff --git a/drivers/char/drm/radeon_cp.c b/drivers/char/drm/radeon_cp.c index 5dc799ab86b..9072e4a1894 100644 --- a/drivers/char/drm/radeon_cp.c +++ b/drivers/char/drm/radeon_cp.c @@ -825,11 +825,19 @@ static u32 RADEON_READ_MCIND(drm_radeon_private_t *dev_priv, int addr) return ret; } +static u32 RS690_READ_MCIND(drm_radeon_private_t *dev_priv, int addr) +{ + RADEON_WRITE(RS690_MC_INDEX, (addr & RS690_MC_INDEX_MASK)); + return RADEON_READ(RS690_MC_DATA); +} + u32 radeon_read_fb_location(drm_radeon_private_t *dev_priv) { if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV515) return RADEON_READ_MCIND(dev_priv, RV515_MC_FB_LOCATION); + else if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RS690) + return RS690_READ_MCIND(dev_priv, RS690_MC_FB_LOCATION); else if ((dev_priv->flags & RADEON_FAMILY_MASK) > CHIP_RV515) return RADEON_READ_MCIND(dev_priv, R520_MC_FB_LOCATION); else @@ -840,6 +848,8 @@ static void radeon_write_fb_location(drm_radeon_private_t *dev_priv, u32 fb_loc) { if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV515) RADEON_WRITE_MCIND(RV515_MC_FB_LOCATION, fb_loc); + else if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RS690) + RS690_WRITE_MCIND(RS690_MC_FB_LOCATION, fb_loc); else if ((dev_priv->flags & RADEON_FAMILY_MASK) > CHIP_RV515) RADEON_WRITE_MCIND(R520_MC_FB_LOCATION, fb_loc); else @@ -850,6 +860,8 @@ static void radeon_write_agp_location(drm_radeon_private_t *dev_priv, u32 agp_lo { if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RV515) RADEON_WRITE_MCIND(RV515_MC_AGP_LOCATION, agp_loc); + else if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RS690) + RS690_WRITE_MCIND(RS690_MC_AGP_LOCATION, agp_loc); else if ((dev_priv->flags & RADEON_FAMILY_MASK) > CHIP_RV515) RADEON_WRITE_MCIND(R520_MC_AGP_LOCATION, agp_loc); else @@ -1362,6 +1374,70 @@ static void radeon_set_igpgart(drm_radeon_private_t * dev_priv, int on) } } +/* Enable or disable RS690 GART on the chip */ +static void radeon_set_rs690gart(drm_radeon_private_t *dev_priv, int on) +{ + u32 temp; + + if (on) { + DRM_DEBUG("programming rs690 gart %08X %08lX %08X\n", + dev_priv->gart_vm_start, + (long)dev_priv->gart_info.bus_addr, + dev_priv->gart_size); + + temp = RS690_READ_MCIND(dev_priv, RS690_MC_MISC_CNTL); + RS690_WRITE_MCIND(RS690_MC_MISC_CNTL, 0x5000); + + RS690_WRITE_MCIND(RS690_MC_AGP_SIZE, + RS690_MC_GART_EN | RS690_MC_AGP_SIZE_32MB); + + temp = RS690_READ_MCIND(dev_priv, RS690_MC_GART_FEATURE_ID); + RS690_WRITE_MCIND(RS690_MC_GART_FEATURE_ID, 0x42040800); + + RS690_WRITE_MCIND(RS690_MC_GART_BASE, + dev_priv->gart_info.bus_addr); + + temp = RS690_READ_MCIND(dev_priv, RS690_MC_AGP_MODE_CONTROL); + RS690_WRITE_MCIND(RS690_MC_AGP_MODE_CONTROL, 0x01400000); + + RS690_WRITE_MCIND(RS690_MC_AGP_BASE, + (unsigned int)dev_priv->gart_vm_start); + + dev_priv->gart_size = 32*1024*1024; + temp = (((dev_priv->gart_vm_start - 1 + dev_priv->gart_size) & + 0xffff0000) | (dev_priv->gart_vm_start >> 16)); + + RS690_WRITE_MCIND(RS690_MC_AGP_LOCATION, temp); + + temp = RS690_READ_MCIND(dev_priv, RS690_MC_AGP_SIZE); + RS690_WRITE_MCIND(RS690_MC_AGP_SIZE, + RS690_MC_GART_EN | RS690_MC_AGP_SIZE_32MB); + + do { + temp = RS690_READ_MCIND(dev_priv, RS690_MC_GART_CACHE_CNTL); + if ((temp & RS690_MC_GART_CLEAR_STATUS) == + RS690_MC_GART_CLEAR_DONE) + break; + DRM_UDELAY(1); + } while (1); + + RS690_WRITE_MCIND(RS690_MC_GART_CACHE_CNTL, + RS690_MC_GART_CC_CLEAR); + do { + temp = RS690_READ_MCIND(dev_priv, RS690_MC_GART_CACHE_CNTL); + if ((temp & RS690_MC_GART_CLEAR_STATUS) == + RS690_MC_GART_CLEAR_DONE) + break; + DRM_UDELAY(1); + } while (1); + + RS690_WRITE_MCIND(RS690_MC_GART_CACHE_CNTL, + RS690_MC_GART_CC_NO_CHANGE); + } else { + RS690_WRITE_MCIND(RS690_MC_AGP_SIZE, RS690_MC_GART_DIS); + } +} + static void radeon_set_pciegart(drm_radeon_private_t * dev_priv, int on) { u32 tmp = RADEON_READ_PCIE(dev_priv, RADEON_PCIE_TX_GART_CNTL); @@ -1396,6 +1472,11 @@ static void radeon_set_pcigart(drm_radeon_private_t * dev_priv, int on) { u32 tmp; + if ((dev_priv->flags & RADEON_FAMILY_MASK) == CHIP_RS690) { + radeon_set_rs690gart(dev_priv, on); + return; + } + if (dev_priv->flags & RADEON_IS_IGPGART) { radeon_set_igpgart(dev_priv, on); return; @@ -1726,6 +1807,7 @@ static int radeon_do_init_cp(struct drm_device * dev, drm_radeon_init_t * init) } else #endif { + dev_priv->gart_info.table_mask = DMA_BIT_MASK(32); /* if we have an offset set from userspace */ if (dev_priv->pcigart_offset_set) { dev_priv->gart_info.bus_addr = diff --git a/drivers/char/drm/radeon_drm.h b/drivers/char/drm/radeon_drm.h index 71e5b21fad2..aab82e121e0 100644 --- a/drivers/char/drm/radeon_drm.h +++ b/drivers/char/drm/radeon_drm.h @@ -225,8 +225,20 @@ typedef union { #define R300_CMD_WAIT 7 # define R300_WAIT_2D 0x1 # define R300_WAIT_3D 0x2 +/* these two defines are DOING IT WRONG - however + * we have userspace which relies on using these. + * The wait interface is backwards compat new + * code should use the NEW_WAIT defines below + * THESE ARE NOT BIT FIELDS + */ # define R300_WAIT_2D_CLEAN 0x3 # define R300_WAIT_3D_CLEAN 0x4 + +# define R300_NEW_WAIT_2D_3D 0x3 +# define R300_NEW_WAIT_2D_2D_CLEAN 0x4 +# define R300_NEW_WAIT_3D_3D_CLEAN 0x6 +# define R300_NEW_WAIT_2D_2D_CLEAN_3D_3D_CLEAN 0x8 + #define R300_CMD_SCRATCH 8 typedef union { diff --git a/drivers/char/drm/radeon_drv.h b/drivers/char/drm/radeon_drv.h index 4434332c79b..173ae620223 100644 --- a/drivers/char/drm/radeon_drv.h +++ b/drivers/char/drm/radeon_drv.h @@ -123,6 +123,7 @@ enum radeon_family { CHIP_R420, CHIP_RV410, CHIP_RS400, + CHIP_RS690, CHIP_RV515, CHIP_R520, CHIP_RV530, @@ -467,6 +468,36 @@ extern int r300_do_cp_cmdbuf(struct drm_device * dev, #define RADEON_IGPGART_ENABLE 0x38 #define RADEON_IGPGART_UNK_39 0x39 +#define RS690_MC_INDEX 0x78 +# define RS690_MC_INDEX_MASK 0x1ff +# define RS690_MC_INDEX_WR_EN (1 << 9) +# define RS690_MC_INDEX_WR_ACK 0x7f +#define RS690_MC_DATA 0x7c + +#define RS690_MC_MISC_CNTL 0x18 +#define RS690_MC_GART_FEATURE_ID 0x2b +#define RS690_MC_GART_BASE 0x2c +#define RS690_MC_GART_CACHE_CNTL 0x2e +# define RS690_MC_GART_CC_NO_CHANGE 0x0 +# define RS690_MC_GART_CC_CLEAR 0x1 +# define RS690_MC_GART_CLEAR_STATUS (1 << 1) +# define RS690_MC_GART_CLEAR_DONE (0 << 1) +# define RS690_MC_GART_CLEAR_PENDING (1 << 1) +#define RS690_MC_AGP_SIZE 0x38 +# define RS690_MC_GART_DIS 0x0 +# define RS690_MC_GART_EN 0x1 +# define RS690_MC_AGP_SIZE_32MB (0 << 1) +# define RS690_MC_AGP_SIZE_64MB (1 << 1) +# define RS690_MC_AGP_SIZE_128MB (2 << 1) +# define RS690_MC_AGP_SIZE_256MB (3 << 1) +# define RS690_MC_AGP_SIZE_512MB (4 << 1) +# define RS690_MC_AGP_SIZE_1GB (5 << 1) +# define RS690_MC_AGP_SIZE_2GB (6 << 1) +#define RS690_MC_AGP_MODE_CONTROL 0x39 +#define RS690_MC_FB_LOCATION 0x100 +#define RS690_MC_AGP_LOCATION 0x101 +#define RS690_MC_AGP_BASE 0x102 + #define R520_MC_IND_INDEX 0x70 #define R520_MC_IND_WR_EN (1<<24) #define R520_MC_IND_DATA 0x74 @@ -1076,6 +1107,13 @@ do { \ RADEON_WRITE(R520_MC_IND_INDEX, 0); \ } while (0) +#define RS690_WRITE_MCIND( addr, val ) \ +do { \ + RADEON_WRITE(RS690_MC_INDEX, RS690_MC_INDEX_WR_EN | ((addr) & RS690_MC_INDEX_MASK)); \ + RADEON_WRITE(RS690_MC_DATA, val); \ + RADEON_WRITE(RS690_MC_INDEX, RS690_MC_INDEX_WR_ACK); \ +} while (0) + #define CP_PACKET0( reg, n ) \ (RADEON_CP_PACKET0 | ((n) << 16) | ((reg) >> 2)) #define CP_PACKET0_TABLE( reg, n ) \ diff --git a/drivers/char/drm/radeon_mem.c b/drivers/char/drm/radeon_mem.c index 78b34fa7c89..4af5286a36f 100644 --- a/drivers/char/drm/radeon_mem.c +++ b/drivers/char/drm/radeon_mem.c @@ -88,7 +88,7 @@ static struct mem_block *alloc_block(struct mem_block *heap, int size, list_for_each(p, heap) { int start = (p->start + mask) & ~mask; - if (p->file_priv == 0 && start + size <= p->start + p->size) + if (p->file_priv == NULL && start + size <= p->start + p->size) return split_block(p, start, size, file_priv); } @@ -113,7 +113,7 @@ static void free_block(struct mem_block *p) /* Assumes a single contiguous range. Needs a special file_priv in * 'heap' to stop it being subsumed. */ - if (p->next->file_priv == 0) { + if (p->next->file_priv == NULL) { struct mem_block *q = p->next; p->size += q->size; p->next = q->next; @@ -121,7 +121,7 @@ static void free_block(struct mem_block *p) drm_free(q, sizeof(*q), DRM_MEM_BUFS); } - if (p->prev->file_priv == 0) { + if (p->prev->file_priv == NULL) { struct mem_block *q = p->prev; q->size += p->size; q->next = p->next; @@ -174,7 +174,7 @@ void radeon_mem_release(struct drm_file *file_priv, struct mem_block *heap) * 'heap' to stop it being subsumed. */ list_for_each(p, heap) { - while (p->file_priv == 0 && p->next->file_priv == 0) { + while (p->file_priv == NULL && p->next->file_priv == NULL) { struct mem_block *q = p->next; p->size += q->size; p->next = q->next; diff --git a/drivers/char/drm/via_dma.c b/drivers/char/drm/via_dma.c index 94baec692b5..7a339dba6a6 100644 --- a/drivers/char/drm/via_dma.c +++ b/drivers/char/drm/via_dma.c @@ -126,6 +126,8 @@ via_cmdbuf_wait(drm_via_private_t * dev_priv, unsigned int size) hw_addr, cur_addr, next_addr); return -1; } + if ((cur_addr < hw_addr) && (next_addr >= hw_addr)) + msleep(1); } while ((cur_addr < hw_addr) && (next_addr >= hw_addr)); return 0; } @@ -416,27 +418,50 @@ static int via_hook_segment(drm_via_private_t * dev_priv, int paused, count; volatile uint32_t *paused_at = dev_priv->last_pause_ptr; uint32_t reader,ptr; + uint32_t diff; paused = 0; via_flush_write_combine(); (void) *(volatile uint32_t *)(via_get_dma(dev_priv) -1); + *paused_at = pause_addr_lo; via_flush_write_combine(); (void) *paused_at; + reader = *(dev_priv->hw_addr_ptr); ptr = ((volatile char *)paused_at - dev_priv->dma_ptr) + dev_priv->dma_offset + (uint32_t) dev_priv->agpAddr + 4; + dev_priv->last_pause_ptr = via_get_dma(dev_priv) - 1; - if ((ptr - reader) <= dev_priv->dma_diff ) { - count = 10000000; - while (!(paused = (VIA_READ(0x41c) & 0x80000000)) && count--); + /* + * If there is a possibility that the command reader will + * miss the new pause address and pause on the old one, + * In that case we need to program the new start address + * using PCI. + */ + + diff = (uint32_t) (ptr - reader) - dev_priv->dma_diff; + count = 10000000; + while(diff == 0 && count--) { + paused = (VIA_READ(0x41c) & 0x80000000); + if (paused) + break; + reader = *(dev_priv->hw_addr_ptr); + diff = (uint32_t) (ptr - reader) - dev_priv->dma_diff; } + paused = VIA_READ(0x41c) & 0x80000000; + if (paused && !no_pci_fire) { reader = *(dev_priv->hw_addr_ptr); - if ((ptr - reader) == dev_priv->dma_diff) { - + diff = (uint32_t) (ptr - reader) - dev_priv->dma_diff; + diff &= (dev_priv->dma_high - 1); + if (diff != 0 && diff < (dev_priv->dma_high >> 1)) { + DRM_ERROR("Paused at incorrect address. " + "0x%08x, 0x%08x 0x%08x\n", + ptr, reader, dev_priv->dma_diff); + } else if (diff == 0) { /* * There is a concern that these writes may stall the PCI bus * if the GPU is not idle. However, idling the GPU first @@ -577,6 +602,7 @@ static void via_cmdbuf_jump(drm_via_private_t * dev_priv) uint32_t pause_addr_lo, pause_addr_hi; uint32_t jump_addr_lo, jump_addr_hi; volatile uint32_t *last_pause_ptr; + uint32_t dma_low_save1, dma_low_save2; agp_base = dev_priv->dma_offset + (uint32_t) dev_priv->agpAddr; via_align_cmd(dev_priv, HC_HAGPBpID_JUMP, 0, &jump_addr_hi, @@ -603,8 +629,29 @@ static void via_cmdbuf_jump(drm_via_private_t * dev_priv) &pause_addr_lo, 0); *last_pause_ptr = pause_addr_lo; + dma_low_save1 = dev_priv->dma_low; - via_hook_segment( dev_priv, jump_addr_hi, jump_addr_lo, 0); + /* + * Now, set a trap that will pause the regulator if it tries to rerun the old + * command buffer. (Which may happen if via_hook_segment detecs a command regulator pause + * and reissues the jump command over PCI, while the regulator has already taken the jump + * and actually paused at the current buffer end). + * There appears to be no other way to detect this condition, since the hw_addr_pointer + * does not seem to get updated immediately when a jump occurs. + */ + + last_pause_ptr = + via_align_cmd(dev_priv, HC_HAGPBpID_PAUSE, 0, &pause_addr_hi, + &pause_addr_lo, 0) - 1; + via_align_cmd(dev_priv, HC_HAGPBpID_PAUSE, 0, &pause_addr_hi, + &pause_addr_lo, 0); + *last_pause_ptr = pause_addr_lo; + + dma_low_save2 = dev_priv->dma_low; + dev_priv->dma_low = dma_low_save1; + via_hook_segment(dev_priv, jump_addr_hi, jump_addr_lo, 0); + dev_priv->dma_low = dma_low_save2; + via_hook_segment(dev_priv, pause_addr_hi, pause_addr_lo, 0); } diff --git a/drivers/char/drm/via_dmablit.c b/drivers/char/drm/via_dmablit.c index 33c5197b73c..409e00afdd0 100644 --- a/drivers/char/drm/via_dmablit.c +++ b/drivers/char/drm/via_dmablit.c @@ -603,7 +603,7 @@ via_build_sg_info(struct drm_device *dev, drm_via_sg_info_t *vsg, drm_via_dmabli * (Not a big limitation anyway.) */ - if ((xfer->mem_stride - xfer->line_length) >= PAGE_SIZE) { + if ((xfer->mem_stride - xfer->line_length) > 2*PAGE_SIZE) { DRM_ERROR("Too large system memory stride. Stride: %d, " "Length: %d\n", xfer->mem_stride, xfer->line_length); return -EINVAL; |