Merge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...

[linux-2.6] / arch / x86 / kernel / amd_iommu.c

diff --git a/arch/x86/kernel/amd_iommu.c b/arch/x86/kernel/amd_iommu.c
index c25210e6ac888e94224b460e6eb82f3c556d7616..69b4d060b21c4c32e8e40b45c9dc2171c1b56505 100644 (file)
--- a/arch/x86/kernel/amd_iommu.c
+++ b/arch/x86/kernel/amd_iommu.c
@@ -29,9 +29,6 @@
 
 #define CMD_SET_TYPE(cmd, t) ((cmd)->data[1] |= ((t) << 28))
 
-#define to_pages(addr, size) \
-        (round_up(((addr) & ~PAGE_MASK) + (size), PAGE_SIZE) >> PAGE_SHIFT)
-
 #define EXIT_LOOP_COUNT 10000000
 
 static DEFINE_RWLOCK(amd_iommu_devtable_lock);
@@ -68,7 +65,7 @@ static int __iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
        u8 *target;
 
        tail = readl(iommu->mmio_base + MMIO_CMD_TAIL_OFFSET);
-       target = (iommu->cmd_buf + tail);
+       target = iommu->cmd_buf + tail;
        memcpy_toio(target, cmd, sizeof(*cmd));
        tail = (tail + sizeof(*cmd)) % iommu->cmd_buf_size;
        head = readl(iommu->mmio_base + MMIO_CMD_HEAD_OFFSET);
@@ -104,16 +101,13 @@ static int iommu_queue_command(struct amd_iommu *iommu, struct iommu_cmd *cmd)
  */
 static int iommu_completion_wait(struct amd_iommu *iommu)
 {
-       int ret;
+       int ret, ready = 0;
+       unsigned status = 0;
        struct iommu_cmd cmd;
-       volatile u64 ready = 0;
-       unsigned long ready_phys = virt_to_phys(&ready);
        unsigned long i = 0;
 
        memset(&cmd, 0, sizeof(cmd));
-       cmd.data[0] = LOW_U32(ready_phys) | CMD_COMPL_WAIT_STORE_MASK;
-       cmd.data[1] = upper_32_bits(ready_phys);
-       cmd.data[2] = 1; /* value written to 'ready' */
+       cmd.data[0] = CMD_COMPL_WAIT_INT_MASK;
        CMD_SET_TYPE(&cmd, CMD_COMPL_WAIT);
 
        iommu->need_sync = 0;
@@ -125,9 +119,15 @@ static int iommu_completion_wait(struct amd_iommu *iommu)
 
        while (!ready && (i < EXIT_LOOP_COUNT)) {
                ++i;
-               cpu_relax();
+               /* wait for the bit to become one */
+               status = readl(iommu->mmio_base + MMIO_STATUS_OFFSET);
+               ready = status & MMIO_STATUS_COM_WAIT_INT_MASK;
        }
 
+       /* set bit back to zero */
+       status &= ~MMIO_STATUS_COM_WAIT_INT_MASK;
+       writel(status, iommu->mmio_base + MMIO_STATUS_OFFSET);
+
        if (unlikely((i == EXIT_LOOP_COUNT) && printk_ratelimit()))
                printk(KERN_WARNING "AMD IOMMU: Completion wait loop failed\n");
 
@@ -164,7 +164,7 @@ static int iommu_queue_inv_iommu_pages(struct amd_iommu *iommu,
        address &= PAGE_MASK;
        CMD_SET_TYPE(&cmd, CMD_INV_IOMMU_PAGES);
        cmd.data[1] |= domid;
-       cmd.data[2] = LOW_U32(address);
+       cmd.data[2] = lower_32_bits(address);
        cmd.data[3] = upper_32_bits(address);
        if (s) /* size bit - we flush more than one 4kb page */
                cmd.data[2] |= CMD_INV_IOMMU_PAGES_SIZE_MASK;
@@ -185,7 +185,7 @@ static int iommu_flush_pages(struct amd_iommu *iommu, u16 domid,
                u64 address, size_t size)
 {
        int s = 0;
-       unsigned pages = to_pages(address, size);
+       unsigned pages = iommu_num_pages(address, size);
 
        address &= PAGE_MASK;
 
@@ -557,8 +557,8 @@ static struct dma_ops_domain *dma_ops_domain_alloc(struct amd_iommu *iommu,
        if (iommu->exclusion_start &&
            iommu->exclusion_start < dma_dom->aperture_size) {
                unsigned long startpage = iommu->exclusion_start >> PAGE_SHIFT;
-               int pages = to_pages(iommu->exclusion_start,
-                               iommu->exclusion_length);
+               int pages = iommu_num_pages(iommu->exclusion_start,
+                                           iommu->exclusion_length);
                dma_ops_reserve_addresses(dma_dom, startpage, pages);
        }
 
@@ -667,7 +667,7 @@ static int get_device_resources(struct device *dev,
        _bdf = calc_devid(pcidev->bus->number, pcidev->devfn);
 
        /* device not translated by any IOMMU in the system? */
-       if (_bdf >= amd_iommu_last_bdf) {
+       if (_bdf > amd_iommu_last_bdf) {
                *iommu = NULL;
                *domain = NULL;
                *bdf = 0xffff;
@@ -767,7 +767,7 @@ static dma_addr_t __map_single(struct device *dev,
        unsigned int pages;
        int i;
 
-       pages = to_pages(paddr, size);
+       pages = iommu_num_pages(paddr, size);
        paddr &= PAGE_MASK;
 
        address = dma_ops_alloc_addresses(dev, dma_dom, pages);
@@ -802,7 +802,7 @@ static void __unmap_single(struct amd_iommu *iommu,
        if ((dma_addr == 0) || (dma_addr + size > dma_dom->aperture_size))
                return;
 
-       pages = to_pages(dma_addr, size);
+       pages = iommu_num_pages(dma_addr, size);
        dma_addr &= PAGE_MASK;
        start = dma_addr;
 
@@ -1085,7 +1085,7 @@ void prealloc_protection_domains(void)
 
        while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) {
                devid = (dev->bus->number << 8) | dev->devfn;
-               if (devid >= amd_iommu_last_bdf)
+               if (devid > amd_iommu_last_bdf)
                        continue;
                devid = amd_iommu_alias_table[devid];
                if (domain_for_device(devid))