iommu/amd: use full 64-bit value in build_completion_wait()

We started using a 64 bit completion value.  Unfortunately, we only
stored the low 32-bits, so a very large completion value would never
be matched in iommu_completion_wait().

Fixes: ef44ccbb9829 ("iommu/amd: Use 4K page for completion wait write-back semaphore")
Signed-off-by: John Sperbeck <jsperbeck@google.com>
Link: https://lore.kernel.org/r/20220801192229.3358786-1-jsperbeck@google.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>

diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c
index 65b8e4fd82177872a81b706b538014f1be75a8b7..828672a46a3d4fbcd8f23b680cb59d5f90ff284d 100644 (file)
--- a/drivers/iommu/amd/iommu.c
+++ b/drivers/iommu/amd/iommu.c
@@ -939,7 +939,8 @@ static void build_completion_wait(struct iommu_cmd *cmd,
        memset(cmd, 0, sizeof(*cmd));
        cmd->data[0] = lower_32_bits(paddr) | CMD_COMPL_WAIT_STORE_MASK;
        cmd->data[1] = upper_32_bits(paddr);
-       cmd->data[2] = data;
+       cmd->data[2] = lower_32_bits(data);
+       cmd->data[3] = upper_32_bits(data);
        CMD_SET_TYPE(cmd, CMD_COMPL_WAIT);
 }