[PATCH 4.9 121/223] intel_th: msu: Fix single mode with disabled IOMMU

[Greg Kroah-Hartman]

From: Alexander Shishkin <alexander.shishkin@xxxxxxxxxxxxxxx>

commit 918b8646497b5dba6ae82d4a7325f01b258972b9 upstream.

Commit 4e0eaf239fb3 ("intel_th: msu: Fix single mode with IOMMU") switched
the single mode code to use dma mapping pages obtained from the page
allocator, but with IOMMU disabled, that may lead to using SWIOTLB bounce
buffers and without additional sync'ing, produces empty trace buffers.

Fix this by using a DMA32 GFP flag to the page allocation in single mode,
as the device supports full 32-bit DMA addressing.

Signed-off-by: Alexander Shishkin <alexander.shishkin@xxxxxxxxxxxxxxx>
Fixes: 4e0eaf239fb3 ("intel_th: msu: Fix single mode with IOMMU")
Reviewed-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
Reported-by: Ammy Yi <ammy.yi@xxxxxxxxx>
Cc: stable <stable@xxxxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20190621161930.60785-4-alexander.shishkin@xxxxxxxxxxxxxxx
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
 drivers/hwtracing/intel_th/msu.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/drivers/hwtracing/intel_th/msu.c
+++ b/drivers/hwtracing/intel_th/msu.c
@@ -638,7 +638,7 @@ static int msc_buffer_contig_alloc(struc
 		goto err_out;
 
 	ret = -ENOMEM;
-	page = alloc_pages(GFP_KERNEL | __GFP_ZERO, order);
+	page = alloc_pages(GFP_KERNEL | __GFP_ZERO | GFP_DMA32, order);
 	if (!page)
 		goto err_free_sgt;