agp/intel-gtt: Don't leak the scratch page
authorDaniel Vetter <daniel.vetter@ffwll.ch>
Wed, 27 Jan 2016 13:37:58 +0000 (14:37 +0100)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Wed, 10 Feb 2016 07:52:08 +0000 (08:52 +0100)
Recently discovered by enabling CONFIG_DMA_API_DEBUG in our CI. By the
looks of it broken since forever.

v2: Don't forget to set the scratch page back to wb (Chris). Reuse
intel_gtt_teardown_scratch_page for that (and fix it up to treat
needs_dmar y/n correctly).

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93793
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1453901881-26425-1-git-send-email-daniel.vetter@ffwll.ch
drivers/char/agp/intel-gtt.c

index 1341a94cc7793aa0425eb83c4e446393be7f60c6..e657f989745e13bfb703989ef06a73fa767acfe2 100644 (file)
@@ -555,8 +555,10 @@ static unsigned int intel_gtt_mappable_entries(void)
 static void intel_gtt_teardown_scratch_page(void)
 {
        set_pages_wb(intel_private.scratch_page, 1);
-       pci_unmap_page(intel_private.pcidev, intel_private.scratch_page_dma,
-                      PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+       if (intel_private.needs_dmar)
+               pci_unmap_page(intel_private.pcidev,
+                              intel_private.scratch_page_dma,
+                              PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
        __free_page(intel_private.scratch_page);
 }
 
@@ -1430,6 +1432,8 @@ void intel_gmch_remove(void)
        if (--intel_private.refcount)
                return;
 
+       if (intel_private.scratch_page)
+               intel_gtt_teardown_scratch_page();
        if (intel_private.pcidev)
                pci_dev_put(intel_private.pcidev);
        if (intel_private.bridge_dev)