6.5-stable review patch. If anyone has any objections, please let me know.
------------------
From: Jason Gunthorpe <jgg@xxxxxxxxxx>
[ Upstream commit f429378a9bf84d79a7e2cae05d2e3384cf7d68ba ]
When test_remove is enabled really_probe() does not properly pair
dma_configure() with dma_remove(), it will end up calling dma_configure()
twice. This corrupts the owner_cnt and renders the group unusable with
VFIO/etc.
Add the missing cleanup before going back to re_probe.
Fixes: 25f3bcfc54bc ("driver core: Add dma_cleanup callback in bus_type")
Reported-by: Zenghui Yu <yuzenghui@xxxxxxxxxx>
Tested-by: Zenghui Yu <yuzenghui@xxxxxxxxxx>
Closes: https://lore.kernel.org/all/6472f254-c3c4-8610-4a37-8d9dfdd54ce8@xxxxxxxxxx/
Signed-off-by: Jason Gunthorpe <jgg@xxxxxxxxxx>
Reviewed-by: Kevin Tian <kevin.tian@xxxxxxxxx>
Link: https://lore.kernel.org/r/0-v2-4deed94e283e+40948-really_probe_dma_cleanup_jgg@xxxxxxxxxx
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/base/dd.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/base/dd.c b/drivers/base/dd.c
index 878aa7646b37e..a528cec24264a 100644
--- a/drivers/base/dd.c
+++ b/drivers/base/dd.c
@@ -693,6 +693,8 @@ static int really_probe(struct device *dev, struct device_driver *drv)
device_remove(dev);
driver_sysfs_remove(dev);
+ if (dev->bus && dev->bus->dma_cleanup)
+ dev->bus->dma_cleanup(dev);
device_unbind_cleanup(dev);
goto re_probe;
--
2.40.1
