From: Xiang Wang <wangx@xxxxxxxxxxx>
There might be such calling sequence for dma channel:
1. register dma device
dma_async_device_register
-> if (dma_has_cap(DMA_PRIVATE, device->cap_mask))
device->privatecnt++;
2. request channel
dma_request_slave_channel
-> of_dma_request_slave_channel
-> of_dma_xlate
-> dma_get_any_slave_channel in dma drivers
note that device->privatecnt is not changed during this.
3. release channel
dma_release_channel
-> if (--chan->device->privatecnt == 0)
dma_cap_clear(DMA_PRIVATE,
chan->device->cap_mask);
So if we request a channel then release it, DMA_PRIVATE will
be cleared unexpectedly. In this patch, we increase privatecnt
in step 2 just like what __dma_request_channel does.
Signed-off-by: Xiang Wang <wangx@xxxxxxxxxxx>
---
drivers/dma/dmaengine.c | 3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c
index ed610b4..01bb372 100644
--- a/drivers/dma/dmaengine.c
+++ b/drivers/dma/dmaengine.c
@@ -557,6 +557,9 @@ struct dma_chan *dma_get_any_slave_channel(struct dma_device *device)
}
}
+ if (chan && dma_has_cap(DMA_PRIVATE, device->cap_mask))
+ device->privatecnt++;
+
mutex_unlock(&dma_list_mutex);
return chan;
--
1.7.5.4
--
To unsubscribe from this list: send the line "unsubscribe dmaengine" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
