Hi Loic,
On 12/10/20 3:31 AM, Loic Poulain wrote:
Instead of duplicating queuing procedure in mhi_queue_dma(),
mhi_queue_buf() and mhi_queue_skb(), add a new generic mhi_queue()
as common helper.
Signed-off-by: Loic Poulain <loic.poulain@xxxxxxxxxx>
---
drivers/bus/mhi/core/main.c | 160 +++++++++++---------------------------------
1 file changed, 38 insertions(+), 122 deletions(-)
diff --git a/drivers/bus/mhi/core/main.c b/drivers/bus/mhi/core/main.c
index 3871ef0..4fa4c88 100644
--- a/drivers/bus/mhi/core/main.c
+++ b/drivers/bus/mhi/core/main.c
@@ -963,118 +963,78 @@ static bool mhi_is_ring_full(struct mhi_controller *mhi_cntrl,
return (tmp == ring->rp);
}
-int mhi_queue_skb(struct mhi_device *mhi_dev, enum dma_data_direction dir,
- struct sk_buff *skb, size_t len, enum mhi_flags mflags)
+static int mhi_queue(struct mhi_device *mhi_dev, struct mhi_buf_info *buf_info,
+ enum dma_data_direction dir, enum mhi_flags mflags)
{
struct mhi_controller *mhi_cntrl = mhi_dev->mhi_cntrl;
struct mhi_chan *mhi_chan = (dir == DMA_TO_DEVICE) ? mhi_dev->ul_chan :
mhi_dev->dl_chan;
struct mhi_ring *tre_ring = &mhi_chan->tre_ring;
- struct mhi_buf_info buf_info = { };
+ unsigned long flags;
int ret;
- /* If MHI host pre-allocates buffers then client drivers cannot queue */
- if (mhi_chan->pre_alloc)
Are we getting rid of auto_queue ?
- return -EINVAL;
+ if (unlikely(MHI_PM_IN_ERROR_STATE(mhi_cntrl->pm_state)))
+ return -EIO;
- if (mhi_is_ring_full(mhi_cntrl, tre_ring))
- return -ENOMEM;
+ read_lock_irqsave(&mhi_cntrl->pm_lock, flags);
This is something used in mhi_queue_buf and rest of the (_skb, _dma)
used read_lock_bh. I see that now you are making it a re-interant as you
identified before :) So irq ver is disabling irq on the core this code
is running and disabling preemption, where as bh ver is preemptiable. Is
that the reason you chose irq instead of bh ?
[..]
Thanks,
Hemant
--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
