Tegra I2C DMA and SMBus blockread updates
v6 cover Tegra I2C DMA and SMBus blockread updates
The patchset contains two independent updates to tegra-i2c driver.
* Update I2C DMA channel usage to use single DMA channel instead of two different hardware channels.
* Fix PEC byte issue during SMBus block read caused by mismanaged msg->len property in i2c_msg struct.
v5->v6:
Updated __u16 to unsigned intv4->v5:
Corrected the comment stylev3->v4:
Combine the two patches to one series to avoid merge conflicts.
Akhil R (2):
i2c: tegra: Fix PEC support for SMBUS block read
i2c: tegra: Share same DMA channel for RX and TX
drivers/i2c/busses/i2c-tegra.c | 109 ++++++++++++++------------------- 1 file changed, 47 insertions(+), 62 deletions(-)
Signature · 1 line
2.17.1
PATCH RESEND v6 1/2 i2c: tegra: Fix PEC support for SMBUS block read
Update the msg->len value correctly for SMBUS block read. The discrepancy went unnoticed as msg->len is used in SMBUS transfers only when a PEC byte is added.
--- drivers/i2c/busses/i2c-tegra.c | 40 +++++++++++++++++++++++----------- 1 file changed, 27 insertions(+), 13 deletions(-)
drivers/i2c/busses/i2c-tegra.c +27 -13
@@ -242,9 +242,10 @@ struct tegra_i2c_hw_feature {* @is_dvc: identifies the DVC I2C controller, has a different register layout * @is_vi: identifies the VI I2C controller, has a different register layout * @msg_complete: transfer completion notifier +* @msg_buf_remaining: size of unsent data in the message buffer +* @msg_len: length of message in current transfer * @msg_err: error code for completed message * @msg_buf: pointer to current message data -* @msg_buf_remaining: size of unsent data in the message buffer * @msg_read: indicates that the transfer is a read access * @timings: i2c timings information like bus frequency * @multimaster_mode: indicates that I2C controller is in multi-master mode
@@ -277,6 +278,7 @@ struct tegra_i2c_dev {struct completion msg_complete;size_t msg_buf_remaining; +unsigned int msg_len;int msg_err;u8 *msg_buf;
@@ -1169,7 +1171,7 @@ static void tegra_i2c_push_packet_header(struct tegra_i2c_dev *i2c_dev,else i2c_writel (i2c_dev, packet_header,I2C_TX_FIFO ); - packet_header = msg->len -1 ; + packet_header = i2c_dev->msg_len -1 ;if (i2c_dev->dma_mode && !i2c_dev->msg_read) *dma_buf++ = packet_header;
@@ -1242,20 +1244,32 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev,return err; i2c_dev->msg_buf = msg->buf; + i2c_dev->msg_len = msg->len; -/* The condition true implies smbus block read and len is already read */ -if (msg->flags &I2C_M_RECV_LEN && end_state !=MSG_END_CONTINUE ) - i2c_dev->msg_buf = msg->buf +1 ; - - i2c_dev->msg_buf_remaining = msg->len; i2c_dev->msg_err =I2C_ERR_NONE ; i2c_dev->msg_read = !!(msg->flags &I2C_M_RD );reinit_completion (&i2c_dev->msg_complete); +/* +* For SMBUS block read command, read only 1 byte in the first transfer. +* Adjust that 1 byte for the next transfer in the msg buffer and msg +* length. +*/ +if (msg->flags &I2C_M_RECV_LEN ) { +if (end_state ==MSG_END_CONTINUE ) { + i2c_dev->msg_len =1 ; + }else { + i2c_dev->msg_buf +=1 ; + i2c_dev->msg_len -=1 ; + } + } + + i2c_dev->msg_buf_remaining = i2c_dev->msg_len; +if (i2c_dev->msg_read) - xfer_size = msg->len; + xfer_size = i2c_dev->msg_len;else - xfer_size = msg->len +I2C_PACKET_HEADER_SIZE ; + xfer_size = i2c_dev->msg_len +I2C_PACKET_HEADER_SIZE ; xfer_size =ALIGN (xfer_size,BYTES_PER_FIFO_WORD );
@@ -1295,7 +1309,7 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev,if (!i2c_dev->msg_read) {if (i2c_dev->dma_mode) {memcpy (i2c_dev->dma_buf +I2C_PACKET_HEADER_SIZE , - msg->buf, msg->len); + msg->buf, i2c_dev->msg_len); dma_sync_single_for_device(i2c_dev->dma_dev, i2c_dev->dma_phys,
@@ -1352,7 +1366,7 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev, i2c_dev->dma_phys, xfer_size, DMA_FROM_DEVICE); -memcpy (i2c_dev->msg_buf, i2c_dev->dma_buf, msg->len); +memcpy (i2c_dev->msg_buf, i2c_dev->dma_buf, i2c_dev->msg_len); } }
@@ -1408,8 +1422,8 @@ static int tegra_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[], ret =tegra_i2c_xfer_msg (i2c_dev, &msgs[i],MSG_END_CONTINUE );if (ret)break ; -/* Set the read byte as msg len */ - msgs[i].len = msgs[i].buf[0 ]; +/* Set the msg length from first byte */ + msgs[i].len += msgs[i].buf[0 ];dev_dbg (i2c_dev->dev,"reading %d bytes\n" , msgs[i].len); } ret =tegra_i2c_xfer_msg (i2c_dev, &msgs[i], end_type);
Signature · 1 line
2.17.1
Wolfram Sang
AppliedRe: [PATCH v6 RESEND 1/2] i2c: tegra: Fix PEC support for SMBUS block read
Quoted 7 lines · Akhil R
Update the
msg->lenvalue correctly for SMBUS block read. The discrepancy went unnoticed asmsg->lenis used in SMBUS transfers only when a PEC byte is added.Fixes d7583c8a5748 ("i2c: tegra: Add SMBus block read function")Signed-off-by Akhil R <[email protected]>Acked-by Thierry Reding <[email protected]>
Applied to for-current, thanks!
PATCH RESEND v6 2/2 i2c: tegra: Share same DMA channel for RX and TX
In reply to Akhil RAllocate only one DMA channel for I2C and share it for both TX and RX instead of using two different DMA hardware channels with the same slave ID. Since I2C supports only half duplex, there is no impact on perf with this.
--- drivers/i2c/busses/i2c-tegra.c | 69 ++++++++++------------------------ 1 file changed, 20 insertions(+), 49 deletions(-)
drivers/i2c/busses/i2c-tegra.c +20 -49
@@ -249,8 +249,7 @@ struct tegra_i2c_hw_feature {* @msg_read: indicates that the transfer is a read access * @timings: i2c timings information like bus frequency * @multimaster_mode: indicates that I2C controller is in multi-master mode -* @tx_dma_chan: DMA transmit channel -* @rx_dma_chan: DMA receive channel +* @dma_chan: DMA channel * @dma_phys: handle to DMA resources * @dma_buf: pointer to allocated DMA buffer * @dma_buf_size: DMA buffer size
@@ -283,8 +282,7 @@ struct tegra_i2c_dev {u8 *msg_buf;struct completion dma_complete; -struct dma_chan *tx_dma_chan; -struct dma_chan *rx_dma_chan; +struct dma_chan *dma_chan;unsigned int dma_buf_size;struct device *dma_dev;dma_addr_t dma_phys;
@@ -393,16 +391,14 @@ static int tegra_i2c_dma_submit(struct tegra_i2c_dev *i2c_dev, size_t len) {struct dma_async_tx_descriptor *dma_desc;enum dma_transfer_direction dir; -struct dma_chan *chan;dev_dbg (i2c_dev->dev,"starting DMA for length: %zu\n" , len);reinit_completion (&i2c_dev->dma_complete); dir = i2c_dev->msg_read ?DMA_DEV_TO_MEM :DMA_MEM_TO_DEV ; - chan = i2c_dev->msg_read ? i2c_dev->rx_dma_chan : i2c_dev->tx_dma_chan; - dma_desc =dmaengine_prep_slave_single (chan, i2c_dev->dma_phys, + dma_desc =dmaengine_prep_slave_single (i2c_dev->dma_chan, i2c_dev->dma_phys, len, dir,DMA_PREP_INTERRUPT |DMA_CTRL_ACK );if (!dma_desc) {
@@ -415,7 +411,7 @@ static int tegra_i2c_dma_submit(struct tegra_i2c_dev *i2c_dev, size_t len) dma_desc->callback_param = i2c_dev;dmaengine_submit (dma_desc); -dma_async_issue_pending (chan); +dma_async_issue_pending (i2c_dev->dma_chan);return 0 ; }
@@ -428,20 +424,14 @@ static void tegra_i2c_release_dma(struct tegra_i2c_dev *i2c_dev) i2c_dev->dma_buf =NULL ; } -if (i2c_dev->tx_dma_chan) { -dma_release_channel (i2c_dev->tx_dma_chan); - i2c_dev->tx_dma_chan =NULL ; - } - -if (i2c_dev->rx_dma_chan) { -dma_release_channel (i2c_dev->rx_dma_chan); - i2c_dev->rx_dma_chan =NULL ; +if (i2c_dev->dma_chan) { +dma_release_channel (i2c_dev->dma_chan); + i2c_dev->dma_chan =NULL ; } }static int tegra_i2c_init_dma (struct tegra_i2c_dev *i2c_dev) { -struct dma_chan *chan;dma_addr_t dma_phys;u32 *dma_buf;int err;
@@ -459,25 +449,18 @@ static int tegra_i2c_init_dma(struct tegra_i2c_dev *i2c_dev)return 0 ; } - chan =dma_request_chan (i2c_dev->dev,"rx" ); -if (IS_ERR (chan)) { - err =PTR_ERR (chan); - goto err_out; - } - - i2c_dev->rx_dma_chan = chan; - - chan =dma_request_chan (i2c_dev->dev,"tx" ); -if (IS_ERR (chan)) { - err =PTR_ERR (chan); +/* +* The same channel will be used for both RX and TX. +* Keeping the name as "tx" for backward compatibility +* with existing devicetrees. +*/ + i2c_dev->dma_chan =dma_request_chan (i2c_dev->dev,"tx" ); +if (IS_ERR (i2c_dev->dma_chan)) { + err =PTR_ERR (i2c_dev->dma_chan); goto err_out; } - i2c_dev->tx_dma_chan = chan; - -WARN_ON (i2c_dev->tx_dma_chan->device != i2c_dev->rx_dma_chan->device); - i2c_dev->dma_dev = chan->device->dev; - + i2c_dev->dma_dev = i2c_dev->dma_chan->device->dev; i2c_dev->dma_buf_size = i2c_dev->hw->quirks->max_write_len +I2C_PACKET_HEADER_SIZE ;
@@ -976,11 +959,7 @@ static irqreturn_t tegra_i2c_isr(int irq, void *dev_id)dvc_writel (i2c_dev,DVC_STATUS_I2C_DONE_INTR ,DVC_STATUS );if (i2c_dev->dma_mode) { -if (i2c_dev->msg_read) -dmaengine_terminate_async (i2c_dev->rx_dma_chan); -else -dmaengine_terminate_async (i2c_dev->tx_dma_chan); - +dmaengine_terminate_async (i2c_dev->dma_chan);complete (&i2c_dev->dma_complete); }
@@ -994,7 +973,6 @@ static void tegra_i2c_config_fifo_trig(struct tegra_i2c_dev *i2c_dev, {struct dma_slave_config slv_config = {0 };u32 val, reg, dma_burst, reg_offset; -struct dma_chan *chan;int err;if (i2c_dev->hw->has_mst_fifo)
@@ -1011,7 +989,6 @@ static void tegra_i2c_config_fifo_trig(struct tegra_i2c_dev *i2c_dev, dma_burst =8 ;if (i2c_dev->msg_read) { - chan = i2c_dev->rx_dma_chan; reg_offset =tegra_i2c_reg_addr (i2c_dev,I2C_RX_FIFO ); slv_config.src_addr = i2c_dev->base_phys + reg_offset;
@@ -1023,7 +1000,6 @@ static void tegra_i2c_config_fifo_trig(struct tegra_i2c_dev *i2c_dev,else val =I2C_FIFO_CONTROL_RX_TRIG (dma_burst); }else { - chan = i2c_dev->tx_dma_chan; reg_offset =tegra_i2c_reg_addr (i2c_dev,I2C_TX_FIFO ); slv_config.dst_addr = i2c_dev->base_phys + reg_offset;
@@ -1037,7 +1013,7 @@ static void tegra_i2c_config_fifo_trig(struct tegra_i2c_dev *i2c_dev, } slv_config.device_fc = true; - err =dmaengine_slave_config (chan, &slv_config); + err =dmaengine_slave_config (i2c_dev->dma_chan, &slv_config);if (err) {dev_err (i2c_dev->dev,"DMA config failed: %d\n" , err);dev_err (i2c_dev->dev,"falling back to PIO\n" );
@@ -1347,13 +1323,8 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev,* performs synchronization after the transfer's termination * and we want to get a completion if transfer succeeded. */ -dmaengine_synchronize (i2c_dev->msg_read ? - i2c_dev->rx_dma_chan : - i2c_dev->tx_dma_chan); - -dmaengine_terminate_sync (i2c_dev->msg_read ? - i2c_dev->rx_dma_chan : - i2c_dev->tx_dma_chan); +dmaengine_synchronize (i2c_dev->dma_chan); +dmaengine_terminate_sync (i2c_dev->dma_chan);if (!time_left && !completion_done (&i2c_dev->dma_complete)) {dev_err (i2c_dev->dev,"DMA transfer timed out\n" );
Signature · 1 line
2.17.1
RE: [PATCH v6 RESEND 2/2] i2c: tegra: Share same DMA channel for RX and TX
Quoted 11 lines
Allocate only one DMA channel for I2C and share it for both TX and RX instead of using two different DMA hardware channels with the same slave ID. Since I2C supports only half duplex, there is no impact on perf with this.
Signed-off-by Akhil R <[email protected]>Acked-by Thierry Reding <[email protected]>--- drivers/i2c/busses/i2c-tegra.c | 69 ++++++++++------------------------ 1 file changed, 20 insertions(+), 49 deletions(-)
Hi Wolfram,
Thanks for applying the patch 1/2 of this series.
I saw that this patch is not applied currently.
Do we have any concerns on applying this?
Regards,
Akhil
Wolfram Sang
AppliedRe: [PATCH v6 RESEND 2/2] i2c: tegra: Share same DMA channel for RX and TX
In reply to Akhil RQuoted 7 lines · Akhil R
Allocate only one DMA channel for I2C and share it for both TX and RX instead of using two different DMA hardware channels with the same slave ID. Since I2C supports only half duplex, there is no impact on perf with this.
Signed-off-by Akhil R <[email protected]>Acked-by Thierry Reding <[email protected]>
Applied to for-next, thanks!
No, there wasn't an issue with this patch. I just needed to make sure patch 1 (which was for-current material) landed upstream to avoid merge conflicts.
RE: [PATCH v6 RESEND 2/2] i2c: tegra: Share same DMA channel for RX and TX
Quoted 6 lines · Akhil R
Quoted 5 lines
Allocate only one DMA channel for I2C and share it for both TX and RX instead of using two different DMA hardware channels with the same slave ID. Since I2C supports only half duplex, there is no impact on perf with this.
Quoted 8 lines
Quoted 2 lines
Signed-off-by Akhil R <[email protected]>Acked-by Thierry Reding <[email protected]>Applied to for-next, thanks!
No, there wasn't an issue with this patch. I just needed to make sure patch 1 (which was for-current material) landed upstream to avoid merge conflicts.
Thank you :)