i2c: aspeed: backport I2C patches from v5.1
cover i2c: aspeed: backport I2C patches from v5.1
This patch series backports changes on Aspeed I2C driver from upstream v5.1 tree to support multi-master bus environment stably.
To enable the multi-master configuration with this patch, each platform's dt overlay should have 'multi-master' setting like below example:
&i2c5 {
multi-master;
status = "okay";
};Jae Hyun Yoo (2):
i2c: aspeed: Remove hard-coded bus timeout value setting
i2c: aspeed: Add multi-master use case support
drivers/i2c/busses/i2c-aspeed.c | 120 +++++++++++++++++++++++++------- 1 file changed, 93 insertions(+), 27 deletions(-)
Signature · 1 line
2.21.0
PATCH 1/2 i2c: aspeed: Remove hard-coded bus timeout value setting
This commit removes hard-coded bus timeout value setting so that it can be set by i2c-core-base.
--- drivers/i2c/busses/i2c-aspeed.c | 1 - 1 file changed, 1 deletion(-)
drivers/i2c/busses/i2c-aspeed.c +0 -1
@@ -930,7 +930,6 @@ static int aspeed_i2c_probe_bus(struct platform_device *pdev)init_completion (&bus->cmd_complete); bus->adap.owner =THIS_MODULE ; bus->adap.retries =0 ; - bus->adap.timeout =5 *HZ ; bus->adap.algo = &aspeed_i2c_algo; bus->adap.dev.parent = &pdev->dev; bus->adap.dev.of_node = pdev->dev.of_node;
Signature · 1 line
2.21.0
PATCH 2/2 i2c: aspeed: Add multi-master use case support
In reply to Jae Hyun YooIn multi-master environment, this driver's master cannot know exactly when a peer master sends data to this driver's slave so cases can be happened that this master tries sending data through the master_xfer function but slave data from a peer master is still being processed or slave xfer is started by a peer immediately after it queues a master command. To support multi-master use cases properly, this H/W provides arbitration in physical level and it provides priority based command handling too to avoid conflicts in multi-master environment, means that if a master and a slave events happen at the same time, H/W will handle a higher priority event first and a pending event will be handled when bus comes back to the idle state.
To support this H/W feature properly, this patch adds the 'pending' state of master and its handling code so that the pending master xfer can be continued after slave operation properly.
--- drivers/i2c/busses/i2c-aspeed.c | 119 +++++++++++++++++++++++++------- 1 file changed, 93 insertions(+), 26 deletions(-)
drivers/i2c/busses/i2c-aspeed.c +93 -26
@@ -117,6 +117,7 @@enum aspeed_i2c_master_state {ASPEED_I2C_MASTER_INACTIVE , +ASPEED_I2C_MASTER_PENDING ,ASPEED_I2C_MASTER_START ,ASPEED_I2C_MASTER_TX_FIRST ,ASPEED_I2C_MASTER_TX ,
@@ -126,12 +127,13 @@ enum aspeed_i2c_master_state { };enum aspeed_i2c_slave_state { -ASPEED_I2C_SLAVE_STOP , +ASPEED_I2C_SLAVE_INACTIVE ,ASPEED_I2C_SLAVE_START ,ASPEED_I2C_SLAVE_READ_REQUESTED ,ASPEED_I2C_SLAVE_READ_PROCESSED ,ASPEED_I2C_SLAVE_WRITE_REQUESTED ,ASPEED_I2C_SLAVE_WRITE_RECEIVED , +ASPEED_I2C_SLAVE_STOP , };struct aspeed_i2c_bus {
@@ -156,6 +158,8 @@ struct aspeed_i2c_bus {int cmd_err;/* Protected only by i2c_lock_bus */ int master_xfer_result; +/* Multi-master */ +bool multi_master;#if IS_ENABLED (CONFIG_I2C_SLAVE )struct i2c_client *slave;enum aspeed_i2c_slave_state slave_state;
@@ -251,7 +255,7 @@ static u32 aspeed_i2c_slave_irq(struct aspeed_i2c_bus *bus, u32 irq_status) }/* Slave is not currently active, irq was for someone else. */ -if (bus->slave_state ==ASPEED_I2C_SLAVE_STOP ) +if (bus->slave_state ==ASPEED_I2C_SLAVE_INACTIVE )return irq_handled; dev_dbg(bus->dev, "slave irq status 0x%08x, cmd 0x%08x\n",
@@ -277,16 +281,15 @@ static u32 aspeed_i2c_slave_irq(struct aspeed_i2c_bus *bus, u32 irq_status) irq_handled |=ASPEED_I2CD_INTR_NORMAL_STOP ; bus->slave_state =ASPEED_I2C_SLAVE_STOP ; } -if (irq_status &ASPEED_I2CD_INTR_TX_NAK ) { +if (irq_status &ASPEED_I2CD_INTR_TX_NAK && + bus->slave_state ==ASPEED_I2C_SLAVE_READ_PROCESSED ) { irq_handled |=ASPEED_I2CD_INTR_TX_NAK ; bus->slave_state =ASPEED_I2C_SLAVE_STOP ; } -if (irq_status &ASPEED_I2CD_INTR_TX_ACK ) - irq_handled |=ASPEED_I2CD_INTR_TX_ACK ;switch (bus->slave_state) {case ASPEED_I2C_SLAVE_READ_REQUESTED : -if (irq_status &ASPEED_I2CD_INTR_TX_ACK ) +if (unlikely (irq_status &ASPEED_I2CD_INTR_TX_ACK ))dev_err (bus->dev,"Unexpected ACK on read request.\n" ); bus->slave_state =ASPEED_I2C_SLAVE_READ_PROCESSED ;i2c_slave_event (slave,I2C_SLAVE_READ_REQUESTED , &value);
@@ -294,9 +297,12 @@ static u32 aspeed_i2c_slave_irq(struct aspeed_i2c_bus *bus, u32 irq_status)writel (ASPEED_I2CD_S_TX_CMD , bus->base +ASPEED_I2C_CMD_REG );break ;case ASPEED_I2C_SLAVE_READ_PROCESSED : -if (!(irq_status &ASPEED_I2CD_INTR_TX_ACK )) +if (unlikely (!(irq_status &ASPEED_I2CD_INTR_TX_ACK ))) {dev_err (bus->dev,"Expected ACK after processed read.\n" ); +break ; + } + irq_handled |=ASPEED_I2CD_INTR_TX_ACK ;i2c_slave_event (slave,I2C_SLAVE_READ_PROCESSED , &value);writel (value, bus->base +ASPEED_I2C_BYTE_BUF_REG );writel (ASPEED_I2CD_S_TX_CMD , bus->base +ASPEED_I2C_CMD_REG );
@@ -310,10 +316,15 @@ static u32 aspeed_i2c_slave_irq(struct aspeed_i2c_bus *bus, u32 irq_status)break ;case ASPEED_I2C_SLAVE_STOP :i2c_slave_event (slave,I2C_SLAVE_STOP , &value); + bus->slave_state =ASPEED_I2C_SLAVE_INACTIVE ; +break ; +case ASPEED_I2C_SLAVE_START : +/* Slave was just started. Waiting for the next event. */ ;break ;default : -dev_err (bus->dev,"unhandled slave_state: %d\n" , +dev_err (bus->dev,"unknown slave_state: %d\n" , bus->slave_state); + bus->slave_state =ASPEED_I2C_SLAVE_INACTIVE ;break ; }
@@ -329,6 +340,17 @@ static void aspeed_i2c_do_start(struct aspeed_i2c_bus *bus)u8 slave_addr =i2c_8bit_addr_from_msg (msg); bus->master_state =ASPEED_I2C_MASTER_START ; + +#if IS_ENABLED (CONFIG_I2C_SLAVE ) +/* +* If it's requested in the middle of a slave session, set the master +* state to 'pending' then H/W will continue handling this master +* command when the bus comes back to the idle state. +*/ +if (bus->slave_state !=ASPEED_I2C_SLAVE_INACTIVE ) + bus->master_state =ASPEED_I2C_MASTER_PENDING ; +#endif /* CONFIG_I2C_SLAVE */ + bus->buf_index =0 ;if (msg->flags &I2C_M_RD ) {
@@ -384,10 +406,6 @@ static u32 aspeed_i2c_master_irq(struct aspeed_i2c_bus *bus, u32 irq_status) bus->master_state =ASPEED_I2C_MASTER_INACTIVE ; irq_handled |=ASPEED_I2CD_INTR_BUS_RECOVER_DONE ; goto out_complete; - }else { -/* Master is not currently active, irq was for someone else. */ -if (bus->master_state ==ASPEED_I2C_MASTER_INACTIVE ) - goto out_no_complete; }/*
@@ -399,11 +417,32 @@ static u32 aspeed_i2c_master_irq(struct aspeed_i2c_bus *bus, u32 irq_status)if (ret) {dev_dbg (bus->dev,"received error interrupt: 0x%08x\n" , irq_status); - bus->cmd_err = ret; - bus->master_state =ASPEED_I2C_MASTER_INACTIVE ; irq_handled |= (irq_status &ASPEED_I2CD_INTR_MASTER_ERRORS ); - goto out_complete; +if (bus->master_state !=ASPEED_I2C_MASTER_INACTIVE ) { + bus->cmd_err = ret; + bus->master_state =ASPEED_I2C_MASTER_INACTIVE ; + goto out_complete; + } + } + +#if IS_ENABLED (CONFIG_I2C_SLAVE ) +/* +* A pending master command will be started by H/W when the bus comes +* back to idle state after completing a slave operation so change the +* master state from 'pending' to 'start' at here if slave is inactive. +*/ +if (bus->master_state ==ASPEED_I2C_MASTER_PENDING ) { +if (bus->slave_state !=ASPEED_I2C_SLAVE_INACTIVE ) + goto out_no_complete; + + bus->master_state =ASPEED_I2C_MASTER_START ; } +#endif /* CONFIG_I2C_SLAVE */ + +/* Master is not currently active, irq was for someone else. */ +if (bus->master_state ==ASPEED_I2C_MASTER_INACTIVE || + bus->master_state ==ASPEED_I2C_MASTER_PENDING ) + goto out_no_complete;/* We are in an invalid state; reset bus to a known state. */ if (!bus->msgs) {
@@ -423,6 +462,20 @@ static u32 aspeed_i2c_master_irq(struct aspeed_i2c_bus *bus, u32 irq_status)* then update the state and handle the new state below. */ if (bus->master_state ==ASPEED_I2C_MASTER_START ) { +#if IS_ENABLED (CONFIG_I2C_SLAVE ) +/* +* If a peer master starts a xfer immediately after it queues a +* master command, change its state to 'pending' then H/W will +* continue the queued master xfer just after completing the +* slave mode session. +*/ +if (unlikely (irq_status &ASPEED_I2CD_INTR_SLAVE_MATCH )) { + bus->master_state =ASPEED_I2C_MASTER_PENDING ; +dev_dbg (bus->dev, +"master goes pending due to a slave start\n" ); + goto out_no_complete; + } +#endif /* CONFIG_I2C_SLAVE */ if (unlikely (!(irq_status &ASPEED_I2CD_INTR_TX_ACK ))) {if (unlikely (!(irq_status &ASPEED_I2CD_INTR_TX_NAK ))) { bus->cmd_err = -ENXIO ;
@@ -566,7 +619,8 @@ static irqreturn_t aspeed_i2c_bus_irq(int irq, void *dev_id)* interrupt bits. Each case needs to be handled using corresponding * handlers depending on the current state. */ -if (bus->master_state !=ASPEED_I2C_MASTER_INACTIVE ) { +if (bus->master_state !=ASPEED_I2C_MASTER_INACTIVE && + bus->master_state !=ASPEED_I2C_MASTER_PENDING ) { irq_handled =aspeed_i2c_master_irq (bus, irq_remaining); irq_remaining &= ~irq_handled;if (irq_remaining)
@@ -601,15 +655,16 @@ static int aspeed_i2c_master_xfer(struct i2c_adapter *adap, {struct aspeed_i2c_bus *bus =i2c_get_adapdata (adap);unsigned long time_left, flags; -int ret =0 ;spin_lock_irqsave (&bus->lock, flags); bus->cmd_err =0 ; -/* If bus is busy, attempt recovery. We assume a single master -* environment. -*/ -if (readl (bus->base +ASPEED_I2C_CMD_REG ) &ASPEED_I2CD_BUS_BUSY_STS ) { +/* If bus is busy in a single master environment, attempt recovery. */ +if (!bus->multi_master && + (readl (bus->base +ASPEED_I2C_CMD_REG ) & +ASPEED_I2CD_BUS_BUSY_STS )) { +int ret; +spin_unlock_irqrestore (&bus->lock, flags); ret =aspeed_i2c_recover_bus (bus);if (ret)
@@ -629,10 +684,20 @@ static int aspeed_i2c_master_xfer(struct i2c_adapter *adap, time_left =wait_for_completion_timeout (&bus->cmd_complete, bus->adap.timeout); -if (time_left ==0 ) +if (time_left ==0 ) { +/* +* If timed out and bus is still busy in a multi master +* environment, attempt recovery at here. +*/ +if (bus->multi_master && + (readl (bus->base +ASPEED_I2C_CMD_REG ) & +ASPEED_I2CD_BUS_BUSY_STS )) +aspeed_i2c_recover_bus (bus); +return -ETIMEDOUT ; -else -return bus->master_xfer_result; + } + +return bus->master_xfer_result; }static u32 aspeed_i2c_functionality (struct i2c_adapter *adap)
@@ -672,7 +737,7 @@ static int aspeed_i2c_reg_slave(struct i2c_client *client)__aspeed_i2c_reg_slave (bus, client->addr); bus->slave = client; - bus->slave_state =ASPEED_I2C_SLAVE_STOP ; + bus->slave_state =ASPEED_I2C_SLAVE_INACTIVE ;spin_unlock_irqrestore (&bus->lock, flags);return 0 ;
@@ -827,7 +892,9 @@ static int aspeed_i2c_init(struct aspeed_i2c_bus *bus,if (ret <0 )return ret; -if (!of_property_read_bool (pdev->dev.of_node,"multi-master" )) +if (of_property_read_bool (pdev->dev.of_node,"multi-master" )) + bus->multi_master = true; +else fun_ctrl_reg |=ASPEED_I2CD_MULTI_MASTER_DIS ;/* Enable Master Mode */
Signature · 1 line
2.21.0
Joel Stanley
maintainerHi Jae,
Quoted 3 lines · Jae Hyun Yoo
This patch series backports changes on Aspeed I2C driver from upstream v5.1 tree to support multi-master bus environment stably.
Thanks. These are applied to dev-5.0.
Cheers,
Joel
Quoted 19 lines
To enable the multi-master configuration with this patch, each platform's dt overlay should have 'multi-master' setting like below example:
&i2c5 { multi-master; status = "okay"; };Jae Hyun Yoo (2):
i2c: aspeed: Remove hard-coded bus timeout value setting
i2c: aspeed: Add multi-master use case supportdrivers/i2c/busses/i2c-aspeed.c | 120 +++++++++++++++++++++++++------- 1 file changed, 93 insertions(+), 27 deletions(-)--
2.21.0