dma: mxs-dma bugfixes and cleanup
Subsystems (2)
v5 cover dma: mxs-dma bugfixes and cleanup
Hi,
v5 is rebased onto your dma_complete branch. It includes one bugfix.
This series contains a number of bugfixes for the mxs DMA engine.
Regards,
Markus Pargmann
Markus Pargmann (5):
dma: mxs-dma: Cleanup interrupt handler
dma: mxs-dma: Report correct residue for cyclic DMA
dma: mxs-dma: Fix channel reset hardware bug
dma: mxs-dma: Update state after channel reset
dma: mxs-dma: Use semaphores for cyclic DMA
drivers/dma/mxs-dma.c | 176 +++++++++++++++++++++++++++++++++++++++----------- 1 file changed, 137 insertions(+), 39 deletions(-)
Signature · 1 line
1.8.4.rc3
PATCH v5 1/5 dma: mxs-dma: Cleanup interrupt handler
The DMA interrupt handler uses its controll registers to handle all available channel interrupts it can find.
This patch changes it to handle only one interrupt by directly mapping irq number to channel. It also includes a cleanup of the ctrl-register usage.
--- drivers/dma/mxs-dma.c | 96 ++++++++++++++++++++++++++++++++------------------- 1 file changed, 60 insertions(+), 36 deletions(-)
drivers/dma/mxs-dma.c +60 -36
@@ -27,6 +27,7 @@#include <linux/of.h> #include <linux/of_device.h> #include <linux/of_dma.h> +#include <linux/list.h> #include <asm/irq.h>
@@ -272,58 +273,81 @@ static void mxs_dma_tasklet(unsigned long data) mxs_chan->desc.callback (mxs_chan->desc.callback_param); } +static int mxs_dma_irq_to_chan (struct mxs_dma_engine *mxs_dma,int irq) +{ +int i; + +for (i =0 ; i != mxs_dma->nr_channels; ++i) +if (mxs_dma->mxs_chans[i].chan_irq == irq) +return i; + +return -EINVAL ; +} +static irqreturn_t mxs_dma_int_handler (int irq,void *dev_id) {struct mxs_dma_engine *mxs_dma = dev_id; -u32 stat1, stat2; +struct mxs_dma_chan *mxs_chan; +u32 completed; +u32 err; +int chan =mxs_dma_irq_to_chan (mxs_dma, irq); + +if (chan <0 ) +return IRQ_NONE ;/* completion status */ - stat1 =readl (mxs_dma->base +HW_APBHX_CTRL1 ); - stat1 &=MXS_DMA_CHANNELS_MASK ; -writel (stat1, mxs_dma->base +HW_APBHX_CTRL1 +STMP_OFFSET_REG_CLR ); + completed =readl (mxs_dma->base +HW_APBHX_CTRL1 ); + completed = (completed >> chan) &0x1 ; + +/* Clear interrupt */ +writel ((1 << chan), + mxs_dma->base +HW_APBHX_CTRL1 +STMP_OFFSET_REG_CLR );/* error status */ - stat2 =readl (mxs_dma->base +HW_APBHX_CTRL2 ); -writel (stat2, mxs_dma->base +HW_APBHX_CTRL2 +STMP_OFFSET_REG_CLR ); + err =readl (mxs_dma->base +HW_APBHX_CTRL2 ); + err &= (1 << (MXS_DMA_CHANNELS + chan)) | (1 << chan); + +/* +* error status bit is in the upper 16 bits, error irq bit in the lower +* 16 bits. We transform it into a simpler error code: +* err: 0x00 = no error, 0x01 = TERMINATION, 0x02 = BUS_ERROR +*/ + err = (err >> (MXS_DMA_CHANNELS + chan)) + (err >> chan); + +/* Clear error irq */ +writel ((1 << chan), + mxs_dma->base +HW_APBHX_CTRL2 +STMP_OFFSET_REG_CLR );/* * When both completion and error of termination bits set at the * same time, we do not take it as an error. IOW, it only becomes -* an error we need to handle here in case of either it's (1) a bus -* error or (2) a termination error with no completion. +* an error we need to handle here in case of either it's a bus +* error or a termination error with no completion. 0x01 is termination +* error, so we can subtract err & completed to get the real error case. */ - stat2 = ((stat2 >>MXS_DMA_CHANNELS ) & stat2) |/* (1) */ - (~(stat2 >>MXS_DMA_CHANNELS ) & stat2 & ~stat1);/* (2) */ - -/* combine error and completion status for checking */ - stat1 = (stat2 <<MXS_DMA_CHANNELS ) | stat1; -while (stat1) { -int channel =fls (stat1) -1 ; -struct mxs_dma_chan *mxs_chan = - &mxs_dma->mxs_chans[channel %MXS_DMA_CHANNELS ]; - -if (channel >=MXS_DMA_CHANNELS ) { -dev_dbg (mxs_dma->dma_device.dev, -"%s: error in channel %d\n" , __func__, - channel -MXS_DMA_CHANNELS ); - mxs_chan->status =DMA_ERROR ; -mxs_dma_reset_chan (mxs_chan); - }else { -if (mxs_chan->flags &MXS_DMA_SG_LOOP ) - mxs_chan->status =DMA_IN_PROGRESS ; -else - mxs_chan->status =DMA_COMPLETE ; - } + err -= err & completed; - stat1 &= ~(1 << channel); + mxs_chan = &mxs_dma->mxs_chans[chan]; -if (mxs_chan->status ==DMA_COMPLETE ) -dma_cookie_complete (&mxs_chan->desc); - -/* schedule tasklet on this channel */ -tasklet_schedule (&mxs_chan->tasklet); +if (err) { +dev_dbg (mxs_dma->dma_device.dev, +"%s: error in channel %d\n" , __func__, + chan); + mxs_chan->status =DMA_ERROR ; +mxs_dma_reset_chan (mxs_chan); + }else { +if (mxs_chan->flags &MXS_DMA_SG_LOOP ) + mxs_chan->status =DMA_IN_PROGRESS ; +else + mxs_chan->status =DMA_COMPLETE ; } +if (mxs_chan->status ==DMA_COMPLETE ) +dma_cookie_complete (&mxs_chan->desc); + +/* schedule tasklet on this channel */ +tasklet_schedule (&mxs_chan->tasklet); +return IRQ_HANDLED ; }
Signature · 1 line
1.8.4.rc3
PATCH v5 2/5 dma: mxs-dma: Report correct residue for cyclic DMA
In reply to Markus PargmannUse the channel's buffer address register to calculate correct residue value for tx_status.
--- drivers/dma/mxs-dma.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-)
drivers/dma/mxs-dma.c +19 -1
@@ -58,6 +58,8 @@ (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x050 : 0x110) + (n) * 0x70)#define HW_APBHX_CHn_SEMA (d, n) \ (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x080 : 0x140) + (n) * 0x70) +#define HW_APBHX_CHn_BAR (d, n) \ + (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x070 : 0x130) + (n) * 0x70)/* * ccw bits definitions
@@ -623,8 +625,24 @@ static enum dma_status mxs_dma_tx_status(struct dma_chan *chan, dma_cookie_t cookie, struct dma_tx_state *txstate) { struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan); +struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma; +u32 residue =0 ; + +if (mxs_chan->status ==DMA_IN_PROGRESS && + mxs_chan->flags &MXS_DMA_SG_LOOP ) { +struct mxs_dma_ccw *last_ccw; +u32 bar; + + last_ccw = &mxs_chan->ccw[mxs_chan->desc_count -1 ]; + residue = last_ccw->xfer_bytes + last_ccw->bufaddr; + + bar =readl (mxs_dma->base + +HW_APBHX_CHn_BAR (mxs_dma, chan->chan_id)); + residue -= bar; + } -dma_set_tx_state (txstate, chan->completed_cookie, chan->cookie,0 ); +dma_set_tx_state (txstate, chan->completed_cookie, chan->cookie, + residue);return mxs_chan->status; }
Signature · 1 line
1.8.4.rc3
PATCH v5 2/5 dma: mxs-dma: Report correct residue for cyclic DMA
Quoted 36 lines · Markus Pargmann
Use the channel's buffer address register to calculate correct residue value for
tx_status.Signed-off-by Markus Pargmann <[email protected]>--- drivers/dma/mxs-dma.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-)drivers/dma/mxs-dma.c +13
-0@@ -58,6 +58,8 @@ (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x050 : 0x110) + (n) * 0x70)#define HW_APBHX_CHn_SEMA (d, n) \ (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x080 : 0x140) + (n) * 0x70) +#define HW_APBHX_CHn_BAR (d, n) \ + (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x070 : 0x130) + (n) * 0x70)/* * ccw bits definitions @@ -623,8 +625,24 @@ static enum dma_status mxs_dma_tx_status(struct dma_chan *chan, dma_cookie_t cookie, struct dma_tx_state *txstate) { struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan); +struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma; +u32 residue =0 ; + +if (mxs_chan->status ==DMA_IN_PROGRESS && + mxs_chan->flags &MXS_DMA_SG_LOOP ) { +struct mxs_dma_ccw *last_ccw; +u32 bar; + + last_ccw = &mxs_chan->ccw[mxs_chan->desc_count -1 ]; + residue = last_ccw->xfer_bytes + last_ccw->bufaddr; +
drivers/dma/mxs-dma.c static enum dma_status mxs_dma_tx_status(struct dma_chan *chan, · line 647
Quoted 10 lines
drivers/dma/mxs-dma.c +6
-1+ bar =readl (mxs_dma->base + +HW_APBHX_CHn_BAR (mxs_dma, chan->chan_id)); + residue -= bar; + } -dma_set_tx_state (txstate, chan->completed_cookie, chan->cookie,0 ); +dma_set_tx_state (txstate, chan->completed_cookie, chan->cookie, + residue);return mxs_chan->status;
This will work for your driver where we have artficial restriction on allowing only 1 descriptor to be submitted at any point of time. I winder what made that restriction?
--
~Vinod
PATCH v5 3/5 dma: mxs-dma: Fix channel reset hardware bug
In reply to Markus PargmannThis is no official errata, but I noticed that the channel reset may stop working if the DMA state engine is in the READ_FLUSH state.
This patch uses the channel debug1 register to wait for the DMA statemachine to leave the READ_FLUSH state. After that we can continue to reset the channel.
Tested on i.MX28.
--- drivers/dma/mxs-dma.c | 29 +++++++++++++++++++++++++++-- 1 file changed, 27 insertions(+), 2 deletions(-)
drivers/dma/mxs-dma.c +27 -2
@@ -60,6 +60,7 @@ (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x080 : 0x140) + (n) * 0x70)#define HW_APBHX_CHn_BAR (d, n) \ (((dma_is_apbh(d) && apbh_is_old(d)) ? 0x070 : 0x130) + (n) * 0x70) +#define HW_APBX_CHn_DEBUG1 (d, n) (0x150 + (n) * 0x70)/* * ccw bits definitions
@@ -204,12 +205,36 @@ static void mxs_dma_reset_chan(struct mxs_dma_chan *mxs_chan)struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;int chan_id = mxs_chan->chan.chan_id; -if (dma_is_apbh (mxs_dma) &&apbh_is_old (mxs_dma)) +if (dma_is_apbh (mxs_dma) &&apbh_is_old (mxs_dma)) {writel (1 << (chan_id +BP_APBH_CTRL0_RESET_CHANNEL ), mxs_dma->base +HW_APBHX_CTRL0 +STMP_OFFSET_REG_SET ); -else + }else { +unsigned long elapsed =0 ; +const unsigned long max_wait =50000 ;/* 50ms */ +void __iomem *reg_dbg1 = mxs_dma->base + +HW_APBX_CHn_DEBUG1 (mxs_dma, chan_id); + +/* +* On i.MX28 APBX, the DMA channel can stop working if we reset +* the channel while it is in READ_FLUSH (0x08) state. +* We wait here until we leave the state. Then we trigger the +* reset. Waiting a maximum of 50ms, the kernel shouldn't crash +* because of this. +*/ +while ((readl (reg_dbg1) &0xf ) ==0x8 && elapsed < max_wait) { +udelay (100 ); + elapsed +=100 ; + } + +if (elapsed >= max_wait) +dev_err (&mxs_chan->mxs_dma->pdev->dev, +"Failed waiting for the DMA channel %d to leave state READ_FLUSH, trying to reset channel in READ_FLUSH state now\n" , + chan_id); + +writel (1 << (chan_id +BP_APBHX_CHANNEL_CTRL_RESET_CHANNEL ), mxs_dma->base +HW_APBHX_CHANNEL_CTRL +STMP_OFFSET_REG_SET ); + } }static void mxs_dma_enable_chan (struct mxs_dma_chan *mxs_chan)
Signature · 1 line
1.8.4.rc3
PATCH v5 4/5 dma: mxs-dma: Update state after channel reset
In reply to Markus PargmannAfter a channel reset, the channel stops running automatically. The state update was missing so that a channel perperation right after a channel reset failed.
--- drivers/dma/mxs-dma.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/dma/mxs-dma.c +3 -1
@@ -235,6 +235,8 @@ static void mxs_dma_reset_chan(struct mxs_dma_chan *mxs_chan)writel (1 << (chan_id +BP_APBHX_CHANNEL_CTRL_RESET_CHANNEL ), mxs_dma->base +HW_APBHX_CHANNEL_CTRL +STMP_OFFSET_REG_SET ); } + + mxs_chan->status =DMA_COMPLETE ; }static void mxs_dma_enable_chan (struct mxs_dma_chan *mxs_chan)
@@ -362,7 +364,7 @@ static irqreturn_t mxs_dma_int_handler(int irq, void *dev_id) chan); mxs_chan->status =DMA_ERROR ;mxs_dma_reset_chan (mxs_chan); - }else { + }else if (mxs_chan->status !=DMA_COMPLETE ) {if (mxs_chan->flags &MXS_DMA_SG_LOOP ) mxs_chan->status =DMA_IN_PROGRESS ;else
Signature · 1 line
1.8.4.rc3
PATCH v5 5/5 dma: mxs-dma: Use semaphores for cyclic DMA
In reply to Markus Pargmannmxs dma channel hardware reset command is not reliable and can cause a channel stall. The only way to fix the channel stall is a DMA engine reset.
To avoid channel resets we use the hardware semaphore counter. For each transmitted segment, the DMA channel will decrease the counter by one. To use this mechanism with cyclic DMA, we need to increase the semaphore counter with each completed DMA command in the interrupt handler. To avoid any interruptions between the DMA transfers, the semaphore counter is initialized with 2. This way the counter can be increased in the interrupt handler without an influence on the transfer of the DMA engine.
When disabling the channel, we stop increasing the semaphore counter in the interrupt handler.
This patch was tested on i.MX28 with the SAIF DMA channel.
--- drivers/dma/mxs-dma.c | 41 +++++++++++++++++++++++++++++++++++------ 1 file changed, 35 insertions(+), 6 deletions(-)
drivers/dma/mxs-dma.c +35 -6
@@ -119,7 +119,9 @@ struct mxs_dma_chan {int desc_count;enum dma_status status;unsigned int flags; +bool reset;#define MXS_DMA_SG_LOOP (1 << 0) +#define MXS_DMA_USE_SEMAPHORE (1 << 1) };#define MXS_DMA_CHANNELS 16
@@ -205,7 +207,17 @@ static void mxs_dma_reset_chan(struct mxs_dma_chan *mxs_chan)struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;int chan_id = mxs_chan->chan.chan_id; -if (dma_is_apbh (mxs_dma) &&apbh_is_old (mxs_dma)) { +/* +* mxs dma channel resets can cause a channel stall. To recover from a +* channel stall, we have to reset the whole DMA engine. To avoid this, +* we use cyclic DMA with semaphores, that are enhanced in +* mxs_dma_int_handler. To reset the channel, we can simply stop writing +* into the semaphore counter. +*/ +if (mxs_chan->flags &MXS_DMA_USE_SEMAPHORE && + mxs_chan->flags &MXS_DMA_SG_LOOP ) { + mxs_chan->reset = true; + }else if (dma_is_apbh (mxs_dma) &&apbh_is_old (mxs_dma)) {writel (1 << (chan_id +BP_APBH_CTRL0_RESET_CHANNEL ), mxs_dma->base +HW_APBHX_CTRL0 +STMP_OFFSET_REG_SET ); }else {
@@ -231,7 +243,6 @@ static void mxs_dma_reset_chan(struct mxs_dma_chan *mxs_chan)"Failed waiting for the DMA channel %d to leave state READ_FLUSH, trying to reset channel in READ_FLUSH state now\n" , chan_id); -writel (1 << (chan_id +BP_APBHX_CHANNEL_CTRL_RESET_CHANNEL ), mxs_dma->base +HW_APBHX_CHANNEL_CTRL +STMP_OFFSET_REG_SET ); }
@@ -249,7 +260,16 @@ static void mxs_dma_enable_chan(struct mxs_dma_chan *mxs_chan) mxs_dma->base + HW_APBHX_CHn_NXTCMDAR(mxs_dma, chan_id));/* write 1 to SEMA to kick off the channel */ -writel (1 , mxs_dma->base +HW_APBHX_CHn_SEMA (mxs_dma, chan_id)); +if (mxs_chan->flags &MXS_DMA_USE_SEMAPHORE && + mxs_chan->flags &MXS_DMA_SG_LOOP ) { +/* A cyclic DMA consists of at least 2 segments, so initialize +* the semaphore with 2 so we have enough time to add 1 to the +* semaphore if we need to */ +writel (2 , mxs_dma->base +HW_APBHX_CHn_SEMA (mxs_dma, chan_id)); + }else { +writel (1 , mxs_dma->base +HW_APBHX_CHn_SEMA (mxs_dma, chan_id)); + } + mxs_chan->reset = false; }static void mxs_dma_disable_chan (struct mxs_dma_chan *mxs_chan)
@@ -365,14 +385,21 @@ static irqreturn_t mxs_dma_int_handler(int irq, void *dev_id) mxs_chan->status =DMA_ERROR ;mxs_dma_reset_chan (mxs_chan); }else if (mxs_chan->status !=DMA_COMPLETE ) { -if (mxs_chan->flags &MXS_DMA_SG_LOOP ) +if (mxs_chan->flags &MXS_DMA_SG_LOOP ) { mxs_chan->status =DMA_IN_PROGRESS ; -else +if (mxs_chan->flags &MXS_DMA_USE_SEMAPHORE ) +writel (1 , mxs_dma->base + +HW_APBHX_CHn_SEMA (mxs_dma, chan)); + }else { mxs_chan->status =DMA_COMPLETE ; + } } -if (mxs_chan->status ==DMA_COMPLETE ) +if (mxs_chan->status ==DMA_COMPLETE ) { +if (mxs_chan->reset) +return IRQ_HANDLED ;dma_cookie_complete (&mxs_chan->desc); + }/* schedule tasklet on this channel */ tasklet_schedule (&mxs_chan->tasklet);
@@ -576,6 +603,7 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_dma_cyclic( mxs_chan->status =DMA_IN_PROGRESS ; mxs_chan->flags |=MXS_DMA_SG_LOOP ; + mxs_chan->flags |=MXS_DMA_USE_SEMAPHORE ;if (num_periods >NUM_CCW ) { dev_err(mxs_dma->dma_device.dev,
@@ -607,6 +635,7 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_dma_cyclic( ccw->bits |=CCW_IRQ ; ccw->bits |=CCW_HALT_ON_TERM ; ccw->bits |=CCW_TERM_FLUSH ; + ccw->bits |=CCW_DEC_SEM ; ccw->bits |=BF_CCW (direction ==DMA_DEV_TO_MEM ?MXS_DMA_CMD_WRITE :MXS_DMA_CMD_READ ,COMMAND );
Signature · 1 line
1.8.4.rc3
v5 cover dma: mxs-dma bugfixes and cleanup
In reply to Markus PargmannQuoted 5 lines · Markus Pargmann
Hi,
v5 is rebased onto your
dma_completebranch. It includes one bugfix.This series contains a number of bugfixes for the mxs DMA engine.
Thanks,
I am going to apply this. But as noted in residue patch I think the restriction of single descriptor is an artficial one and we cna do better in SW. This also bring the abuse the completion status to track the channel status which shouldnt be required/gone once we do multiple descriptors. But those should be handled as another patch series!
--
~Vinod
Quoted 19 lines
Regards,
Markus Pargmann
Markus Pargmann (5):
dma: mxs-dma: Cleanup interrupt handler
dma: mxs-dma: Report correct residue for cyclic DMA
dma: mxs-dma: Fix channel reset hardware bug
dma: mxs-dma: Update state after channel reset
dma: mxs-dma: Use semaphores for cyclic DMAdrivers/dma/mxs-dma.c | 176 +++++++++++++++++++++++++++++++++++++++----------- 1 file changed, 137 insertions(+), 39 deletions(-)Signature · 1 line
1.8.4.rc3