linux_dsm_epyc7002/drivers/dma/omap-dma.c

1664 lines
40 KiB
C
Raw Normal View History

/*
* OMAP DMAengine support
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
#include <linux/delay.h>
#include <linux/dmaengine.h>
#include <linux/dma-mapping.h>
#include <linux/dmapool.h>
#include <linux/err.h>
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/list.h>
#include <linux/module.h>
#include <linux/omap-dma.h>
#include <linux/platform_device.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/of_dma.h>
#include <linux/of_device.h>
#include "virt-dma.h"
#define OMAP_SDMA_REQUESTS 127
#define OMAP_SDMA_CHANNELS 32
struct omap_dmadev {
struct dma_device ddev;
spinlock_t lock;
void __iomem *base;
const struct omap_dma_reg *reg_map;
struct omap_system_dma_plat_info *plat;
bool legacy;
bool ll123_supported;
struct dma_pool *desc_pool;
unsigned dma_requests;
spinlock_t irq_lock;
uint32_t irq_enable_mask;
struct omap_chan **lch_map;
};
struct omap_chan {
struct virt_dma_chan vc;
void __iomem *channel_base;
const struct omap_dma_reg *reg_map;
uint32_t ccr;
struct dma_slave_config cfg;
unsigned dma_sig;
bool cyclic;
bool paused;
bool running;
int dma_ch;
struct omap_desc *desc;
unsigned sgidx;
};
#define DESC_NXT_SV_REFRESH (0x1 << 24)
#define DESC_NXT_SV_REUSE (0x2 << 24)
#define DESC_NXT_DV_REFRESH (0x1 << 26)
#define DESC_NXT_DV_REUSE (0x2 << 26)
#define DESC_NTYPE_TYPE2 (0x2 << 29)
/* Type 2 descriptor with Source or Destination address update */
struct omap_type2_desc {
uint32_t next_desc;
uint32_t en;
uint32_t addr; /* src or dst */
uint16_t fn;
uint16_t cicr;
int16_t cdei;
int16_t csei;
int32_t cdfi;
int32_t csfi;
} __packed;
struct omap_sg {
dma_addr_t addr;
uint32_t en; /* number of elements (24-bit) */
uint32_t fn; /* number of frames (16-bit) */
int32_t fi; /* for double indexing */
int16_t ei; /* for double indexing */
/* Linked list */
struct omap_type2_desc *t2_desc;
dma_addr_t t2_desc_paddr;
};
struct omap_desc {
struct virt_dma_desc vd;
bool using_ll;
enum dma_transfer_direction dir;
dma_addr_t dev_addr;
int32_t fi; /* for OMAP_DMA_SYNC_PACKET / double indexing */
int16_t ei; /* for double indexing */
uint8_t es; /* CSDP_DATA_TYPE_xxx */
uint32_t ccr; /* CCR value */
uint16_t clnk_ctrl; /* CLNK_CTRL value */
uint16_t cicr; /* CICR value */
uint32_t csdp; /* CSDP value */
unsigned sglen;
struct omap_sg sg[0];
};
enum {
CAPS_0_SUPPORT_LL123 = BIT(20), /* Linked List type1/2/3 */
CAPS_0_SUPPORT_LL4 = BIT(21), /* Linked List type4 */
CCR_FS = BIT(5),
CCR_READ_PRIORITY = BIT(6),
CCR_ENABLE = BIT(7),
CCR_AUTO_INIT = BIT(8), /* OMAP1 only */
CCR_REPEAT = BIT(9), /* OMAP1 only */
CCR_OMAP31_DISABLE = BIT(10), /* OMAP1 only */
CCR_SUSPEND_SENSITIVE = BIT(8), /* OMAP2+ only */
CCR_RD_ACTIVE = BIT(9), /* OMAP2+ only */
CCR_WR_ACTIVE = BIT(10), /* OMAP2+ only */
CCR_SRC_AMODE_CONSTANT = 0 << 12,
CCR_SRC_AMODE_POSTINC = 1 << 12,
CCR_SRC_AMODE_SGLIDX = 2 << 12,
CCR_SRC_AMODE_DBLIDX = 3 << 12,
CCR_DST_AMODE_CONSTANT = 0 << 14,
CCR_DST_AMODE_POSTINC = 1 << 14,
CCR_DST_AMODE_SGLIDX = 2 << 14,
CCR_DST_AMODE_DBLIDX = 3 << 14,
CCR_CONSTANT_FILL = BIT(16),
CCR_TRANSPARENT_COPY = BIT(17),
CCR_BS = BIT(18),
CCR_SUPERVISOR = BIT(22),
CCR_PREFETCH = BIT(23),
CCR_TRIGGER_SRC = BIT(24),
CCR_BUFFERING_DISABLE = BIT(25),
CCR_WRITE_PRIORITY = BIT(26),
CCR_SYNC_ELEMENT = 0,
CCR_SYNC_FRAME = CCR_FS,
CCR_SYNC_BLOCK = CCR_BS,
CCR_SYNC_PACKET = CCR_BS | CCR_FS,
CSDP_DATA_TYPE_8 = 0,
CSDP_DATA_TYPE_16 = 1,
CSDP_DATA_TYPE_32 = 2,
CSDP_SRC_PORT_EMIFF = 0 << 2, /* OMAP1 only */
CSDP_SRC_PORT_EMIFS = 1 << 2, /* OMAP1 only */
CSDP_SRC_PORT_OCP_T1 = 2 << 2, /* OMAP1 only */
CSDP_SRC_PORT_TIPB = 3 << 2, /* OMAP1 only */
CSDP_SRC_PORT_OCP_T2 = 4 << 2, /* OMAP1 only */
CSDP_SRC_PORT_MPUI = 5 << 2, /* OMAP1 only */
CSDP_SRC_PACKED = BIT(6),
CSDP_SRC_BURST_1 = 0 << 7,
CSDP_SRC_BURST_16 = 1 << 7,
CSDP_SRC_BURST_32 = 2 << 7,
CSDP_SRC_BURST_64 = 3 << 7,
CSDP_DST_PORT_EMIFF = 0 << 9, /* OMAP1 only */
CSDP_DST_PORT_EMIFS = 1 << 9, /* OMAP1 only */
CSDP_DST_PORT_OCP_T1 = 2 << 9, /* OMAP1 only */
CSDP_DST_PORT_TIPB = 3 << 9, /* OMAP1 only */
CSDP_DST_PORT_OCP_T2 = 4 << 9, /* OMAP1 only */
CSDP_DST_PORT_MPUI = 5 << 9, /* OMAP1 only */
CSDP_DST_PACKED = BIT(13),
CSDP_DST_BURST_1 = 0 << 14,
CSDP_DST_BURST_16 = 1 << 14,
CSDP_DST_BURST_32 = 2 << 14,
CSDP_DST_BURST_64 = 3 << 14,
CSDP_WRITE_NON_POSTED = 0 << 16,
CSDP_WRITE_POSTED = 1 << 16,
CSDP_WRITE_LAST_NON_POSTED = 2 << 16,
CICR_TOUT_IE = BIT(0), /* OMAP1 only */
CICR_DROP_IE = BIT(1),
CICR_HALF_IE = BIT(2),
CICR_FRAME_IE = BIT(3),
CICR_LAST_IE = BIT(4),
CICR_BLOCK_IE = BIT(5),
CICR_PKT_IE = BIT(7), /* OMAP2+ only */
CICR_TRANS_ERR_IE = BIT(8), /* OMAP2+ only */
CICR_SUPERVISOR_ERR_IE = BIT(10), /* OMAP2+ only */
CICR_MISALIGNED_ERR_IE = BIT(11), /* OMAP2+ only */
CICR_DRAIN_IE = BIT(12), /* OMAP2+ only */
CICR_SUPER_BLOCK_IE = BIT(14), /* OMAP2+ only */
CLNK_CTRL_ENABLE_LNK = BIT(15),
CDP_DST_VALID_INC = 0 << 0,
CDP_DST_VALID_RELOAD = 1 << 0,
CDP_DST_VALID_REUSE = 2 << 0,
CDP_SRC_VALID_INC = 0 << 2,
CDP_SRC_VALID_RELOAD = 1 << 2,
CDP_SRC_VALID_REUSE = 2 << 2,
CDP_NTYPE_TYPE1 = 1 << 4,
CDP_NTYPE_TYPE2 = 2 << 4,
CDP_NTYPE_TYPE3 = 3 << 4,
CDP_TMODE_NORMAL = 0 << 8,
CDP_TMODE_LLIST = 1 << 8,
CDP_FAST = BIT(10),
};
static const unsigned es_bytes[] = {
[CSDP_DATA_TYPE_8] = 1,
[CSDP_DATA_TYPE_16] = 2,
[CSDP_DATA_TYPE_32] = 4,
};
static struct of_dma_filter_info omap_dma_info = {
.filter_fn = omap_dma_filter_fn,
};
static inline struct omap_dmadev *to_omap_dma_dev(struct dma_device *d)
{
return container_of(d, struct omap_dmadev, ddev);
}
static inline struct omap_chan *to_omap_dma_chan(struct dma_chan *c)
{
return container_of(c, struct omap_chan, vc.chan);
}
static inline struct omap_desc *to_omap_dma_desc(struct dma_async_tx_descriptor *t)
{
return container_of(t, struct omap_desc, vd.tx);
}
static void omap_dma_desc_free(struct virt_dma_desc *vd)
{
struct omap_desc *d = to_omap_dma_desc(&vd->tx);
if (d->using_ll) {
struct omap_dmadev *od = to_omap_dma_dev(vd->tx.chan->device);
int i;
for (i = 0; i < d->sglen; i++) {
if (d->sg[i].t2_desc)
dma_pool_free(od->desc_pool, d->sg[i].t2_desc,
d->sg[i].t2_desc_paddr);
}
}
kfree(d);
}
static void omap_dma_fill_type2_desc(struct omap_desc *d, int idx,
enum dma_transfer_direction dir, bool last)
{
struct omap_sg *sg = &d->sg[idx];
struct omap_type2_desc *t2_desc = sg->t2_desc;
if (idx)
d->sg[idx - 1].t2_desc->next_desc = sg->t2_desc_paddr;
if (last)
t2_desc->next_desc = 0xfffffffc;
t2_desc->en = sg->en;
t2_desc->addr = sg->addr;
t2_desc->fn = sg->fn & 0xffff;
t2_desc->cicr = d->cicr;
if (!last)
t2_desc->cicr &= ~CICR_BLOCK_IE;
switch (dir) {
case DMA_DEV_TO_MEM:
t2_desc->cdei = sg->ei;
t2_desc->csei = d->ei;
t2_desc->cdfi = sg->fi;
t2_desc->csfi = d->fi;
t2_desc->en |= DESC_NXT_DV_REFRESH;
t2_desc->en |= DESC_NXT_SV_REUSE;
break;
case DMA_MEM_TO_DEV:
t2_desc->cdei = d->ei;
t2_desc->csei = sg->ei;
t2_desc->cdfi = d->fi;
t2_desc->csfi = sg->fi;
t2_desc->en |= DESC_NXT_SV_REFRESH;
t2_desc->en |= DESC_NXT_DV_REUSE;
break;
default:
return;
}
t2_desc->en |= DESC_NTYPE_TYPE2;
}
static void omap_dma_write(uint32_t val, unsigned type, void __iomem *addr)
{
switch (type) {
case OMAP_DMA_REG_16BIT:
writew_relaxed(val, addr);
break;
case OMAP_DMA_REG_2X16BIT:
writew_relaxed(val, addr);
writew_relaxed(val >> 16, addr + 2);
break;
case OMAP_DMA_REG_32BIT:
writel_relaxed(val, addr);
break;
default:
WARN_ON(1);
}
}
static unsigned omap_dma_read(unsigned type, void __iomem *addr)
{
unsigned val;
switch (type) {
case OMAP_DMA_REG_16BIT:
val = readw_relaxed(addr);
break;
case OMAP_DMA_REG_2X16BIT:
val = readw_relaxed(addr);
val |= readw_relaxed(addr + 2) << 16;
break;
case OMAP_DMA_REG_32BIT:
val = readl_relaxed(addr);
break;
default:
WARN_ON(1);
val = 0;
}
return val;
}
static void omap_dma_glbl_write(struct omap_dmadev *od, unsigned reg, unsigned val)
{
const struct omap_dma_reg *r = od->reg_map + reg;
WARN_ON(r->stride);
omap_dma_write(val, r->type, od->base + r->offset);
}
static unsigned omap_dma_glbl_read(struct omap_dmadev *od, unsigned reg)
{
const struct omap_dma_reg *r = od->reg_map + reg;
WARN_ON(r->stride);
return omap_dma_read(r->type, od->base + r->offset);
}
static void omap_dma_chan_write(struct omap_chan *c, unsigned reg, unsigned val)
{
const struct omap_dma_reg *r = c->reg_map + reg;
omap_dma_write(val, r->type, c->channel_base + r->offset);
}
static unsigned omap_dma_chan_read(struct omap_chan *c, unsigned reg)
{
const struct omap_dma_reg *r = c->reg_map + reg;
return omap_dma_read(r->type, c->channel_base + r->offset);
}
static void omap_dma_clear_csr(struct omap_chan *c)
{
if (dma_omap1())
omap_dma_chan_read(c, CSR);
else
omap_dma_chan_write(c, CSR, ~0);
}
static unsigned omap_dma_get_csr(struct omap_chan *c)
{
unsigned val = omap_dma_chan_read(c, CSR);
if (!dma_omap1())
omap_dma_chan_write(c, CSR, val);
return val;
}
static void omap_dma_assign(struct omap_dmadev *od, struct omap_chan *c,
unsigned lch)
{
c->channel_base = od->base + od->plat->channel_stride * lch;
od->lch_map[lch] = c;
}
static void omap_dma_start(struct omap_chan *c, struct omap_desc *d)
{
struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
uint16_t cicr = d->cicr;
if (__dma_omap15xx(od->plat->dma_attr))
omap_dma_chan_write(c, CPC, 0);
else
omap_dma_chan_write(c, CDAC, 0);
omap_dma_clear_csr(c);
if (d->using_ll) {
uint32_t cdp = CDP_TMODE_LLIST | CDP_NTYPE_TYPE2 | CDP_FAST;
if (d->dir == DMA_DEV_TO_MEM)
cdp |= (CDP_DST_VALID_RELOAD | CDP_SRC_VALID_REUSE);
else
cdp |= (CDP_DST_VALID_REUSE | CDP_SRC_VALID_RELOAD);
omap_dma_chan_write(c, CDP, cdp);
omap_dma_chan_write(c, CNDP, d->sg[0].t2_desc_paddr);
omap_dma_chan_write(c, CCDN, 0);
omap_dma_chan_write(c, CCFN, 0xffff);
omap_dma_chan_write(c, CCEN, 0xffffff);
cicr &= ~CICR_BLOCK_IE;
} else if (od->ll123_supported) {
omap_dma_chan_write(c, CDP, 0);
}
/* Enable interrupts */
omap_dma_chan_write(c, CICR, cicr);
/* Enable channel */
omap_dma_chan_write(c, CCR, d->ccr | CCR_ENABLE);
c->running = true;
}
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
static void omap_dma_drain_chan(struct omap_chan *c)
{
int i;
u32 val;
/* Wait for sDMA FIFO to drain */
for (i = 0; ; i++) {
val = omap_dma_chan_read(c, CCR);
if (!(val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE)))
break;
if (i > 100)
break;
udelay(5);
}
if (val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE))
dev_err(c->vc.chan.device->dev,
"DMA drain did not complete on lch %d\n",
c->dma_ch);
}
static int omap_dma_stop(struct omap_chan *c)
{
struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
uint32_t val;
/* disable irq */
omap_dma_chan_write(c, CICR, 0);
omap_dma_clear_csr(c);
val = omap_dma_chan_read(c, CCR);
if (od->plat->errata & DMA_ERRATA_i541 && val & CCR_TRIGGER_SRC) {
uint32_t sysconfig;
sysconfig = omap_dma_glbl_read(od, OCP_SYSCONFIG);
val = sysconfig & ~DMA_SYSCONFIG_MIDLEMODE_MASK;
val |= DMA_SYSCONFIG_MIDLEMODE(DMA_IDLEMODE_NO_IDLE);
omap_dma_glbl_write(od, OCP_SYSCONFIG, val);
val = omap_dma_chan_read(c, CCR);
val &= ~CCR_ENABLE;
omap_dma_chan_write(c, CCR, val);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
if (!(c->ccr & CCR_BUFFERING_DISABLE))
omap_dma_drain_chan(c);
omap_dma_glbl_write(od, OCP_SYSCONFIG, sysconfig);
} else {
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
if (!(val & CCR_ENABLE))
return -EINVAL;
val &= ~CCR_ENABLE;
omap_dma_chan_write(c, CCR, val);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
if (!(c->ccr & CCR_BUFFERING_DISABLE))
omap_dma_drain_chan(c);
}
mb();
if (!__dma_omap15xx(od->plat->dma_attr) && c->cyclic) {
val = omap_dma_chan_read(c, CLNK_CTRL);
if (dma_omap1())
val |= 1 << 14; /* set the STOP_LNK bit */
else
val &= ~CLNK_CTRL_ENABLE_LNK;
omap_dma_chan_write(c, CLNK_CTRL, val);
}
c->running = false;
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
return 0;
}
static void omap_dma_start_sg(struct omap_chan *c, struct omap_desc *d)
{
struct omap_sg *sg = d->sg + c->sgidx;
unsigned cxsa, cxei, cxfi;
if (d->dir == DMA_DEV_TO_MEM || d->dir == DMA_MEM_TO_MEM) {
cxsa = CDSA;
cxei = CDEI;
cxfi = CDFI;
} else {
cxsa = CSSA;
cxei = CSEI;
cxfi = CSFI;
}
omap_dma_chan_write(c, cxsa, sg->addr);
omap_dma_chan_write(c, cxei, sg->ei);
omap_dma_chan_write(c, cxfi, sg->fi);
omap_dma_chan_write(c, CEN, sg->en);
omap_dma_chan_write(c, CFN, sg->fn);
omap_dma_start(c, d);
c->sgidx++;
}
static void omap_dma_start_desc(struct omap_chan *c)
{
struct virt_dma_desc *vd = vchan_next_desc(&c->vc);
struct omap_desc *d;
unsigned cxsa, cxei, cxfi;
if (!vd) {
c->desc = NULL;
return;
}
list_del(&vd->node);
c->desc = d = to_omap_dma_desc(&vd->tx);
c->sgidx = 0;
/*
* This provides the necessary barrier to ensure data held in
* DMA coherent memory is visible to the DMA engine prior to
* the transfer starting.
*/
mb();
omap_dma_chan_write(c, CCR, d->ccr);
if (dma_omap1())
omap_dma_chan_write(c, CCR2, d->ccr >> 16);
if (d->dir == DMA_DEV_TO_MEM || d->dir == DMA_MEM_TO_MEM) {
cxsa = CSSA;
cxei = CSEI;
cxfi = CSFI;
} else {
cxsa = CDSA;
cxei = CDEI;
cxfi = CDFI;
}
omap_dma_chan_write(c, cxsa, d->dev_addr);
omap_dma_chan_write(c, cxei, d->ei);
omap_dma_chan_write(c, cxfi, d->fi);
omap_dma_chan_write(c, CSDP, d->csdp);
omap_dma_chan_write(c, CLNK_CTRL, d->clnk_ctrl);
omap_dma_start_sg(c, d);
}
static void omap_dma_callback(int ch, u16 status, void *data)
{
struct omap_chan *c = data;
struct omap_desc *d;
unsigned long flags;
spin_lock_irqsave(&c->vc.lock, flags);
d = c->desc;
if (d) {
if (c->cyclic) {
vchan_cyclic_callback(&d->vd);
} else if (d->using_ll || c->sgidx == d->sglen) {
omap_dma_start_desc(c);
vchan_cookie_complete(&d->vd);
} else {
omap_dma_start_sg(c, d);
}
}
spin_unlock_irqrestore(&c->vc.lock, flags);
}
static irqreturn_t omap_dma_irq(int irq, void *devid)
{
struct omap_dmadev *od = devid;
unsigned status, channel;
spin_lock(&od->irq_lock);
status = omap_dma_glbl_read(od, IRQSTATUS_L1);
status &= od->irq_enable_mask;
if (status == 0) {
spin_unlock(&od->irq_lock);
return IRQ_NONE;
}
while ((channel = ffs(status)) != 0) {
unsigned mask, csr;
struct omap_chan *c;
channel -= 1;
mask = BIT(channel);
status &= ~mask;
c = od->lch_map[channel];
if (c == NULL) {
/* This should never happen */
dev_err(od->ddev.dev, "invalid channel %u\n", channel);
continue;
}
csr = omap_dma_get_csr(c);
omap_dma_glbl_write(od, IRQSTATUS_L1, mask);
omap_dma_callback(channel, csr, c);
}
spin_unlock(&od->irq_lock);
return IRQ_HANDLED;
}
static int omap_dma_alloc_chan_resources(struct dma_chan *chan)
{
struct omap_dmadev *od = to_omap_dma_dev(chan->device);
struct omap_chan *c = to_omap_dma_chan(chan);
struct device *dev = od->ddev.dev;
int ret;
if (od->legacy) {
ret = omap_request_dma(c->dma_sig, "DMA engine",
omap_dma_callback, c, &c->dma_ch);
} else {
ret = omap_request_dma(c->dma_sig, "DMA engine", NULL, NULL,
&c->dma_ch);
}
dev_dbg(dev, "allocating channel %u for %u\n", c->dma_ch, c->dma_sig);
if (ret >= 0) {
omap_dma_assign(od, c, c->dma_ch);
if (!od->legacy) {
unsigned val;
spin_lock_irq(&od->irq_lock);
val = BIT(c->dma_ch);
omap_dma_glbl_write(od, IRQSTATUS_L1, val);
od->irq_enable_mask |= val;
omap_dma_glbl_write(od, IRQENABLE_L1, od->irq_enable_mask);
val = omap_dma_glbl_read(od, IRQENABLE_L0);
val &= ~BIT(c->dma_ch);
omap_dma_glbl_write(od, IRQENABLE_L0, val);
spin_unlock_irq(&od->irq_lock);
}
}
if (dma_omap1()) {
if (__dma_omap16xx(od->plat->dma_attr)) {
c->ccr = CCR_OMAP31_DISABLE;
/* Duplicate what plat-omap/dma.c does */
c->ccr |= c->dma_ch + 1;
} else {
c->ccr = c->dma_sig & 0x1f;
}
} else {
c->ccr = c->dma_sig & 0x1f;
c->ccr |= (c->dma_sig & ~0x1f) << 14;
}
if (od->plat->errata & DMA_ERRATA_IFRAME_BUFFERING)
c->ccr |= CCR_BUFFERING_DISABLE;
return ret;
}
static void omap_dma_free_chan_resources(struct dma_chan *chan)
{
struct omap_dmadev *od = to_omap_dma_dev(chan->device);
struct omap_chan *c = to_omap_dma_chan(chan);
if (!od->legacy) {
spin_lock_irq(&od->irq_lock);
od->irq_enable_mask &= ~BIT(c->dma_ch);
omap_dma_glbl_write(od, IRQENABLE_L1, od->irq_enable_mask);
spin_unlock_irq(&od->irq_lock);
}
c->channel_base = NULL;
od->lch_map[c->dma_ch] = NULL;
vchan_free_chan_resources(&c->vc);
omap_free_dma(c->dma_ch);
dev_dbg(od->ddev.dev, "freeing channel %u used for %u\n", c->dma_ch,
c->dma_sig);
c->dma_sig = 0;
}
static size_t omap_dma_sg_size(struct omap_sg *sg)
{
return sg->en * sg->fn;
}
static size_t omap_dma_desc_size(struct omap_desc *d)
{
unsigned i;
size_t size;
for (size = i = 0; i < d->sglen; i++)
size += omap_dma_sg_size(&d->sg[i]);
return size * es_bytes[d->es];
}
static size_t omap_dma_desc_size_pos(struct omap_desc *d, dma_addr_t addr)
{
unsigned i;
size_t size, es_size = es_bytes[d->es];
for (size = i = 0; i < d->sglen; i++) {
size_t this_size = omap_dma_sg_size(&d->sg[i]) * es_size;
if (size)
size += this_size;
else if (addr >= d->sg[i].addr &&
addr < d->sg[i].addr + this_size)
size += d->sg[i].addr + this_size - addr;
}
return size;
}
/*
* OMAP 3.2/3.3 erratum: sometimes 0 is returned if CSAC/CDAC is
* read before the DMA controller finished disabling the channel.
*/
static uint32_t omap_dma_chan_read_3_3(struct omap_chan *c, unsigned reg)
{
struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
uint32_t val;
val = omap_dma_chan_read(c, reg);
if (val == 0 && od->plat->errata & DMA_ERRATA_3_3)
val = omap_dma_chan_read(c, reg);
return val;
}
static dma_addr_t omap_dma_get_src_pos(struct omap_chan *c)
{
struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
dma_addr_t addr, cdac;
if (__dma_omap15xx(od->plat->dma_attr)) {
addr = omap_dma_chan_read(c, CPC);
} else {
addr = omap_dma_chan_read_3_3(c, CSAC);
cdac = omap_dma_chan_read_3_3(c, CDAC);
/*
* CDAC == 0 indicates that the DMA transfer on the channel has
* not been started (no data has been transferred so far).
* Return the programmed source start address in this case.
*/
if (cdac == 0)
addr = omap_dma_chan_read(c, CSSA);
}
if (dma_omap1())
addr |= omap_dma_chan_read(c, CSSA) & 0xffff0000;
return addr;
}
static dma_addr_t omap_dma_get_dst_pos(struct omap_chan *c)
{
struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
dma_addr_t addr;
if (__dma_omap15xx(od->plat->dma_attr)) {
addr = omap_dma_chan_read(c, CPC);
} else {
addr = omap_dma_chan_read_3_3(c, CDAC);
/*
* CDAC == 0 indicates that the DMA transfer on the channel
* has not been started (no data has been transferred so
* far). Return the programmed destination start address in
* this case.
*/
if (addr == 0)
addr = omap_dma_chan_read(c, CDSA);
}
if (dma_omap1())
addr |= omap_dma_chan_read(c, CDSA) & 0xffff0000;
return addr;
}
static enum dma_status omap_dma_tx_status(struct dma_chan *chan,
dma_cookie_t cookie, struct dma_tx_state *txstate)
{
struct omap_chan *c = to_omap_dma_chan(chan);
struct virt_dma_desc *vd;
enum dma_status ret;
unsigned long flags;
ret = dma_cookie_status(chan, cookie, txstate);
if (!c->paused && c->running) {
uint32_t ccr = omap_dma_chan_read(c, CCR);
/*
* The channel is no longer active, set the return value
* accordingly
*/
if (!(ccr & CCR_ENABLE))
ret = DMA_COMPLETE;
}
if (ret == DMA_COMPLETE || !txstate)
return ret;
spin_lock_irqsave(&c->vc.lock, flags);
vd = vchan_find_desc(&c->vc, cookie);
if (vd) {
txstate->residue = omap_dma_desc_size(to_omap_dma_desc(&vd->tx));
} else if (c->desc && c->desc->vd.tx.cookie == cookie) {
struct omap_desc *d = c->desc;
dma_addr_t pos;
if (d->dir == DMA_MEM_TO_DEV)
pos = omap_dma_get_src_pos(c);
else if (d->dir == DMA_DEV_TO_MEM || d->dir == DMA_MEM_TO_MEM)
pos = omap_dma_get_dst_pos(c);
else
pos = 0;
txstate->residue = omap_dma_desc_size_pos(d, pos);
} else {
txstate->residue = 0;
}
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
if (ret == DMA_IN_PROGRESS && c->paused)
ret = DMA_PAUSED;
spin_unlock_irqrestore(&c->vc.lock, flags);
return ret;
}
static void omap_dma_issue_pending(struct dma_chan *chan)
{
struct omap_chan *c = to_omap_dma_chan(chan);
unsigned long flags;
spin_lock_irqsave(&c->vc.lock, flags);
if (vchan_issue_pending(&c->vc) && !c->desc)
omap_dma_start_desc(c);
spin_unlock_irqrestore(&c->vc.lock, flags);
}
static struct dma_async_tx_descriptor *omap_dma_prep_slave_sg(
struct dma_chan *chan, struct scatterlist *sgl, unsigned sglen,
enum dma_transfer_direction dir, unsigned long tx_flags, void *context)
{
struct omap_dmadev *od = to_omap_dma_dev(chan->device);
struct omap_chan *c = to_omap_dma_chan(chan);
enum dma_slave_buswidth dev_width;
struct scatterlist *sgent;
struct omap_desc *d;
dma_addr_t dev_addr;
unsigned i, es, en, frame_bytes;
bool ll_failed = false;
u32 burst;
u32 port_window, port_window_bytes;
if (dir == DMA_DEV_TO_MEM) {
dev_addr = c->cfg.src_addr;
dev_width = c->cfg.src_addr_width;
burst = c->cfg.src_maxburst;
port_window = c->cfg.src_port_window_size;
} else if (dir == DMA_MEM_TO_DEV) {
dev_addr = c->cfg.dst_addr;
dev_width = c->cfg.dst_addr_width;
burst = c->cfg.dst_maxburst;
port_window = c->cfg.dst_port_window_size;
} else {
dev_err(chan->device->dev, "%s: bad direction?\n", __func__);
return NULL;
}
/* Bus width translates to the element size (ES) */
switch (dev_width) {
case DMA_SLAVE_BUSWIDTH_1_BYTE:
es = CSDP_DATA_TYPE_8;
break;
case DMA_SLAVE_BUSWIDTH_2_BYTES:
es = CSDP_DATA_TYPE_16;
break;
case DMA_SLAVE_BUSWIDTH_4_BYTES:
es = CSDP_DATA_TYPE_32;
break;
default: /* not reached */
return NULL;
}
/* Now allocate and setup the descriptor. */
d = kzalloc(sizeof(*d) + sglen * sizeof(d->sg[0]), GFP_ATOMIC);
if (!d)
return NULL;
d->dir = dir;
d->dev_addr = dev_addr;
d->es = es;
/* When the port_window is used, one frame must cover the window */
if (port_window) {
burst = port_window;
port_window_bytes = port_window * es_bytes[es];
d->ei = 1;
/*
* One frame covers the port_window and by configure
* the source frame index to be -1 * (port_window - 1)
* we instruct the sDMA that after a frame is processed
* it should move back to the start of the window.
*/
d->fi = -(port_window_bytes - 1);
}
d->ccr = c->ccr | CCR_SYNC_FRAME;
if (dir == DMA_DEV_TO_MEM) {
d->csdp = CSDP_DST_BURST_64 | CSDP_DST_PACKED;
d->ccr |= CCR_DST_AMODE_POSTINC;
if (port_window) {
d->ccr |= CCR_SRC_AMODE_DBLIDX;
if (port_window_bytes >= 64)
d->csdp |= CSDP_SRC_BURST_64;
else if (port_window_bytes >= 32)
d->csdp |= CSDP_SRC_BURST_32;
else if (port_window_bytes >= 16)
d->csdp |= CSDP_SRC_BURST_16;
} else {
d->ccr |= CCR_SRC_AMODE_CONSTANT;
}
} else {
d->csdp = CSDP_SRC_BURST_64 | CSDP_SRC_PACKED;
d->ccr |= CCR_SRC_AMODE_POSTINC;
if (port_window) {
d->ccr |= CCR_DST_AMODE_DBLIDX;
if (port_window_bytes >= 64)
d->csdp |= CSDP_DST_BURST_64;
else if (port_window_bytes >= 32)
d->csdp |= CSDP_DST_BURST_32;
else if (port_window_bytes >= 16)
d->csdp |= CSDP_DST_BURST_16;
} else {
d->ccr |= CCR_DST_AMODE_CONSTANT;
}
}
d->cicr = CICR_DROP_IE | CICR_BLOCK_IE;
d->csdp |= es;
if (dma_omap1()) {
d->cicr |= CICR_TOUT_IE;
if (dir == DMA_DEV_TO_MEM)
d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_TIPB;
else
d->csdp |= CSDP_DST_PORT_TIPB | CSDP_SRC_PORT_EMIFF;
} else {
if (dir == DMA_DEV_TO_MEM)
d->ccr |= CCR_TRIGGER_SRC;
d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
if (port_window)
d->csdp |= CSDP_WRITE_LAST_NON_POSTED;
}
if (od->plat->errata & DMA_ERRATA_PARALLEL_CHANNELS)
d->clnk_ctrl = c->dma_ch;
/*
* Build our scatterlist entries: each contains the address,
* the number of elements (EN) in each frame, and the number of
* frames (FN). Number of bytes for this entry = ES * EN * FN.
*
* Burst size translates to number of elements with frame sync.
* Note: DMA engine defines burst to be the number of dev-width
* transfers.
*/
en = burst;
frame_bytes = es_bytes[es] * en;
if (sglen >= 2)
d->using_ll = od->ll123_supported;
for_each_sg(sgl, sgent, sglen, i) {
struct omap_sg *osg = &d->sg[i];
osg->addr = sg_dma_address(sgent);
osg->en = en;
osg->fn = sg_dma_len(sgent) / frame_bytes;
if (d->using_ll) {
osg->t2_desc = dma_pool_alloc(od->desc_pool, GFP_ATOMIC,
&osg->t2_desc_paddr);
if (!osg->t2_desc) {
dev_err(chan->device->dev,
"t2_desc[%d] allocation failed\n", i);
ll_failed = true;
d->using_ll = false;
continue;
}
omap_dma_fill_type2_desc(d, i, dir, (i == sglen - 1));
}
}
d->sglen = sglen;
/* Release the dma_pool entries if one allocation failed */
if (ll_failed) {
for (i = 0; i < d->sglen; i++) {
struct omap_sg *osg = &d->sg[i];
if (osg->t2_desc) {
dma_pool_free(od->desc_pool, osg->t2_desc,
osg->t2_desc_paddr);
osg->t2_desc = NULL;
}
}
}
return vchan_tx_prep(&c->vc, &d->vd, tx_flags);
}
static struct dma_async_tx_descriptor *omap_dma_prep_dma_cyclic(
struct dma_chan *chan, dma_addr_t buf_addr, size_t buf_len,
size_t period_len, enum dma_transfer_direction dir, unsigned long flags)
{
struct omap_dmadev *od = to_omap_dma_dev(chan->device);
struct omap_chan *c = to_omap_dma_chan(chan);
enum dma_slave_buswidth dev_width;
struct omap_desc *d;
dma_addr_t dev_addr;
unsigned es;
u32 burst;
if (dir == DMA_DEV_TO_MEM) {
dev_addr = c->cfg.src_addr;
dev_width = c->cfg.src_addr_width;
burst = c->cfg.src_maxburst;
} else if (dir == DMA_MEM_TO_DEV) {
dev_addr = c->cfg.dst_addr;
dev_width = c->cfg.dst_addr_width;
burst = c->cfg.dst_maxburst;
} else {
dev_err(chan->device->dev, "%s: bad direction?\n", __func__);
return NULL;
}
/* Bus width translates to the element size (ES) */
switch (dev_width) {
case DMA_SLAVE_BUSWIDTH_1_BYTE:
es = CSDP_DATA_TYPE_8;
break;
case DMA_SLAVE_BUSWIDTH_2_BYTES:
es = CSDP_DATA_TYPE_16;
break;
case DMA_SLAVE_BUSWIDTH_4_BYTES:
es = CSDP_DATA_TYPE_32;
break;
default: /* not reached */
return NULL;
}
/* Now allocate and setup the descriptor. */
d = kzalloc(sizeof(*d) + sizeof(d->sg[0]), GFP_ATOMIC);
if (!d)
return NULL;
d->dir = dir;
d->dev_addr = dev_addr;
d->fi = burst;
d->es = es;
d->sg[0].addr = buf_addr;
d->sg[0].en = period_len / es_bytes[es];
d->sg[0].fn = buf_len / period_len;
d->sglen = 1;
d->ccr = c->ccr;
if (dir == DMA_DEV_TO_MEM)
d->ccr |= CCR_DST_AMODE_POSTINC | CCR_SRC_AMODE_CONSTANT;
else
d->ccr |= CCR_DST_AMODE_CONSTANT | CCR_SRC_AMODE_POSTINC;
d->cicr = CICR_DROP_IE;
if (flags & DMA_PREP_INTERRUPT)
d->cicr |= CICR_FRAME_IE;
d->csdp = es;
if (dma_omap1()) {
d->cicr |= CICR_TOUT_IE;
if (dir == DMA_DEV_TO_MEM)
d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_MPUI;
else
d->csdp |= CSDP_DST_PORT_MPUI | CSDP_SRC_PORT_EMIFF;
} else {
if (burst)
d->ccr |= CCR_SYNC_PACKET;
else
d->ccr |= CCR_SYNC_ELEMENT;
dmaengine: omap-dma: Enable packed accesses for cyclic transfers The L3 throughput can be higher than expected when packed access is not enabled. The ratio depends on the number of bytes in a transaction and the EMIF interface width. The throughput was measured for the following settings/cases: * Case 1: Burst size of 64 bytes, packed access disabled * Case 2: Burst size of 64 bytes, packed access enabled * Case 3: Burst disabled, packed access disabled Throughput measurements were done during McASP-based audio playback on the Jacinto6 EVM using the omapconf tool [1]: $ omapconf trace bw -m sdma_rd --------------------------------------------------------- Throughput (MB/s) Audio parameters Case 1 Case 2 Case 3 --------------------------------------------------------- 44.1kHz, 16-bits, stereo 1.41 0.18 1.41 44.1kHz, 32-bits, stereo 1.41 0.35 1.41 44.1kHz, 16-bits, 4-chan 2.82 0.35 2.82 44.1kHz, 16-bits, 6-chan 4.23 0.53 4.23 44.1kHz, 16-bits, 8-chan 5.64 0.71 5.64 --------------------------------------------------------- From above measurements, case 2 is the only one that delivers the expected throughput for the given audio parameters. For that reason, the packed accesses are now enabled. It's worth to mention that packed accesses cannot be enabled for all addressing modes. In cyclic transfers, it can be enabled in the source for MEM_TO_DEV and in dest for DEV_TO_MEM, as they use post-increment mode which supports packed accesses. Peter Ujfalusi: From the TRM regarding to this: "NOTE: Except in the constant addressing mode, the source or destination must be specified as packed for burst transactions to occur." So w/o the packed setting the burst on the MEM side was not enabled, this explains the numbers. [1] https://github.com/omapconf/omapconf Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2015-09-14 19:31:05 +07:00
if (dir == DMA_DEV_TO_MEM) {
d->ccr |= CCR_TRIGGER_SRC;
dmaengine: omap-dma: Enable packed accesses for cyclic transfers The L3 throughput can be higher than expected when packed access is not enabled. The ratio depends on the number of bytes in a transaction and the EMIF interface width. The throughput was measured for the following settings/cases: * Case 1: Burst size of 64 bytes, packed access disabled * Case 2: Burst size of 64 bytes, packed access enabled * Case 3: Burst disabled, packed access disabled Throughput measurements were done during McASP-based audio playback on the Jacinto6 EVM using the omapconf tool [1]: $ omapconf trace bw -m sdma_rd --------------------------------------------------------- Throughput (MB/s) Audio parameters Case 1 Case 2 Case 3 --------------------------------------------------------- 44.1kHz, 16-bits, stereo 1.41 0.18 1.41 44.1kHz, 32-bits, stereo 1.41 0.35 1.41 44.1kHz, 16-bits, 4-chan 2.82 0.35 2.82 44.1kHz, 16-bits, 6-chan 4.23 0.53 4.23 44.1kHz, 16-bits, 8-chan 5.64 0.71 5.64 --------------------------------------------------------- From above measurements, case 2 is the only one that delivers the expected throughput for the given audio parameters. For that reason, the packed accesses are now enabled. It's worth to mention that packed accesses cannot be enabled for all addressing modes. In cyclic transfers, it can be enabled in the source for MEM_TO_DEV and in dest for DEV_TO_MEM, as they use post-increment mode which supports packed accesses. Peter Ujfalusi: From the TRM regarding to this: "NOTE: Except in the constant addressing mode, the source or destination must be specified as packed for burst transactions to occur." So w/o the packed setting the burst on the MEM side was not enabled, this explains the numbers. [1] https://github.com/omapconf/omapconf Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2015-09-14 19:31:05 +07:00
d->csdp |= CSDP_DST_PACKED;
} else {
d->csdp |= CSDP_SRC_PACKED;
}
d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
d->csdp |= CSDP_DST_BURST_64 | CSDP_SRC_BURST_64;
}
if (__dma_omap15xx(od->plat->dma_attr))
d->ccr |= CCR_AUTO_INIT | CCR_REPEAT;
else
d->clnk_ctrl = c->dma_ch | CLNK_CTRL_ENABLE_LNK;
c->cyclic = true;
return vchan_tx_prep(&c->vc, &d->vd, flags);
}
static struct dma_async_tx_descriptor *omap_dma_prep_dma_memcpy(
struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
size_t len, unsigned long tx_flags)
{
struct omap_chan *c = to_omap_dma_chan(chan);
struct omap_desc *d;
uint8_t data_type;
d = kzalloc(sizeof(*d) + sizeof(d->sg[0]), GFP_ATOMIC);
if (!d)
return NULL;
data_type = __ffs((src | dest | len));
if (data_type > CSDP_DATA_TYPE_32)
data_type = CSDP_DATA_TYPE_32;
d->dir = DMA_MEM_TO_MEM;
d->dev_addr = src;
d->fi = 0;
d->es = data_type;
d->sg[0].en = len / BIT(data_type);
d->sg[0].fn = 1;
d->sg[0].addr = dest;
d->sglen = 1;
d->ccr = c->ccr;
d->ccr |= CCR_DST_AMODE_POSTINC | CCR_SRC_AMODE_POSTINC;
d->cicr = CICR_DROP_IE | CICR_FRAME_IE;
d->csdp = data_type;
if (dma_omap1()) {
d->cicr |= CICR_TOUT_IE;
d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_EMIFF;
} else {
d->csdp |= CSDP_DST_PACKED | CSDP_SRC_PACKED;
d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
d->csdp |= CSDP_DST_BURST_64 | CSDP_SRC_BURST_64;
}
return vchan_tx_prep(&c->vc, &d->vd, tx_flags);
}
static struct dma_async_tx_descriptor *omap_dma_prep_dma_interleaved(
struct dma_chan *chan, struct dma_interleaved_template *xt,
unsigned long flags)
{
struct omap_chan *c = to_omap_dma_chan(chan);
struct omap_desc *d;
struct omap_sg *sg;
uint8_t data_type;
size_t src_icg, dst_icg;
/* Slave mode is not supported */
if (is_slave_direction(xt->dir))
return NULL;
if (xt->frame_size != 1 || xt->numf == 0)
return NULL;
d = kzalloc(sizeof(*d) + sizeof(d->sg[0]), GFP_ATOMIC);
if (!d)
return NULL;
data_type = __ffs((xt->src_start | xt->dst_start | xt->sgl[0].size));
if (data_type > CSDP_DATA_TYPE_32)
data_type = CSDP_DATA_TYPE_32;
sg = &d->sg[0];
d->dir = DMA_MEM_TO_MEM;
d->dev_addr = xt->src_start;
d->es = data_type;
sg->en = xt->sgl[0].size / BIT(data_type);
sg->fn = xt->numf;
sg->addr = xt->dst_start;
d->sglen = 1;
d->ccr = c->ccr;
src_icg = dmaengine_get_src_icg(xt, &xt->sgl[0]);
dst_icg = dmaengine_get_dst_icg(xt, &xt->sgl[0]);
if (src_icg) {
d->ccr |= CCR_SRC_AMODE_DBLIDX;
d->ei = 1;
d->fi = src_icg;
} else if (xt->src_inc) {
d->ccr |= CCR_SRC_AMODE_POSTINC;
d->fi = 0;
} else {
dev_err(chan->device->dev,
"%s: SRC constant addressing is not supported\n",
__func__);
kfree(d);
return NULL;
}
if (dst_icg) {
d->ccr |= CCR_DST_AMODE_DBLIDX;
sg->ei = 1;
sg->fi = dst_icg;
} else if (xt->dst_inc) {
d->ccr |= CCR_DST_AMODE_POSTINC;
sg->fi = 0;
} else {
dev_err(chan->device->dev,
"%s: DST constant addressing is not supported\n",
__func__);
kfree(d);
return NULL;
}
d->cicr = CICR_DROP_IE | CICR_FRAME_IE;
d->csdp = data_type;
if (dma_omap1()) {
d->cicr |= CICR_TOUT_IE;
d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_EMIFF;
} else {
d->csdp |= CSDP_DST_PACKED | CSDP_SRC_PACKED;
d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
d->csdp |= CSDP_DST_BURST_64 | CSDP_SRC_BURST_64;
}
return vchan_tx_prep(&c->vc, &d->vd, flags);
}
static int omap_dma_slave_config(struct dma_chan *chan, struct dma_slave_config *cfg)
{
struct omap_chan *c = to_omap_dma_chan(chan);
if (cfg->src_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES ||
cfg->dst_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES)
return -EINVAL;
memcpy(&c->cfg, cfg, sizeof(c->cfg));
return 0;
}
static int omap_dma_terminate_all(struct dma_chan *chan)
{
struct omap_chan *c = to_omap_dma_chan(chan);
unsigned long flags;
LIST_HEAD(head);
spin_lock_irqsave(&c->vc.lock, flags);
/*
* Stop DMA activity: we assume the callback will not be called
* after omap_dma_stop() returns (even if it does, it will see
* c->desc is NULL and exit.)
*/
if (c->desc) {
omap_dma_desc_free(&c->desc->vd);
c->desc = NULL;
/* Avoid stopping the dma twice */
if (!c->paused)
omap_dma_stop(c);
}
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
c->cyclic = false;
c->paused = false;
vchan_get_all_descriptors(&c->vc, &head);
spin_unlock_irqrestore(&c->vc.lock, flags);
vchan_dma_desc_free_list(&c->vc, &head);
return 0;
}
static void omap_dma_synchronize(struct dma_chan *chan)
{
struct omap_chan *c = to_omap_dma_chan(chan);
vchan_synchronize(&c->vc);
}
static int omap_dma_pause(struct dma_chan *chan)
{
struct omap_chan *c = to_omap_dma_chan(chan);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
struct omap_dmadev *od = to_omap_dma_dev(chan->device);
unsigned long flags;
int ret = -EINVAL;
bool can_pause = false;
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
spin_lock_irqsave(&od->irq_lock, flags);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
if (!c->desc)
goto out;
if (c->cyclic)
can_pause = true;
/*
* We do not allow DMA_MEM_TO_DEV transfers to be paused.
* From the AM572x TRM, 16.1.4.18 Disabling a Channel During Transfer:
* "When a channel is disabled during a transfer, the channel undergoes
* an abort, unless it is hardware-source-synchronized ".
* A source-synchronised channel is one where the fetching of data is
* under control of the device. In other words, a device-to-memory
* transfer. So, a destination-synchronised channel (which would be a
* memory-to-device transfer) undergoes an abort if the the CCR_ENABLE
* bit is cleared.
* From 16.1.4.20.4.6.2 Abort: "If an abort trigger occurs, the channel
* aborts immediately after completion of current read/write
* transactions and then the FIFO is cleaned up." The term "cleaned up"
* is not defined. TI recommends to check that RD_ACTIVE and WR_ACTIVE
* are both clear _before_ disabling the channel, otherwise data loss
* will occur.
* The problem is that if the channel is active, then device activity
* can result in DMA activity starting between reading those as both
* clear and the write to DMA_CCR to clear the enable bit hitting the
* hardware. If the DMA hardware can't drain the data in its FIFO to the
* destination, then data loss "might" occur (say if we write to an UART
* and the UART is not accepting any further data).
*/
else if (c->desc->dir == DMA_DEV_TO_MEM)
can_pause = true;
if (can_pause && !c->paused) {
ret = omap_dma_stop(c);
if (!ret)
c->paused = true;
}
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
out:
spin_unlock_irqrestore(&od->irq_lock, flags);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
return ret;
}
static int omap_dma_resume(struct dma_chan *chan)
{
struct omap_chan *c = to_omap_dma_chan(chan);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
struct omap_dmadev *od = to_omap_dma_dev(chan->device);
unsigned long flags;
int ret = -EINVAL;
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
spin_lock_irqsave(&od->irq_lock, flags);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
if (c->paused && c->desc) {
mb();
/* Restore channel link register */
omap_dma_chan_write(c, CLNK_CTRL, c->desc->clnk_ctrl);
omap_dma_start(c, c->desc);
c->paused = false;
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
ret = 0;
}
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
spin_unlock_irqrestore(&od->irq_lock, flags);
dmaengine: omap-dma: add support for pause of non-cyclic transfers This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R <vigneshr@ti.com> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
2016-10-14 12:00:47 +07:00
return ret;
}
static int omap_dma_chan_init(struct omap_dmadev *od)
{
struct omap_chan *c;
c = kzalloc(sizeof(*c), GFP_KERNEL);
if (!c)
return -ENOMEM;
c->reg_map = od->reg_map;
c->vc.desc_free = omap_dma_desc_free;
vchan_init(&c->vc, &od->ddev);
return 0;
}
static void omap_dma_free(struct omap_dmadev *od)
{
while (!list_empty(&od->ddev.channels)) {
struct omap_chan *c = list_first_entry(&od->ddev.channels,
struct omap_chan, vc.chan.device_node);
list_del(&c->vc.chan.device_node);
tasklet_kill(&c->vc.task);
kfree(c);
}
}
#define OMAP_DMA_BUSWIDTHS (BIT(DMA_SLAVE_BUSWIDTH_1_BYTE) | \
BIT(DMA_SLAVE_BUSWIDTH_2_BYTES) | \
BIT(DMA_SLAVE_BUSWIDTH_4_BYTES))
static int omap_dma_probe(struct platform_device *pdev)
{
struct omap_dmadev *od;
struct resource *res;
int rc, i, irq;
u32 lch_count;
od = devm_kzalloc(&pdev->dev, sizeof(*od), GFP_KERNEL);
if (!od)
return -ENOMEM;
res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
od->base = devm_ioremap_resource(&pdev->dev, res);
if (IS_ERR(od->base))
return PTR_ERR(od->base);
od->plat = omap_get_plat_info();
if (!od->plat)
return -EPROBE_DEFER;
od->reg_map = od->plat->reg_map;
dma_cap_set(DMA_SLAVE, od->ddev.cap_mask);
dma_cap_set(DMA_CYCLIC, od->ddev.cap_mask);
dma_cap_set(DMA_MEMCPY, od->ddev.cap_mask);
dma_cap_set(DMA_INTERLEAVE, od->ddev.cap_mask);
od->ddev.device_alloc_chan_resources = omap_dma_alloc_chan_resources;
od->ddev.device_free_chan_resources = omap_dma_free_chan_resources;
od->ddev.device_tx_status = omap_dma_tx_status;
od->ddev.device_issue_pending = omap_dma_issue_pending;
od->ddev.device_prep_slave_sg = omap_dma_prep_slave_sg;
od->ddev.device_prep_dma_cyclic = omap_dma_prep_dma_cyclic;
od->ddev.device_prep_dma_memcpy = omap_dma_prep_dma_memcpy;
od->ddev.device_prep_interleaved_dma = omap_dma_prep_dma_interleaved;
od->ddev.device_config = omap_dma_slave_config;
od->ddev.device_pause = omap_dma_pause;
od->ddev.device_resume = omap_dma_resume;
od->ddev.device_terminate_all = omap_dma_terminate_all;
od->ddev.device_synchronize = omap_dma_synchronize;
od->ddev.src_addr_widths = OMAP_DMA_BUSWIDTHS;
od->ddev.dst_addr_widths = OMAP_DMA_BUSWIDTHS;
od->ddev.directions = BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV);
od->ddev.residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
od->ddev.dev = &pdev->dev;
INIT_LIST_HEAD(&od->ddev.channels);
spin_lock_init(&od->lock);
spin_lock_init(&od->irq_lock);
/* Number of DMA requests */
od->dma_requests = OMAP_SDMA_REQUESTS;
if (pdev->dev.of_node && of_property_read_u32(pdev->dev.of_node,
"dma-requests",
&od->dma_requests)) {
dev_info(&pdev->dev,
"Missing dma-requests property, using %u.\n",
OMAP_SDMA_REQUESTS);
}
/* Number of available logical channels */
if (!pdev->dev.of_node) {
lch_count = od->plat->dma_attr->lch_count;
if (unlikely(!lch_count))
lch_count = OMAP_SDMA_CHANNELS;
} else if (of_property_read_u32(pdev->dev.of_node, "dma-channels",
&lch_count)) {
dev_info(&pdev->dev,
"Missing dma-channels property, using %u.\n",
OMAP_SDMA_CHANNELS);
lch_count = OMAP_SDMA_CHANNELS;
}
od->lch_map = devm_kcalloc(&pdev->dev, lch_count, sizeof(*od->lch_map),
GFP_KERNEL);
if (!od->lch_map)
return -ENOMEM;
for (i = 0; i < od->dma_requests; i++) {
rc = omap_dma_chan_init(od);
if (rc) {
omap_dma_free(od);
return rc;
}
}
irq = platform_get_irq(pdev, 1);
if (irq <= 0) {
dev_info(&pdev->dev, "failed to get L1 IRQ: %d\n", irq);
od->legacy = true;
} else {
/* Disable all interrupts */
od->irq_enable_mask = 0;
omap_dma_glbl_write(od, IRQENABLE_L1, 0);
rc = devm_request_irq(&pdev->dev, irq, omap_dma_irq,
IRQF_SHARED, "omap-dma-engine", od);
if (rc)
return rc;
}
if (omap_dma_glbl_read(od, CAPS_0) & CAPS_0_SUPPORT_LL123)
od->ll123_supported = true;
od->ddev.filter.map = od->plat->slave_map;
od->ddev.filter.mapcnt = od->plat->slavecnt;
od->ddev.filter.fn = omap_dma_filter_fn;
if (od->ll123_supported) {
od->desc_pool = dma_pool_create(dev_name(&pdev->dev),
&pdev->dev,
sizeof(struct omap_type2_desc),
4, 0);
if (!od->desc_pool) {
dev_err(&pdev->dev,
"unable to allocate descriptor pool\n");
od->ll123_supported = false;
}
}
rc = dma_async_device_register(&od->ddev);
if (rc) {
pr_warn("OMAP-DMA: failed to register slave DMA engine device: %d\n",
rc);
omap_dma_free(od);
return rc;
}
platform_set_drvdata(pdev, od);
if (pdev->dev.of_node) {
omap_dma_info.dma_cap = od->ddev.cap_mask;
/* Device-tree DMA controller registration */
rc = of_dma_controller_register(pdev->dev.of_node,
of_dma_simple_xlate, &omap_dma_info);
if (rc) {
pr_warn("OMAP-DMA: failed to register DMA controller\n");
dma_async_device_unregister(&od->ddev);
omap_dma_free(od);
}
}
dev_info(&pdev->dev, "OMAP DMA engine driver%s\n",
od->ll123_supported ? " (LinkedList1/2/3 supported)" : "");
return rc;
}
static int omap_dma_remove(struct platform_device *pdev)
{
struct omap_dmadev *od = platform_get_drvdata(pdev);
int irq;
if (pdev->dev.of_node)
of_dma_controller_free(pdev->dev.of_node);
irq = platform_get_irq(pdev, 1);
devm_free_irq(&pdev->dev, irq, od);
dma_async_device_unregister(&od->ddev);
if (!od->legacy) {
/* Disable all interrupts */
omap_dma_glbl_write(od, IRQENABLE_L0, 0);
}
if (od->ll123_supported)
dma_pool_destroy(od->desc_pool);
omap_dma_free(od);
return 0;
}
static const struct of_device_id omap_dma_match[] = {
{ .compatible = "ti,omap2420-sdma", },
{ .compatible = "ti,omap2430-sdma", },
{ .compatible = "ti,omap3430-sdma", },
{ .compatible = "ti,omap3630-sdma", },
{ .compatible = "ti,omap4430-sdma", },
{},
};
MODULE_DEVICE_TABLE(of, omap_dma_match);
static struct platform_driver omap_dma_driver = {
.probe = omap_dma_probe,
.remove = omap_dma_remove,
.driver = {
.name = "omap-dma-engine",
.of_match_table = of_match_ptr(omap_dma_match),
},
};
bool omap_dma_filter_fn(struct dma_chan *chan, void *param)
{
if (chan->device->dev->driver == &omap_dma_driver.driver) {
struct omap_dmadev *od = to_omap_dma_dev(chan->device);
struct omap_chan *c = to_omap_dma_chan(chan);
unsigned req = *(unsigned *)param;
if (req <= od->dma_requests) {
c->dma_sig = req;
return true;
}
}
return false;
}
EXPORT_SYMBOL_GPL(omap_dma_filter_fn);
static int omap_dma_init(void)
{
return platform_driver_register(&omap_dma_driver);
}
subsys_initcall(omap_dma_init);
static void __exit omap_dma_exit(void)
{
platform_driver_unregister(&omap_dma_driver);
}
module_exit(omap_dma_exit);
MODULE_AUTHOR("Russell King");
MODULE_LICENSE("GPL");