Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-11-23 12:42:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

dmaengine fixes for v5.10-rc5

Browse Source 
Core:
 *) channel_register error handling
 
 Driver fixes for:
 *) idxd: wq config registers programming and mapping of portal size
 *) ioatdma: unused fn removal
 *) pl330: fix burst size
 *) ti: pm fix on busy and -Wenum-conversion warns
 *) xilinx: SG capability check, usage of xilinx_aximcdma_tx_segment,
 readl_poll_timeout_atomic variant
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCAAdFiEE+vs47OPLdNbVcHzyfBQHDyUjg0cFAl+32mIACgkQfBQHDyUj
 g0ddOw/9GWhL1vkW+0JTeVSNDVT/GajwXONLbi3brSStlKViG0NrusmIEadNOgAR
 QCj//oww7rkT5jK9KVy577y2Pbns+jZYJYsgUkSQxjaXLztvdMn9ujlP2cpcdo6A
 JZ0+YvMastIAUO2XWVxON7nhhVDcOIUdRlEXR/j0Dllk5NIQInNDyEHi76JQgUfb
 5UV8wZSBO1QbgXogqE8KDnkOsiZEhUfl2Ah0wdYkPtr90GyrBQg8qocR7sfH8idk
 1cJ5bA0UeQX5fpHhIM2dEwtc/115QXJIDOP8u6xMN3StHp9ce+/ghYsBONdPYUN9
 NaRVxs2fyHxp8kx5qz76xZoyIHIZq1Tfyx2oYhTEKUmMWGJYXA4tUSJ071+Wq/eg
 fFwd6u557b1TpUJoU6HmOEQlAYIUXbDMO62pgZj1T/hn6BOuj7SC4v7CzDcucykv
 96Cgj/B6ArFaQEmni4R6XhBCD+vD1Vv/CrayOeBZ0VoZAZLTyH/TZXDKSWQzu+Hl
 KVMHeqR6O/DIUOFFWrz6cYaVynSQecbk5mQwlkWD2G6HzRzBJT2FcBAvNUS/4z6E
 8ie2EQBatjHzsJXWZNUVN2XPUzrJFq+gfY/TCh+2ZNVNvcI5Z6pBfpr0DQGMScbQ
 sf4RiEo7mQ608D04nvD6Sco6QBHdEJYSc86v5qfeTykN5dfilwU=
 =V+mh
 -----END PGP SIGNATURE-----

Merge tag 'dmaengine-fix-5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine

Pull dmaengine fixes from Vinod Koul:
 "A solitary core fix and a few driver fixes:

  Core:

   - channel_register error handling

  Driver fixes:

   - idxd: wq config registers programming and mapping of portal size

   - ioatdma: unused fn removal

   - pl330: fix burst size

   - ti: pm fix on busy and -Wenum-conversion warns

   - xilinx: SG capability check, usage of xilinx_aximcdma_tx_segment,
     readl_poll_timeout_atomic variant"

* tag 'dmaengine-fix-5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine:
  dmaengine: fix error codes in channel_register()
  dmaengine: pl330: _prep_dma_memcpy: Fix wrong burst size
  dmaengine: ioatdma: remove unused function missed during dma_v2 removal
  dmaengine: idxd: fix mapping of portal size
  dmaengine: ti: omap-dma: Block PM if SDMA is busy to fix audio
  dmaengine: xilinx_dma: Fix SG capability check for MCDMA
  dmaengine: xilinx_dma: Fix usage of xilinx_aximcdma_tx_segment
  dmaengine: xilinx_dma: use readl_poll_timeout_atomic variant
  dmaengine: ti: k3-udma: fix -Wenum-conversion warning
  dmaengine: idxd: fix wq config registers offset programming
This commit is contained in:
Linus Torvalds 2020-11-20 10:23:49 -08:00
commit bd4d74e8f8
11 changed files with 111 additions and 63 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
drivers/dma/dmaengine.c
View File

@ -1039,16 +1039,15 @@ static int get_dma_id(struct dma_device *device)
static int __dma_async_device_channel_register(struct dma_device *device,
struct dma_chan *chan)
{
int rc = 0;
int rc;
chan->local = alloc_percpu(typeof(*chan->local));
if (!chan->local)
goto err_out;
return -ENOMEM;
chan->dev = kzalloc(sizeof(*chan->dev), GFP_KERNEL);
if (!chan->dev) {
free_percpu(chan->local);
chan->local = NULL;
goto err_out;
rc = -ENOMEM;
goto err_free_local;
}
/*
@ -1061,7 +1060,8 @@ static int __dma_async_device_channel_register(struct dma_device *device,
if (chan->chan_id < 0) {
pr_err("%s: unable to alloc ida for chan: %d\n",
__func__, chan->chan_id);
goto err_out;
rc = chan->chan_id;
goto err_free_dev;
}
chan->dev->device.class = &dma_devclass;
@ -1082,9 +1082,10 @@ static int __dma_async_device_channel_register(struct dma_device *device,
mutex_lock(&device->chan_mutex);
ida_free(&device->chan_ida, chan->chan_id);
mutex_unlock(&device->chan_mutex);
err_out:
free_percpu(chan->local);
err_free_dev:
kfree(chan->dev);
err_free_local:
free_percpu(chan->local);
return rc;
}

31
drivers/dma/idxd/device.c
View File

@ -271,7 +271,7 @@ int idxd_wq_map_portal(struct idxd_wq *wq)
resource_size_t start;
start = pci_resource_start(pdev, IDXD_WQ_BAR);
start = start + wq->id * IDXD_PORTAL_SIZE;
start += idxd_get_wq_portal_full_offset(wq->id, IDXD_PORTAL_LIMITED);
wq->dportal = devm_ioremap(dev, start, IDXD_PORTAL_SIZE);
if (!wq->dportal)
@ -295,7 +295,7 @@ void idxd_wq_disable_cleanup(struct idxd_wq *wq)
int i, wq_offset;
lockdep_assert_held(&idxd->dev_lock);
memset(&wq->wqcfg, 0, sizeof(wq->wqcfg));
memset(wq->wqcfg, 0, idxd->wqcfg_size);
wq->type = IDXD_WQT_NONE;
wq->size = 0;
wq->group = NULL;
@ -304,8 +304,8 @@ void idxd_wq_disable_cleanup(struct idxd_wq *wq)
clear_bit(WQ_FLAG_DEDICATED, &wq->flags);
memset(wq->name, 0, WQ_NAME_SIZE);
for (i = 0; i < 8; i++) {
wq_offset = idxd->wqcfg_offset + wq->id * 32 + i * sizeof(u32);
for (i = 0; i < WQCFG_STRIDES(idxd); i++) {
wq_offset = WQCFG_OFFSET(idxd, wq->id, i);
iowrite32(0, idxd->reg_base + wq_offset);
dev_dbg(dev, "WQ[%d][%d][%#x]: %#x\n",
wq->id, i, wq_offset,
@ -539,10 +539,10 @@ static int idxd_wq_config_write(struct idxd_wq *wq)
if (!wq->group)
return 0;
memset(&wq->wqcfg, 0, sizeof(union wqcfg));
memset(wq->wqcfg, 0, idxd->wqcfg_size);
/* byte 0-3 */
wq->wqcfg.wq_size = wq->size;
wq->wqcfg->wq_size = wq->size;
if (wq->size == 0) {
dev_warn(dev, "Incorrect work queue size: 0\n");
@ -550,22 +550,21 @@ static int idxd_wq_config_write(struct idxd_wq *wq)
}
/* bytes 4-7 */
wq->wqcfg.wq_thresh = wq->threshold;
wq->wqcfg->wq_thresh = wq->threshold;
/* byte 8-11 */
wq->wqcfg.priv = !!(wq->type == IDXD_WQT_KERNEL);
wq->wqcfg.mode = 1;
wq->wqcfg.priority = wq->priority;
wq->wqcfg->priv = !!(wq->type == IDXD_WQT_KERNEL);
wq->wqcfg->mode = 1;
wq->wqcfg->priority = wq->priority;
/* bytes 12-15 */
wq->wqcfg.max_xfer_shift = ilog2(wq->max_xfer_bytes);
wq->wqcfg.max_batch_shift = ilog2(wq->max_batch_size);
wq->wqcfg->max_xfer_shift = ilog2(wq->max_xfer_bytes);
wq->wqcfg->max_batch_shift = ilog2(wq->max_batch_size);
dev_dbg(dev, "WQ %d CFGs\n", wq->id);
for (i = 0; i < 8; i++) {
wq_offset = idxd->wqcfg_offset + wq->id * 32 + i * sizeof(u32);
iowrite32(wq->wqcfg.bits[i], idxd->reg_base + wq_offset);
for (i = 0; i < WQCFG_STRIDES(idxd); i++) {
wq_offset = WQCFG_OFFSET(idxd, wq->id, i);
iowrite32(wq->wqcfg->bits[i], idxd->reg_base + wq_offset);
dev_dbg(dev, "WQ[%d][%d][%#x]: %#x\n",
wq->id, i, wq_offset,
ioread32(idxd->reg_base + wq_offset));

3
drivers/dma/idxd/idxd.h
View File

@ -103,7 +103,7 @@ struct idxd_wq {
u32 priority;
enum idxd_wq_state state;
unsigned long flags;
union wqcfg wqcfg;
union wqcfg *wqcfg;
u32 vec_ptr; /* interrupt steering */
struct dsa_hw_desc **hw_descs;
int num_descs;
@ -183,6 +183,7 @@ struct idxd_device {
int max_wq_size;
int token_limit;
int nr_tokens; /* non-reserved tokens */
unsigned int wqcfg_size;
union sw_err_reg sw_err;
wait_queue_head_t cmd_waitq;

5
drivers/dma/idxd/init.c
View File

@ -178,6 +178,9 @@ static int idxd_setup_internals(struct idxd_device *idxd)
wq->idxd_cdev.minor = -1;
wq->max_xfer_bytes = idxd->max_xfer_bytes;
wq->max_batch_size = idxd->max_batch_size;
wq->wqcfg = devm_kzalloc(dev, idxd->wqcfg_size, GFP_KERNEL);
if (!wq->wqcfg)
return -ENOMEM;
}
for (i = 0; i < idxd->max_engines; i++) {
@ -251,6 +254,8 @@ static void idxd_read_caps(struct idxd_device *idxd)
dev_dbg(dev, "total workqueue size: %u\n", idxd->max_wq_size);
idxd->max_wqs = idxd->hw.wq_cap.num_wqs;
dev_dbg(dev, "max workqueues: %u\n", idxd->max_wqs);
idxd->wqcfg_size = 1 << (idxd->hw.wq_cap.wqcfg_size + IDXD_WQCFG_MIN);
dev_dbg(dev, "wqcfg size: %u\n", idxd->wqcfg_size);
/* reading operation capabilities */
for (i = 0; i < 4; i++) {

25
drivers/dma/idxd/registers.h
View File

@ -8,7 +8,7 @@
#define IDXD_MMIO_BAR 0
#define IDXD_WQ_BAR 2
#define IDXD_PORTAL_SIZE 0x4000
#define IDXD_PORTAL_SIZE PAGE_SIZE
/* MMIO Device BAR0 Registers */
#define IDXD_VER_OFFSET 0x00
@ -43,7 +43,8 @@ union wq_cap_reg {
struct {
u64 total_wq_size:16;
u64 num_wqs:8;
u64 rsvd:24;
u64 wqcfg_size:4;
u64 rsvd:20;
u64 shared_mode:1;
u64 dedicated_mode:1;
u64 rsvd2:1;
@ -55,6 +56,7 @@ union wq_cap_reg {
u64 bits;
} __packed;
#define IDXD_WQCAP_OFFSET 0x20
#define IDXD_WQCFG_MIN 5
union group_cap_reg {
struct {
@ -333,4 +335,23 @@ union wqcfg {
};
u32 bits[8];
} __packed;
/*
* This macro calculates the offset into the WQCFG register
* idxd - struct idxd *
* n - wq id
* ofs - the index of the 32b dword for the config register
*
* The WQCFG register block is divided into groups per each wq. The n index
* allows us to move to the register group that's for that particular wq.
* Each register is 32bits. The ofs gives us the number of register to access.
*/
#define WQCFG_OFFSET(_idxd_dev, n, ofs) \
({\
typeof(_idxd_dev) __idxd_dev = (_idxd_dev); \
(__idxd_dev)->wqcfg_offset + (n) * (__idxd_dev)->wqcfg_size + sizeof(u32) * (ofs); \
})
#define WQCFG_STRIDES(_idxd_dev) ((_idxd_dev)->wqcfg_size / sizeof(u32))
#endif

2
drivers/dma/idxd/submit.c
View File

@ -74,7 +74,7 @@ int idxd_submit_desc(struct idxd_wq *wq, struct idxd_desc *desc)
if (idxd->state != IDXD_DEV_ENABLED)
return -EIO;
portal = wq->dportal + idxd_get_wq_portal_offset(IDXD_PORTAL_UNLIMITED);
portal = wq->dportal;
/*
* The wmb() flushes writes to coherent DMA data before possibly
* triggering a DMA read. The wmb() is necessary even on UP because

10
drivers/dma/ioat/dca.c
View File

@ -40,16 +40,6 @@
#define DCA2_TAG_MAP_BYTE3 0x82
#define DCA2_TAG_MAP_BYTE4 0x82
/* verify if tag map matches expected values */
static inline int dca2_tag_map_valid(u8 *tag_map)
{
return ((tag_map[0] == DCA2_TAG_MAP_BYTE0) &&
(tag_map[1] == DCA2_TAG_MAP_BYTE1) &&
(tag_map[2] == DCA2_TAG_MAP_BYTE2) &&
(tag_map[3] == DCA2_TAG_MAP_BYTE3) &&
(tag_map[4] == DCA2_TAG_MAP_BYTE4));
}
/*
* "Legacy" DCA systems do not implement the DCA register set in the
* I/OAT device. Software needs direct support for their tag mappings.

2
drivers/dma/pl330.c
View File

@ -2799,7 +2799,7 @@ pl330_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dst,
* If burst size is smaller than bus width then make sure we only
* transfer one at a time to avoid a burst stradling an MFIFO entry.
*/
if (desc->rqcfg.brst_size * 8 < pl330->pcfg.data_bus_width)
if (burst * 8 < pl330->pcfg.data_bus_width)
desc->rqcfg.brst_len = 1;
desc->bytes_requested = len;

2
drivers/dma/ti/k3-udma-private.c
View File

@ -83,7 +83,7 @@ EXPORT_SYMBOL(xudma_rflow_is_gp);
#define XUDMA_GET_PUT_RESOURCE(res) \
struct udma_##res *xudma_##res##_get(struct udma_dev *ud, int id) \
{ \
return __udma_reserve_##res(ud, false, id); \
return __udma_reserve_##res(ud, UDMA_TP_NORMAL, id); \
} \
EXPORT_SYMBOL(xudma_##res##_get); \
\

37
drivers/dma/ti/omap-dma.c
View File

@ -1522,29 +1522,38 @@ static void omap_dma_free(struct omap_dmadev *od)
}
}
/* Currently used by omap2 & 3 to block deeper SoC idle states */
static bool omap_dma_busy(struct omap_dmadev *od)
{
struct omap_chan *c;
int lch = -1;
while (1) {
lch = find_next_bit(od->lch_bitmap, od->lch_count, lch + 1);
if (lch >= od->lch_count)
break;
c = od->lch_map[lch];
if (!c)
continue;
if (omap_dma_chan_read(c, CCR) & CCR_ENABLE)
return true;
}
return false;
}
/* Currently only used for omap2. For omap1, also a check for lcd_dma is needed */
static int omap_dma_busy_notifier(struct notifier_block *nb,
unsigned long cmd, void *v)
{
struct omap_dmadev *od;
struct omap_chan *c;
int lch = -1;
od = container_of(nb, struct omap_dmadev, nb);
switch (cmd) {
case CPU_CLUSTER_PM_ENTER:
while (1) {
lch = find_next_bit(od->lch_bitmap, od->lch_count,
lch + 1);
if (lch >= od->lch_count)
break;
c = od->lch_map[lch];
if (!c)
continue;
if (omap_dma_chan_read(c, CCR) & CCR_ENABLE)
return NOTIFY_BAD;
}
if (omap_dma_busy(od))
return NOTIFY_BAD;
break;
case CPU_CLUSTER_PM_ENTER_FAILED:
case CPU_CLUSTER_PM_EXIT:
@ -1595,6 +1604,8 @@ static int omap_dma_context_notifier(struct notifier_block *nb,
switch (cmd) {
case CPU_CLUSTER_PM_ENTER:
if (omap_dma_busy(od))
return NOTIFY_BAD;
omap_dma_context_save(od);
break;
case CPU_CLUSTER_PM_ENTER_FAILED:

40
drivers/dma/xilinx/xilinx_dma.c
View File

@ -517,8 +517,8 @@ struct xilinx_dma_device {
#define to_dma_tx_descriptor(tx) \
container_of(tx, struct xilinx_dma_tx_descriptor, async_tx)
#define xilinx_dma_poll_timeout(chan, reg, val, cond, delay_us, timeout_us) \
readl_poll_timeout(chan->xdev->regs + chan->ctrl_offset + reg, val, \
cond, delay_us, timeout_us)
readl_poll_timeout_atomic(chan->xdev->regs + chan->ctrl_offset + reg, \
val, cond, delay_us, timeout_us)
/* IO accessors */
static inline u32 dma_read(struct xilinx_dma_chan *chan, u32 reg)
@ -948,8 +948,10 @@ static u32 xilinx_dma_get_residue(struct xilinx_dma_chan *chan,
{
struct xilinx_cdma_tx_segment *cdma_seg;
struct xilinx_axidma_tx_segment *axidma_seg;
struct xilinx_aximcdma_tx_segment *aximcdma_seg;
struct xilinx_cdma_desc_hw *cdma_hw;
struct xilinx_axidma_desc_hw *axidma_hw;
struct xilinx_aximcdma_desc_hw *aximcdma_hw;
struct list_head *entry;
u32 residue = 0;
@ -961,13 +963,23 @@ static u32 xilinx_dma_get_residue(struct xilinx_dma_chan *chan,
cdma_hw = &cdma_seg->hw;
residue += (cdma_hw->control - cdma_hw->status) &
chan->xdev->max_buffer_len;
} else {
} else if (chan->xdev->dma_config->dmatype ==
XDMA_TYPE_AXIDMA) {
axidma_seg = list_entry(entry,
struct xilinx_axidma_tx_segment,
node);
axidma_hw = &axidma_seg->hw;
residue += (axidma_hw->control - axidma_hw->status) &
chan->xdev->max_buffer_len;
} else {
aximcdma_seg =
list_entry(entry,
struct xilinx_aximcdma_tx_segment,
node);
aximcdma_hw = &aximcdma_seg->hw;
residue +=
(aximcdma_hw->control - aximcdma_hw->status) &
chan->xdev->max_buffer_len;
}
}
@ -1135,7 +1147,7 @@ static int xilinx_dma_alloc_chan_resources(struct dma_chan *dchan)
upper_32_bits(chan->seg_p + sizeof(*chan->seg_mv) *
((i + 1) % XILINX_DMA_NUM_DESCS));
chan->seg_mv[i].phys = chan->seg_p +
sizeof(*chan->seg_v) * i;
sizeof(*chan->seg_mv) * i;
list_add_tail(&chan->seg_mv[i].node,
&chan->free_seg_list);
}
@ -1560,7 +1572,7 @@ static void xilinx_dma_start_transfer(struct xilinx_dma_chan *chan)
static void xilinx_mcdma_start_transfer(struct xilinx_dma_chan *chan)
{
struct xilinx_dma_tx_descriptor *head_desc, *tail_desc;
struct xilinx_axidma_tx_segment *tail_segment;
struct xilinx_aximcdma_tx_segment *tail_segment;
u32 reg;
/*
@ -1582,7 +1594,7 @@ static void xilinx_mcdma_start_transfer(struct xilinx_dma_chan *chan)
tail_desc = list_last_entry(&chan->pending_list,
struct xilinx_dma_tx_descriptor, node);
tail_segment = list_last_entry(&tail_desc->segments,
struct xilinx_axidma_tx_segment, node);
struct xilinx_aximcdma_tx_segment, node);
reg = dma_ctrl_read(chan, XILINX_MCDMA_CHAN_CR_OFFSET(chan->tdest));
@ -1864,6 +1876,7 @@ static void append_desc_queue(struct xilinx_dma_chan *chan,
struct xilinx_vdma_tx_segment *tail_segment;
struct xilinx_dma_tx_descriptor *tail_desc;
struct xilinx_axidma_tx_segment *axidma_tail_segment;
struct xilinx_aximcdma_tx_segment *aximcdma_tail_segment;
struct xilinx_cdma_tx_segment *cdma_tail_segment;
if (list_empty(&chan->pending_list))
@ -1885,11 +1898,17 @@ static void append_desc_queue(struct xilinx_dma_chan *chan,
struct xilinx_cdma_tx_segment,
node);
cdma_tail_segment->hw.next_desc = (u32)desc->async_tx.phys;
} else {
} else if (chan->xdev->dma_config->dmatype == XDMA_TYPE_AXIDMA) {
axidma_tail_segment = list_last_entry(&tail_desc->segments,
struct xilinx_axidma_tx_segment,
node);
axidma_tail_segment->hw.next_desc = (u32)desc->async_tx.phys;
} else {
aximcdma_tail_segment =
list_last_entry(&tail_desc->segments,
struct xilinx_aximcdma_tx_segment,
node);
aximcdma_tail_segment->hw.next_desc = (u32)desc->async_tx.phys;
}
/*
@ -2836,10 +2855,11 @@ static int xilinx_dma_chan_probe(struct xilinx_dma_device *xdev,
chan->stop_transfer = xilinx_dma_stop_transfer;
}
/* check if SG is enabled (only for AXIDMA and CDMA) */
/* check if SG is enabled (only for AXIDMA, AXIMCDMA, and CDMA) */
if (xdev->dma_config->dmatype != XDMA_TYPE_VDMA) {
if (dma_ctrl_read(chan, XILINX_DMA_REG_DMASR) &
XILINX_DMA_DMASR_SG_MASK)
if (xdev->dma_config->dmatype == XDMA_TYPE_AXIMCDMA ||
dma_ctrl_read(chan, XILINX_DMA_REG_DMASR) &
XILINX_DMA_DMASR_SG_MASK)
chan->has_sg = true;
dev_dbg(chan->dev, "ch %d: SG %s\n", chan->id,
chan->has_sg ? "enabled" : "disabled");