dmaengine: ste_dma40: Fix V4B event group mapping

DMA40 physical resources are connected in pairs to event groups. The
mapping repeats after each complete set of groups.

D40_GROUP_SIZE and D40_PHYS_TO_GROUP() encode the four-group V4A layout.
On V4B, which has five groups, resources 8 and 9 are consequently offered
both to event groups 0 and 4. The GCC clock bookkeeping also treats these
resources as group 0 instead of group 4.

Derive the physical-resource stride and reverse mapping from the
variant-specific number of event groups. This preserves the V4A mapping
and gives V4B a ten-resource stride: resources 8 and 9 belong only to
group 4, while resources 10 and 11 repeat group 0.

Fixes: 3cb645dc85a0 ("dmaengine: ste_dma40: support more than 128 event lines")
Assisted-by: LLM
Signed-off-by: Linus Walleij <linusw@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260927-dma40-fixes-v7-21-89f595e8851d@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

2 files changed, 33 insertions, 16 deletions

diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
index 0e1c0da058a0..208e877d93d7 100644
--- a/drivers/dma/ste_dma40.c
+++ b/drivers/dma/ste_dma40.c
@@ -85,6 +85,7 @@ struct stedma40_platform_data {
 /* Number of event groups per hardware register layout */
 #define D40_EVENT_GROUPS_V4A 4
 #define D40_EVENT_GROUPS_V4B 5
+#define D40_PHYS_PER_EVENT_GROUP 2
 
 /* Max number of logical channels per physical channel */
 #define D40_MAX_LOG_CHAN_PER_PHY 32
@@ -2022,6 +2023,12 @@ static bool d40_alloc_mask_free(struct d40_phy_res *phy, bool is_src,
 	return is_free;
 }
 
+static int d40_phy_to_group(struct d40_base *base, int phy)
+{
+	return (phy / D40_PHYS_PER_EVENT_GROUP) %
+	       base->gen_dmac.num_event_groups;
+}
+
 static int d40_allocate_channel(struct d40_chan *d40c, bool *first_phy_user)
 {
 	int dev_type = d40c->dma_cfg.dev_type;
@@ -2032,11 +2039,14 @@ static int d40_allocate_channel(struct d40_chan *d40c, bool *first_phy_user)
 	int j;
 	int log_num;
 	int num_phy_chans;
+	int phy_group_stride;
 	bool is_src;
 	bool is_log = d40c->dma_cfg.mode == STEDMA40_MODE_LOGICAL;
 
 	phys = d40c->base->phy_res;
 	num_phy_chans = d40c->base->num_phy_chans;
+	phy_group_stride = D40_PHYS_PER_EVENT_GROUP *
+			   d40c->base->gen_dmac.num_event_groups;
 
 	if (d40c->dma_cfg.dir == DMA_DEV_TO_MEM) {
 		log_num = 2 * dev_type;
@@ -2070,13 +2080,18 @@ static int d40_allocate_channel(struct d40_chan *d40c, bool *first_phy_user)
 				}
 			}
 		} else
-			for (j = 0; j < d40c->base->num_phy_chans;
-			     j += D40_GROUP_SIZE) {
-				int phy_num = j  + event_group * 2;
-				if (phy_num + 1 >= num_phy_chans)
+			for (j = 0; j < num_phy_chans;
+			     j += phy_group_stride) {
+				int phy_num = j + event_group *
+					      D40_PHYS_PER_EVENT_GROUP;
+
+				if (phy_num + D40_PHYS_PER_EVENT_GROUP >
+				    num_phy_chans)
 					break;
 
-				for (i = phy_num; i < phy_num + 2; i++) {
+				for (i = phy_num;
+				     i < phy_num + D40_PHYS_PER_EVENT_GROUP;
+				     i++) {
 					if (d40_alloc_mask_set(&phys[i],
 							       is_src,
 							       0,
@@ -2095,9 +2110,10 @@ found_phy:
 		return -EINVAL;
 
 	/* Find logical channel */
-	for (j = 0; j < d40c->base->num_phy_chans; j += D40_GROUP_SIZE) {
-		int phy_num = j + event_group * 2;
-		if (phy_num + 1 >= num_phy_chans)
+	for (j = 0; j < num_phy_chans; j += phy_group_stride) {
+		int phy_num = j + event_group * D40_PHYS_PER_EVENT_GROUP;
+
+		if (phy_num + D40_PHYS_PER_EVENT_GROUP > num_phy_chans)
 			break;
 
 		if (d40c->dma_cfg.use_fixed_channel) {
@@ -2124,14 +2140,17 @@ found_phy:
 		 * channels.
 		 */
 		if (is_src) {
-			for (i = phy_num; i < phy_num + 2; i++) {
+			for (i = phy_num;
+			     i < phy_num + D40_PHYS_PER_EVENT_GROUP;
+			     i++) {
 				if (d40_alloc_mask_set(&phys[i], is_src,
 						       event_line, is_log,
 						       first_phy_user))
 					goto found_log;
 			}
 		} else {
-			for (i = phy_num + 1; i >= phy_num; i--) {
+			for (i = phy_num + D40_PHYS_PER_EVENT_GROUP - 1;
+			     i >= phy_num; i--) {
 				if (d40_alloc_mask_set(&phys[i], is_src,
 						       event_line, is_log,
 						       first_phy_user))
@@ -3302,9 +3321,9 @@ static int __init d40_phy_res_init(struct d40_base *base)
 			base->phy_res[i].allocated_src = D40_ALLOC_PHY;
 			base->phy_res[i].allocated_dst = D40_ALLOC_PHY;
 			base->phy_res[i].reserved = true;
-			gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(i),
+			gcc |= D40_DREG_GCC_EVTGRP_ENA(d40_phy_to_group(base, i),
 						       D40_DREG_GCC_SRC);
-			gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(i),
+			gcc |= D40_DREG_GCC_EVTGRP_ENA(d40_phy_to_group(base, i),
 						       D40_DREG_GCC_DST);
 
 
@@ -3324,9 +3343,9 @@ static int __init d40_phy_res_init(struct d40_base *base)
 		base->phy_res[chan].allocated_src = D40_ALLOC_PHY;
 		base->phy_res[chan].allocated_dst = D40_ALLOC_PHY;
 		base->phy_res[chan].reserved = true;
-		gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(chan),
+		gcc |= D40_DREG_GCC_EVTGRP_ENA(d40_phy_to_group(base, chan),
 					       D40_DREG_GCC_SRC);
-		gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(chan),
+		gcc |= D40_DREG_GCC_EVTGRP_ENA(d40_phy_to_group(base, chan),
 					       D40_DREG_GCC_DST);
 		num_phy_chans_avail--;
 	}
diff --git a/drivers/dma/ste_dma40_ll.h b/drivers/dma/ste_dma40_ll.h
index 2e30e9a94a1e..55090cb31057 100644
--- a/drivers/dma/ste_dma40_ll.h
+++ b/drivers/dma/ste_dma40_ll.h
@@ -16,8 +16,6 @@
 
 #define D40_TYPE_TO_GROUP(type) (type / 16)
 #define D40_TYPE_TO_EVENT(type) (type % 16)
-#define D40_GROUP_SIZE 8
-#define D40_PHYS_TO_GROUP(phys) ((phys & (D40_GROUP_SIZE - 1)) / 2)
 
 /* Most bits of the CFG register are the same in log as in phy mode */
 #define D40_SREG_CFG_MST_POS		15