Skip to contentSkip to conversation

soc: TI knav_qmss: fix dma_addr_t printing

PATCH In review Arnd Bergmann 2 participants 4 messages Dec 8, 2015 TI Keystone multicore navigator drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 soc: TI knav_qmss: fix dma_addr_t printing

The knav_qmss driver is currently broken when CONFIG_LPAE is set, which is a bit surprising because I'd expect that any serious users of this platforms would have more than 2GB of RAM and require LPAE.

The compiler clearly warns about an incorrect use of dma_addr_t in the debug kernel messages:

ti/knav_qmss_queue.c: In function 'knav_queue_setup_region':
ti/knav_qmss_queue.c:1025:117: warning: format '%x' expects argument of type 'unsigned int', but argument 9 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=] ti/knav_qmss_queue.c:1025:117: warning: format '%x' expects argument of type 'unsigned int', but argument 10 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=] ti/knav_qmss_queue.c: In function 'knav_queue_setup_link_ram':
ti/knav_qmss_queue.c:1175:118: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=]

This patch changes all the debugging output to use the correct %pad format string that works with both 32-bit and 64-bit dma_addr_t. As the variable naming is somewhat confusing here, I also change all *_phys names to *_dma when they refer to bus addresses that are used for DMA rather than a physical memory address as seen from the CPU. This is particularly important on keystone, because the two things are not the same there.

---
 drivers/soc/ti/knav_qmss.h       |  4 ++--
 drivers/soc/ti/knav_qmss_acc.c   | 14 +++++++-------
 drivers/soc/ti/knav_qmss_queue.c | 22 +++++++++++-----------
 3 files changed, 20 insertions(+), 20 deletions(-)
drivers/soc/ti/knav_qmss.h +2 -2
index 6ff936cacb70..905b974d1bdc 100644
@@ -93,13 +93,13 @@ struct knav_reg_pdsp_regs {
 struct knav_reg_acc_command {
 	u32		command;
 	u32		queue_mask;
-	u32		list_phys;
+	u32		list_dma;
 	u32		queue_num;
 	u32		timer_config;
 };
 
 struct knav_link_ram_block {
-	dma_addr_t	 phys;
+	dma_addr_t	 dma;
 	void		*virt;
 	size_t		 size;
 };
drivers/soc/ti/knav_qmss_acc.c +7 -7
index d2d48f2802bc..0612ebae0a09 100644
@@ -122,8 +122,8 @@ static irqreturn_t knav_acc_int_handler(int irq, void *_instdata)
 	channel = acc->channel;
 	list_dma = acc->list_dma[acc->list_index];
 	list_cpu = acc->list_cpu[acc->list_index];
-	dev_dbg(kdev->dev, "acc-irq: channel %d, list %d, virt %p, phys %x\n",
-		channel, acc->list_index, list_cpu, list_dma);
+	dev_dbg(kdev->dev, "acc-irq: channel %d, list %d, virt %p, dma %pad\n",
+		channel, acc->list_index, list_cpu, &list_dma);
 	if (atomic_read(&acc->retrigger_count)) {
 		atomic_dec(&acc->retrigger_count);
 		__knav_acc_notify(range, acc);
@@ -297,12 +297,12 @@ knav_acc_write(struct knav_device *kdev, struct knav_pdsp_info *pdsp,
 	u32 result;
 
 	dev_dbg(kdev->dev, "acc command %08x %08x %08x %08x %08x\n",
-		cmd->command, cmd->queue_mask, cmd->list_phys,
+		cmd->command, cmd->queue_mask, cmd->list_dma,
 		cmd->queue_num, cmd->timer_config);
 
 	writel_relaxed(cmd->timer_config, &pdsp->acc_command->timer_config);
 	writel_relaxed(cmd->queue_num, &pdsp->acc_command->queue_num);
-	writel_relaxed(cmd->list_phys, &pdsp->acc_command->list_phys);
+	writel_relaxed(cmd->list_dma, &pdsp->acc_command->list_dma);
 	writel_relaxed(cmd->queue_mask, &pdsp->acc_command->queue_mask);
 	writel_relaxed(cmd->command, &pdsp->acc_command->command);
 
@@ -337,7 +337,7 @@ static void knav_acc_setup_cmd(struct knav_device *kdev,
 	memset(cmd, 0, sizeof(*cmd));
 	cmd->command    = acc->channel;
 	cmd->queue_mask = queue_mask;
-	cmd->list_phys  = acc->list_dma[0];
+	cmd->list_dma   = (u32)acc->list_dma[0];
 	cmd->queue_num  = info->list_entries << 16;
 	cmd->queue_num |= queue_base;
 
@@ -591,8 +591,8 @@ int knav_init_acc_range(struct knav_device *kdev,
 		acc->list_cpu[1] = list_mem + list_size;
 		acc->list_dma[0] = list_dma;
 		acc->list_dma[1] = list_dma + list_size;
-		dev_dbg(kdev->dev, "%s: channel %d, phys %08x, virt %8p\n",
-			acc->name, acc->channel, list_dma, list_mem);
+		dev_dbg(kdev->dev, "%s: channel %d, dma %pad, virt %8p\n",
+			acc->name, acc->channel, &list_dma, list_mem);
 	}
 
 	range->ops = &knav_acc_range_ops;
drivers/soc/ti/knav_qmss_queue.c +11 -11
index 8c03a80b482d..b73e3534f67b 100644
@@ -1023,9 +1023,9 @@ static void knav_queue_setup_region(struct knav_device *kdev,
 	list_add(&pool->region_inst, &region->pools);
 
 	dev_dbg(kdev->dev,
-		"region %s (%d): size:%d, link:%d@%d, phys:%08x-%08x, virt:%p-%p\n",
+		"region %s (%d): size:%d, link:%d@%d, dma:%pad-%pad, virt:%p-%p\n",
 		region->name, id, region->desc_size, region->num_desc,
-		region->link_index, region->dma_start, region->dma_end,
+		region->link_index, &region->dma_start, &region->dma_end,
 		region->virt_start, region->virt_end);
 
 	hw_desc_size = (region->desc_size / 16) - 1;
@@ -1033,7 +1033,7 @@ static void knav_queue_setup_region(struct knav_device *kdev,
 
 	for_each_qmgr(kdev, qmgr) {
 		regs = qmgr->reg_region + id;
-		writel_relaxed(region->dma_start, &regs->base);
+		writel_relaxed((u32)region->dma_start, &regs->base);
 		writel_relaxed(region->link_index, &regs->start_index);
 		writel_relaxed(hw_desc_size << 16 | hw_num_desc,
 			       &regs->size_count);
@@ -1145,14 +1145,14 @@ static int knav_get_link_ram(struct knav_device *kdev,
 			 * queue_base specified => using internal or onchip
 			 * link ram WARNING - we do not "reserve" this block
 			 */
-			block->phys = (dma_addr_t)temp[0];
+			block->dma = (dma_addr_t)temp[0];
 			block->virt = NULL;
 			block->size = temp[1];
 		} else {
 			block->size = temp[1];
 			/* queue_base not specific => allocate requested size */
 			block->virt = dmam_alloc_coherent(kdev->dev,
-						  8 * block->size, &block->phys,
+						  8 * block->size, &block->dma,
 						  GFP_KERNEL);
 			if (!block->virt) {
 				dev_err(kdev->dev, "failed to alloc linkram\n");
@@ -1172,18 +1172,18 @@ static int knav_queue_setup_link_ram(struct knav_device *kdev)
 
 	for_each_qmgr(kdev, qmgr) {
 		block = &kdev->link_rams[0];
-		dev_dbg(kdev->dev, "linkram0: phys:%x, virt:%p, size:%x\n",
-			block->phys, block->virt, block->size);
-		writel_relaxed(block->phys, &qmgr->reg_config->link_ram_base0);
+		dev_dbg(kdev->dev, "linkram0: dma:%pad, virt:%p, size:%x\n",
+			&block->dma, block->virt, block->size);
+		writel_relaxed((u32)block->dma, &qmgr->reg_config->link_ram_base0);
 		writel_relaxed(block->size, &qmgr->reg_config->link_ram_size0);
 
 		block++;
 		if (!block->size)
 			continue;
 
-		dev_dbg(kdev->dev, "linkram1: phys:%x, virt:%p, size:%x\n",
-			block->phys, block->virt, block->size);
-		writel_relaxed(block->phys, &qmgr->reg_config->link_ram_base1);
+		dev_dbg(kdev->dev, "linkram1: dma:%pad, virt:%p, size:%x\n",
+			&block->dma, block->virt, block->size);
+		writel_relaxed(block->dma, &qmgr->reg_config->link_ram_base1);
 	}
 
 	return 0;
Signature · 1 line
2.1.0.rc2
#1 lore raw
Quoted 5 lines · Arnd Bergmann

The knav_qmss driver is currently broken when CONFIG_LPAE is set, which is a bit surprising because I'd expect that any serious users of this platforms would have more than 2GB of RAM and require LPAE.

Well it has been working with LPAE enable always. The production systems do but IIRC, some of the Eval board were still build with just 2 GB memory.

Quoted 9 lines

The compiler clearly warns about an incorrect use of dma_addr_t in the debug kernel messages:

ti/knav_qmss_queue.c: In function 'knav_queue_setup_region':
ti/knav_qmss_queue.c:1025:117: warning: format '%x' expects argument of type 'unsigned int', but argument 9 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=] ti/knav_qmss_queue.c:1025:117: warning: format '%x' expects argument of type 'unsigned int', but argument 10 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=] ti/knav_qmss_queue.c: In function 'knav_queue_setup_link_ram':
ti/knav_qmss_queue.c:1175:118: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=]

%x should have been %llx or %pad as your patch does now.

Quoted 8 lines

This patch changes all the debugging output to use the correct %pad format string that works with both 32-bit and 64-bit dma_addr_t. As the variable naming is somewhat confusing here, I also change all *_phys names to *_dma when they refer to bus addresses that are used for DMA rather than a physical memory address as seen from the CPU. This is particularly important on keystone, because the two things are not the same there.

Right. Rename sounds good.

Quoted 2 lines

---

Thanks for the patch Arnd !!

Wingman/Murali,
Can you guys please test this patch and update the thread ? I will queue this up for next merge window.

#2 lore raw
Quoted 9 lines · On Tuesday 08 December 2015 08:53:53 santosh shilimkar wrote:
Quoted 5 lines · Arnd Bergmann

The knav_qmss driver is currently broken when CONFIG_LPAE is set, which is a bit surprising because I'd expect that any serious users of this platforms would have more than 2GB of RAM and require LPAE.

Well it has been working with LPAE enable always. The production systems do but IIRC, some of the Eval board were still build with just 2 GB memory.

Maybe that just means that the debug code is never used and we can remove all the dev_dbg instead? That might actually be nicer as it would reduce the amount of code in your driver (slightly).

Arnd

#3 lore raw
Quoted 7 lines · Arnd Bergmann
Quoted 6 lines · On Tuesday 08 December 2015 08:53:53 santosh shilimkar wrote:
Quoted 5 lines · Arnd Bergmann

The knav_qmss driver is currently broken when CONFIG_LPAE is set, which is a bit surprising because I'd expect that any serious users of this platforms would have more than 2GB of RAM and require LPAE.

Quoted 8 lines
Quoted 3 lines

Well it has been working with LPAE enable always. The production systems do but IIRC, some of the Eval board were still build with just 2 GB memory.

Maybe that just means that the debug code is never used and we can remove all the dev_dbg instead? That might actually be nicer as it would reduce the amount of code in your driver (slightly).

It was added in intention so that it is useful. I agree if it is not used or useful, we should remove it.

Regards,
Santosh