fsl: esdhc: consolidate fsl_esdhc_cfg structure

PATCH Superseded Peng Fan 3 participants 10 messages Mar 15, 2016 – Apr 6, 2016

lore

Share & tools

lore

Peng Fan

author

2016-03-15 09:57 UTC #0 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

We can use phys_addr_to for esdhc_base to discard
the #ifdef.

Signed-off-by Peng Fan <[email protected]>

Cc York Sun <[email protected]>

Cc Yangbo Lu <[email protected]>

Cc Eric Nelson <[email protected]>

Cc Fabio Estevam <[email protected]>

Cc Pantelis Antoniou <[email protected]>

Cc Tom Rini <[email protected]>

---

V2:
Split this patch from the V1 patch set.

 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

include/fsl_esdhc.h +1 -5

index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 
 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
-	u64	esdhc_base;
-#else
-	u32	esdhc_base;
-#endif
+	phys_addr_t esdhc_base;
 	u32	sdhc_clk;
 	u8	max_bus_width;
 	struct mmc_config cfg;

Signature · 1 line

2.6.2

Tom Rini

2016-03-15 13:58 UTC #1 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

Quoted 10 lines · Peng Fan

We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>

Reviewed-by Tom Rini <[email protected]>

Signature · 8 lines

Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20160315/d38ec529/attachment.sig>

york sun

2016-03-29 00:33 UTC #2 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

In reply to Peng Fan

Quoted 26 lines · Peng Fan

We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>
---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
include/fsl_esdhc.h +0 -0
index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 

include/fsl_esdhc.h line 174

Quoted 10 lines

include/fsl_esdhc.h +1 -5

 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
-	u64	esdhc_base;
-#else
-	u32	esdhc_base;
-#endif
+	phys_addr_t esdhc_base;
 	u32	sdhc_clk;
 	u8	max_bus_width;
 	struct mmc_config cfg;

Peng,

I thought this change is trivial and should be OK. But it turns out this change brings comping warning to many PPC boards

drivers/mmc/fsl_esdhc.c:184:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]

For assignment like this

struct fsl_esdhc *regs = (struct fsl_esdhc *)cfg->esdhc_base;

The phys_addr is not u32 for many targets.

York

Peng Fan

author

2016-03-29 01:22 UTC #3 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

Hi York,

Quoted 40 lines · york sun

Quoted 26 lines · Peng Fan
We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>
---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
include/fsl_esdhc.h +0 -0
index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 
include/fsl_esdhc.h line 174
Quoted 10 lines
include/fsl_esdhc.h +1 -5
 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
-	u64	esdhc_base;
-#else
-	u32	esdhc_base;
-#endif
+	phys_addr_t esdhc_base;
 	u32	sdhc_clk;
 	u8	max_bus_width;
 	struct mmc_config cfg;
Peng,

Quoted 5 lines

I thought this change is trivial and should be OK. But it turns out this change brings comping warning to many PPC boards
drivers/mmc/fsl_esdhc.c:184:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]

You can apply https://patchwork.ozlabs.org/patch/601919/ and retest. I think the reason is that to PPC 64bit, "typedef unsigned long long phys_addr_t" I think "esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" can fix it.

Do I need to send a single V3 patch to fix the warning, or you apply the driver model V3 patch, if the driver model V3 can fix it?

Regards,
Peng.

Quoted 9 lines

For assignment like this
struct fsl_esdhc *regs = (struct fsl_esdhc *)cfg->esdhc_base;
The phys_addr is not u32 for many targets.
York

york sun

2016-03-29 01:35 UTC #4 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

Quoted 49 lines · Peng Fan

Hi York,
Quoted 40 lines · york sun
Quoted 26 lines · Peng Fan
We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>
---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
include/fsl_esdhc.h +0 -0
index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 
include/fsl_esdhc.h line 174
Quoted 10 lines
include/fsl_esdhc.h +1 -5
 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
-	u64	esdhc_base;
-#else
-	u32	esdhc_base;
-#endif
+	phys_addr_t esdhc_base;
 	u32	sdhc_clk;
 	u8	max_bus_width;
 	struct mmc_config cfg;
Peng,
Quoted 5 lines
I thought this change is trivial and should be OK. But it turns out this change brings comping warning to many PPC boards
drivers/mmc/fsl_esdhc.c:184:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]

Quoted 8 lines

You can apply https://patchwork.ozlabs.org/patch/601919/ and retest. I think the reason is that to PPC 64bit, "typedef unsigned long long phys_addr_t" I think "esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" can fix it.
Do I need to send a single V3 patch to fix the warning, or you apply the driver model V3 patch, if the driver model V3 can fix it?

I am concerned about git bisect. It is not good to have a patch with compiling warning. I prefer you fix it.

York

Peng Fan

author

2016-03-29 04:37 UTC #5 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

Hi York,

Quoted 55 lines · york sun

Quoted 54 lines · Peng Fan
Hi York,
Quoted 40 lines · york sun
Quoted 26 lines · Peng Fan
We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>
---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
include/fsl_esdhc.h +0 -0
index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 
include/fsl_esdhc.h line 174
Quoted 10 lines
include/fsl_esdhc.h +1 -5
 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
-	u64	esdhc_base;
-#else
-	u32	esdhc_base;
-#endif
+	phys_addr_t esdhc_base;
 	u32	sdhc_clk;
 	u8	max_bus_width;
 	struct mmc_config cfg;
Peng,
Quoted 5 lines
I thought this change is trivial and should be OK. But it turns out this change brings comping warning to many PPC boards
drivers/mmc/fsl_esdhc.c:184:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
You can apply https://patchwork.ozlabs.org/patch/601919/ and retest. I think the reason is that to PPC 64bit, "typedef unsigned long long phys_addr_t" I think "esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" can fix it.

Quoted 6 lines

Quoted 3 lines
Do I need to send a single V3 patch to fix the warning, or you apply the driver model V3 patch, if the driver model V3 can fix it?
I am concerned about git bisect. It is not good to have a patch with compiling warning. I prefer you fix it.

You can first apply https://patchwork.ozlabs.org/patch/601919/ V3 version for driver model patch. In this patch I have such code:
"esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" If you think this line code is ok, then you can apply this current patch. If not, I think I may also fix this.

If this is ok for you, no need for me to send V3 :)

Thanks,
Peng.

Quoted 4 lines

York

York Sun

2016-03-29 04:49 UTC #6 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

Quoted 65 lines · Peng Fan

Hi York,
Quoted 55 lines · york sun
Quoted 54 lines · Peng Fan
Hi York,
Quoted 40 lines · york sun
Quoted 26 lines · Peng Fan
We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>
---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
include/fsl_esdhc.h +0 -0
index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 
include/fsl_esdhc.h line 174
Quoted 10 lines
include/fsl_esdhc.h +1 -5
 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
-	u64	esdhc_base;
-#else
-	u32	esdhc_base;
-#endif
+	phys_addr_t esdhc_base;
 	u32	sdhc_clk;
 	u8	max_bus_width;
 	struct mmc_config cfg;
Peng,
Quoted 5 lines
I thought this change is trivial and should be OK. But it turns out this change brings comping warning to many PPC boards
drivers/mmc/fsl_esdhc.c:184:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
You can apply https://patchwork.ozlabs.org/patch/601919/ and retest. I think the reason is that to PPC 64bit, "typedef unsigned long long phys_addr_t" I think "esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" can fix it.
Quoted 6 lines
Quoted 3 lines
Do I need to send a single V3 patch to fix the warning, or you apply the driver model V3 patch, if the driver model V3 can fix it?
I am concerned about git bisect. It is not good to have a patch with compiling warning. I prefer you fix it.

Quoted 8 lines

You can first apply https://patchwork.ozlabs.org/patch/601919/ V3 version for driver model patch. In this patch I have such code:
"esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" If you think this line code is ok, then you can apply this current patch. If not, I think I may also fix this.
If this is ok for you, no need for me to send V3 :)

Will try tomorrow.

York

Peng Fan

author

2016-04-05 03:01 UTC #7 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

Hi York,

Quoted 72 lines · York Sun

Quoted 71 lines · Peng Fan
Hi York,
Quoted 55 lines · york sun
Quoted 54 lines · Peng Fan
Hi York,
Quoted 40 lines · york sun
Quoted 26 lines · Peng Fan
We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>
---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
include/fsl_esdhc.h +0 -0
index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 
include/fsl_esdhc.h line 174
Quoted 10 lines
include/fsl_esdhc.h +1 -5
 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
-	u64	esdhc_base;
-#else
-	u32	esdhc_base;
-#endif
+	phys_addr_t esdhc_base;
 	u32	sdhc_clk;
 	u8	max_bus_width;
 	struct mmc_config cfg;
Peng,
Quoted 5 lines
I thought this change is trivial and should be OK. But it turns out this change brings comping warning to many PPC boards
drivers/mmc/fsl_esdhc.c:184:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
You can apply https://patchwork.ozlabs.org/patch/601919/ and retest. I think the reason is that to PPC 64bit, "typedef unsigned long long phys_addr_t" I think "esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" can fix it.
Quoted 6 lines
Quoted 3 lines
Do I need to send a single V3 patch to fix the warning, or you apply the driver model V3 patch, if the driver model V3 can fix it?
I am concerned about git bisect. It is not good to have a patch with compiling warning. I prefer you fix it.
You can first apply https://patchwork.ozlabs.org/patch/601919/ V3 version for driver model patch. In this patch I have such code:
"esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" If you think this line code is ok, then you can apply this current patch. If not, I think I may also fix this.

Quoted 4 lines

Quoted 2 lines
If this is ok for you, no need for me to send V3 :)
Will try tomorrow.

Is there any update on this? If this patch with the driver model patch are ok in your side, will you pick the two patches?

Thanks,
Peng.

Quoted 4 lines

York

york sun

2016-04-05 03:20 UTC #8 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

Yes they are good. They are in my test queue.

York

-------- Original Message --------
From: Peng Fan <[email protected]>
Sent: Monday, April 4, 2016 08:05 PM
To: york sun <[email protected]>
Subject: Re: [PATCH V2] fsl: esdhc: consolidate fsl_esdhc_cfg structure
CC: sbabic at denx.de,trini at konsulko.com,u-boot at lists.denx.de,Yangbo Lu <[email protected]>,Eric Nelson <[email protected]>,Fabio Estevam <[email protected]>,Pantelis Antoniou <[email protected]>

Hi York,

Quoted 72 lines · York Sun

Quoted 71 lines · Peng Fan
Hi York,
Quoted 55 lines · york sun
Quoted 54 lines · Peng Fan
Hi York,
Quoted 40 lines · york sun
Quoted 26 lines · Peng Fan
We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>
---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
include/fsl_esdhc.h +0 -0
index 073048f..fa760a5 100644
@@ -168,11 +168,7 @@
 #define ESDHC_VENDORSPEC_VSELECT 0x00000002 /* Use 1.8V */
 
include/fsl_esdhc.h line 174
Quoted 10 lines
include/fsl_esdhc.h +1 -5
 struct fsl_esdhc_cfg {
-#ifdef CONFIG_FSL_LAYERSCAPE
- u64 esdhc_base;
-#else
- u32 esdhc_base;
-#endif
+ phys_addr_t esdhc_base;
 u32 sdhc_clk;
 u8 max_bus_width;
 struct mmc_config cfg;
Peng,
Quoted 5 lines
I thought this change is trivial and should be OK. But it turns out this change brings comping warning to many PPC boards
drivers/mmc/fsl_esdhc.c:184:27: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
You can apply https://patchwork.ozlabs.org/patch/601919/ and retest. I think the reason is that to PPC 64bit, "typedef unsigned long long phys_addr_t" I think "esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" can fix it.
Quoted 6 lines
Quoted 3 lines
Do I need to send a single V3 patch to fix the warning, or you apply the driver model V3 patch, if the driver model V3 can fix it?
I am concerned about git bisect. It is not good to have a patch with compiling warning. I prefer you fix it.
You can first apply https://patchwork.ozlabs.org/patch/601919/ V3 version for driver model patch. In this patch I have such code:
"esdhc_regs = (struct fsl_esdhc *)(unsigned long)(cfg->esdhc_base);" If you think this line code is ok, then you can apply this current patch. If not, I think I may also fix this.

Quoted 4 lines

Quoted 2 lines
If this is ok for you, no need for me to send V3 :)
Will try tomorrow.

Is there any update on this? If this patch with the driver model patch are ok in your side, will you pick the two patches?

Thanks,
Peng.

Quoted 4 lines

York

York Sun

2016-04-06 17:22 UTC #9 lore raw

PATCH v2 1/1 fsl: esdhc: consolidate fsl_esdhc_cfg structure

In reply to Peng Fan

Quoted 10 lines · Peng Fan

We can use phys_addr_to for esdhc_base to discard
the #ifdef.
Signed-off-by Peng Fan <[email protected]>
Cc York Sun <[email protected]>
Cc Yangbo Lu <[email protected]>
Cc Eric Nelson <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc Pantelis Antoniou <[email protected]>
Cc Tom Rini <[email protected]>

Quoted 8 lines

---
V2:
Split this patch from the V1 patch set.
 include/fsl_esdhc.h | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

Applied to u-boot-fsl-qoriq master. Awaiting upstream. Thanks.

York