Skip to contentSkip to conversation

palm_bk3710 buildfix

PATCH In review David Brownell 3 participants 11 messages Jan 18, 2009 – Jan 19, 2009 libata

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 palm_bk3710 buildfix

From: David Brownell <[email protected]>
Subject: drivers/ide/palm_bk3710.c buildfix

CC drivers/ide/palm_bk3710.o
drivers/ide/palm_bk3710.c: In function 'palm_bk3710_probe':
drivers/ide/palm_bk3710.c:382: warning: assignment makes integer from pointer without a cast

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

---
 drivers/ide/palm_bk3710.c |   11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
drivers/ide/palm_bk3710.c +7 -4
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
-	unsigned long base, rate;
+	void __iomem *base;
+	unsigned long rate;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -382,11 +383,13 @@ static int __init palm_bk3710_probe(stru
 	base = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(base);
 
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
-		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
-	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;
+		hw.io_ports_array[i] = (unsigned long)
+				(base + IDE_PALM_ATA_PRI_REG_OFFSET + i);
+	hw.io_ports.ctl_addr = (unsigned long)
+			(base + IDE_PALM_ATA_PRI_CTL_OFFSET);
 	hw.irq = irq->start;
 	hw.dev = &pdev->dev;
 	hw.chipset = ide_palm3710;
#1 lore raw

Hello.

Quoted 6 lines · David Brownell

CC drivers/ide/palm_bk3710.o
drivers/ide/palm_bk3710.c: In function 'palm_bk3710_probe':
drivers/ide/palm_bk3710.c:382: warning: assignment makes integer from pointer without a cast

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O ports.

Quoted 18 lines
drivers/ide/palm_bk3710.c +2 -1
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
-	unsigned long base, rate;
+	void __iomem *base;
+	unsigned long rate;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -382,11 +383,13 @@ static int __init palm_bk3710_probe(stru
 	base = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */

drivers/ide/palm_bk3710.c static int __init palm_bk3710_probe(stru · line 392

Quoted 10 lines
drivers/ide/palm_bk3710.c +5 -3
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(base);
 
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
-		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
-	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;
+		hw.io_ports_array[i] = (unsigned long)
+				(base + IDE_PALM_ATA_PRI_REG_OFFSET + i);
+	hw.io_ports.ctl_addr = (unsigned long)
+			(base + IDE_PALM_ATA_PRI_CTL_OFFSET);

Ugh. I suggest adding another variable...

drivers/ide/palm_bk3710.c +4 -2
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
 	unsigned long base, rate;
+	void __iomem *regs;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -379,11 +380,11 @@ static int __init palm_bk3710_probe(stru
 		return -EBUSY;
 	}
 
- 	base = IO_ADDRESS(mem->start);
+ 	regs = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(regs);
 
+	base = (unsigned long)regs;
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
 		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;

hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;

MBR, Sergei

#2 lore raw
Quoted 4 lines · Sergei Shtylyov
Quoted 2 lines

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

    It cannot use pointers of course -- as the addresses can be I/O ports.

It could use the ioread*() calls, which take pointers and handle both types of I/O in a type-safe manner.

- Dave

#3 lore raw

Hello.

Quoted 11 lines · David Brownell
Quoted 5 lines · Sergei Shtylyov
Quoted 3 lines

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O ports.

It could use the ioread*() calls, which take pointers and handle both types of I/O in a type-safe manner.

Probably they can... though using those would only slow down register
access on x86.

Quoted 1 line

- Dave

WBR, Sergei

#4 lore raw
Quoted 7 lines · Hello, I wrote:
Quoted 7 lines
Quoted 6 lines
Quoted 3 lines

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O
ports.

Quoted 6 lines
Quoted 3 lines

It could use the ioread*() calls, which take pointers and handle both types of I/O in a type-safe manner.

Probably they can... though using those would only slow down
register access on x86.

... and cause code bloat due to switching to real function calls from
inlines which now should render into in/out isns.

Quoted 1 line

- Dave

WBR, Sergei

#5 lore raw
Quoted 13 lines · Sergei Shtylyov
Quoted 11 lines · Hello, I wrote:
Quoted 10 lines
Quoted 6 lines
Quoted 3 lines

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O
ports.

It could use the ioread*() calls, which take pointers and handle both types of I/O in a type-safe manner.

Quoted 5 lines
Quoted 2 lines

Probably they can... though using those would only slow down
register access on x86.

... and cause code bloat due to switching to real function calls from
inlines which now should render into in/out isns.

OTOH it would allow to unify a lot of 'if (hwif->host_flags & IDE_HFLAG_MMIO)' cases so the code may even shrink. I also suspect that register access slow down is most likely negligable (we can always measure and check it).

The real reason for not using ioread*() and co. is that not all archs support them (though I checked some time ago, maybe they do now?) and we don't want to break IDE support for them.

Thanks,
Bart

#6 lore raw
Quoted 9 lines · Bartlomiej Zolnierkiewicz
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O
ports.

It could use the ioread*() calls, which take pointers and handle both types of I/O in a type-safe manner.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Probably they can... though using those would only slow down
register access on x86.

... and cause code bloat due to switching to real function calls from
inlines which now should render into in/out isns.

OTOH it would allow to unify a lot of 'if (hwif->host_flags & IDE_HFLAG_MMIO)' cases so the code may even shrink.

It won't shrink as much as it will grow.

Quoted 2 lines

I also suspect that register access slow
down is most likely negligable (we can always measure and check it).

I'd agree here but code bloat won't be as negligible considering the IDE
code's predominant use of I/O ports on x86.

Quoted 3 lines

The real reason for not using ioread*() and co. is that not all archs support them (though I checked some time ago, maybe they do now?) and we don't want to break IDE support for them.

Sure. :-)

Quoted 2 lines

Thanks,
Bart

MBR, Sergei

#7 lore raw
Quoted 18 lines · Hello, I wrote:
Quoted 9 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O
ports.

It could use the ioread*() calls, which take pointers and handle both types of I/O in a type-safe manner.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

Probably they can... though using those would only slow down
register access on x86.

... and cause code bloat due to switching to real function calls
from inlines which now should render into in/out isns.

OTOH it would allow to unify a lot of 'if (hwif->host_flags & IDE_HFLAG_MMIO)' cases so the code may even shrink.

Quoted 7 lines

It won't shrink as much as it will grow.

Quoted 2 lines

I also suspect that register access slow
down is most likely negligable (we can always measure and check it).

I'd agree here but code bloat won't be as negligible considering the
IDE code's predominant use of I/O ports on x86.

After looking at the current code, the generic IDE code would actually
benefit from such change, while the most of the controller drivers will only be subject to the code bloat.

MBR, Sergei

#8 lore raw
In reply to Sergei Shtylyov
Quoted 60 lines · Sergei Shtylyov

Hello.

Quoted 6 lines · David Brownell

CC drivers/ide/palm_bk3710.o
drivers/ide/palm_bk3710.c: In function 'palm_bk3710_probe':
drivers/ide/palm_bk3710.c:382: warning: assignment makes integer from pointer without a cast

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O ports.

Quoted 18 lines
drivers/ide/palm_bk3710.c +2 -1
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
-	unsigned long base, rate;
+	void __iomem *base;
+	unsigned long rate;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -382,11 +383,13 @@ static int __init palm_bk3710_probe(stru
 	base = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */

drivers/ide/palm_bk3710.c static int __init palm_bk3710_probe(stru · line 392

Quoted 10 lines
drivers/ide/palm_bk3710.c +5 -3
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(base);
 
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
-		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
-	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;
+		hw.io_ports_array[i] = (unsigned long)
+				(base + IDE_PALM_ATA_PRI_REG_OFFSET + i);
+	hw.io_ports.ctl_addr = (unsigned long)
+			(base + IDE_PALM_ATA_PRI_CTL_OFFSET);

Ugh. I suggest adding another variable...

drivers/ide/palm_bk3710.c +1 -1
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
  {
  	struct clk *clk;
  	struct resource *mem, *irq;
  	unsigned long base, rate;
+	void __iomem *regs;
  	int i, rc;
  	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -379,11 +380,11 @@ static int __init palm_bk3710_probe(stru
  		return -EBUSY;
  	}
 
- 	base = IO_ADDRESS(mem->start);

drivers/ide/palm_bk3710.c static int __init palm_bk3710_probe(stru

Quoted 10 lines
drivers/ide/palm_bk3710.c +3 -1
+ 	regs = IO_ADDRESS(mem->start);
 
  	/* Configure the Palm Chip controller */
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(regs);
 
+	base = (unsigned long)regs;
  	for (i = 0; i < IDE_NR_PORTS - 2; i++)
  		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
  	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;

I applied original patch since having one additional cast has overally lower complexity than having an additional variable...

#9 lore raw

Hello.

Quoted 62 lines · Bartlomiej Zolnierkiewicz
Quoted 62 lines
Quoted 6 lines

CC drivers/ide/palm_bk3710.o
drivers/ide/palm_bk3710.c: In function 'palm_bk3710_probe':
drivers/ide/palm_bk3710.c:382: warning: assignment makes integer from pointer without a cast

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O ports.

Quoted 18 lines
drivers/ide/palm_bk3710.c +2 -1
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
-	unsigned long base, rate;
+	void __iomem *base;
+	unsigned long rate;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -382,11 +383,13 @@ static int __init palm_bk3710_probe(stru
 	base = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */

drivers/ide/palm_bk3710.c static int __init palm_bk3710_probe(stru · line 392

Quoted 10 lines
drivers/ide/palm_bk3710.c +5 -3
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(base);
 
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
-		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
-	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;
+		hw.io_ports_array[i] = (unsigned long)
+				(base + IDE_PALM_ATA_PRI_REG_OFFSET + i);
+	hw.io_ports.ctl_addr = (unsigned long)
+			(base + IDE_PALM_ATA_PRI_CTL_OFFSET);

Ugh. I suggest adding another variable...

drivers/ide/palm_bk3710.c +3 -2
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
 	unsigned long base, rate;
+	void __iomem *regs;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -379,11 +380,11 @@ static int __init palm_bk3710_probe(stru
 		return -EBUSY;
 	}
 
- 	base = IO_ADDRESS(mem->start);
+ 	regs = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(regs);
 
Quoted 7 lines

drivers/ide/palm_bk3710.c static int __init palm_bk3710_probe(stru

Quoted 4 lines
drivers/ide/palm_bk3710.c +1 -0
+	base = (unsigned long)regs;
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
 		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
 	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;

I applied original patch since having one additional cast has overally lower complexity than having an additional variable...

Grr... as if the stack use aren't optimized by gcc based on the variable
lifetimes. I can recast the patch myself BTW.

MBR, Sergei

#10 lore raw
Quoted 74 lines · Sergei Shtylyov

Hello.

Quoted 62 lines · Bartlomiej Zolnierkiewicz
Quoted 62 lines
Quoted 6 lines

CC drivers/ide/palm_bk3710.o
drivers/ide/palm_bk3710.c: In function 'palm_bk3710_probe':
drivers/ide/palm_bk3710.c:382: warning: assignment makes integer from pointer without a cast

Someone should fix hw_regs_t to neither be a typedef, nor use "unsigned long" where it should use "void __iomem *".

It cannot use pointers of course -- as the addresses can be I/O ports.

Quoted 18 lines
drivers/ide/palm_bk3710.c +2 -1
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
-	unsigned long base, rate;
+	void __iomem *base;
+	unsigned long rate;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -382,11 +383,13 @@ static int __init palm_bk3710_probe(stru
 	base = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */

drivers/ide/palm_bk3710.c static int __init palm_bk3710_probe(stru · line 392

Quoted 10 lines
drivers/ide/palm_bk3710.c +5 -3
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(base);
 
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
-		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
-	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;
+		hw.io_ports_array[i] = (unsigned long)
+				(base + IDE_PALM_ATA_PRI_REG_OFFSET + i);
+	hw.io_ports.ctl_addr = (unsigned long)
+			(base + IDE_PALM_ATA_PRI_CTL_OFFSET);

Ugh. I suggest adding another variable...

drivers/ide/palm_bk3710.c +3 -2
@@ -346,7 +346,8 @@ static int __init palm_bk3710_probe(stru
 {
 	struct clk *clk;
 	struct resource *mem, *irq;
 	unsigned long base, rate;
+	void __iomem *regs;
 	int i, rc;
 	hw_regs_t hw, *hws[] = { &hw, NULL, NULL, NULL };
 
@@ -379,11 +380,11 @@ static int __init palm_bk3710_probe(stru
 		return -EBUSY;
 	}
 
- 	base = IO_ADDRESS(mem->start);
+ 	regs = IO_ADDRESS(mem->start);
 
 	/* Configure the Palm Chip controller */
-	palm_bk3710_chipinit((void __iomem *)base);
+	palm_bk3710_chipinit(regs);
 
Quoted 7 lines

drivers/ide/palm_bk3710.c static int __init palm_bk3710_probe(stru

Quoted 4 lines
drivers/ide/palm_bk3710.c +1 -0
+	base = (unsigned long)regs;
 	for (i = 0; i < IDE_NR_PORTS - 2; i++)
 		hw.io_ports_array[i] = base + IDE_PALM_ATA_PRI_REG_OFFSET + i;
 	hw.io_ports.ctl_addr = base + IDE_PALM_ATA_PRI_CTL_OFFSET;

I applied original patch since having one additional cast has overally lower complexity than having an additional variable...

Quoted 2 lines

Grr... as if the stack use aren't optimized by gcc based on the variable
lifetimes. I can recast the patch myself BTW.

I meant code complexity in the maintainance / readability context here.

You know what this variable is for (at least now but a year from now it may result in a brief WTF moment when looking at base/regs) and that it is short lived but some random person who will need to update this driver won't...

Thanks,
Bart