linux-next: mtd tree build failure

linux-next: mtd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 937 bytes --]

Hi David,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/mtd/nand/gpio.c: In function 'gpio_nand_writebuf':
drivers/mtd/nand/gpio.c:90: error: implicit declaration of function 'writesb'
drivers/mtd/nand/gpio.c: In function 'gpio_nand_readbuf':
drivers/mtd/nand/gpio.c:97: error: implicit declaration of function 'readsb'
drivers/mtd/nand/gpio.c: In function 'gpio_nand_writebuf16':
drivers/mtd/nand/gpio.c:123: error: implicit declaration of function 'writesw'
drivers/mtd/nand/gpio.c: In function 'gpio_nand_readbuf16':
drivers/mtd/nand/gpio.c:138: error: implicit declaration of function 'readsw'

Caused by commit aaf7ea20000436df3cbb397ccb734ad1e2e5164d ("[MTD] [NAND]
GPIO NAND flash driver").  I assume this needs to depend on some
architecture/platform so I reverted it for now.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mtd tree build failure

[David Woodhouse]

On Mon, 2008-10-20 at 16:35 +1100, Stephen Rothwell wrote:
> Caused by commit aaf7ea20000436df3cbb397ccb734ad1e2e5164d ("[MTD]
> [NAND] GPIO NAND flash driver").  I assume this needs to depend on
> some architecture/platform so I reverted it for now.

Hm, indeed. I thought it _did_ depend on ARM -- did I apply the wrong
version of the patch?

-- 
David Woodhouse                            Open Source Technology Centre
David.Woodhouse@intel.com                              Intel Corporation

Re: linux-next: mtd tree build failure

[Mike Rapoport]

David Woodhouse wrote:
> On Mon, 2008-10-20 at 16:35 +1100, Stephen Rothwell wrote:
>> Caused by commit aaf7ea20000436df3cbb397ccb734ad1e2e5164d ("[MTD]
>> [NAND] GPIO NAND flash driver").  I assume this needs to depend on
>> some architecture/platform so I reverted it for now.
> 
> Hm, indeed. I thought it _did_ depend on ARM -- did I apply the wrong
> version of the patch?
> 

You applied the correct version of the patch. It did depend on ARM at the
beginning, but after discussion it was decided to make it generic so that other
architecture would be able to use it. I just haven't thought that readsw/writesw
might be not available on all architectures.


-- 
Sincerely yours,
Mike.

Re: linux-next: mtd tree build failure

[David Woodhouse]

On Mon, 2008-10-20 at 10:33 +0200, Mike Rapoport wrote:
> You applied the correct version of the patch. It did depend on ARM at the
> beginning, but after discussion it was decided to make it generic so that other
> architecture would be able to use it. I just haven't thought that readsw/writesw
> might be not available on all architectures.

I've made it depend on ARM again for now. For 2.6.29, let's look at
making use ioread8_rep().

-- 
David Woodhouse                            Open Source Technology Centre
David.Woodhouse@intel.com                              Intel Corporation

[PATCH] MTD: NAND: make gpio_nand use io{read,write}{8,16}_rep (was: Re: linux-next: mtd tree build failure)

[Mike Rapoport]

David Woodhouse wrote:
> On Mon, 2008-10-20 at 10:33 +0200, Mike Rapoport wrote:
>> You applied the correct version of the patch. It did depend on ARM at the
>> beginning, but after discussion it was decided to make it generic so that other
>> architecture would be able to use it. I just haven't thought that readsw/writesw
>> might be not available on all architectures.
> 
> I've made it depend on ARM again for now. For 2.6.29, let's look at
> making use ioread8_rep().

The below patch makes GPIO NAND use io{read,write}{8,16}_rep to allow
architectures other than ARM use the driver.
I've build-tested it for x86_64 with allmodconfig and for powerpc with
ep8248e_defconfig and manually enabled GPIO NAND driver.

Signed-off-by: Mike Rapoport <mike@compulab.co.il>
---
 drivers/mtd/nand/Kconfig |    2 +-
 drivers/mtd/nand/gpio.c  |   16 ++++++++--------
 2 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/drivers/mtd/nand/Kconfig b/drivers/mtd/nand/Kconfig
index 1c2e945..b9eed99 100644
--- a/drivers/mtd/nand/Kconfig
+++ b/drivers/mtd/nand/Kconfig
@@ -58,7 +58,7 @@ config MTD_NAND_H1900

 config MTD_NAND_GPIO
 	tristate "GPIO NAND Flash driver"
-	depends on GENERIC_GPIO && ARM
+	depends on GENERIC_GPIO
 	help
 	  This enables a GPIO based NAND flash driver.

diff --git a/drivers/mtd/nand/gpio.c b/drivers/mtd/nand/gpio.c
index 8f902e7..c9f26fa 100644
--- a/drivers/mtd/nand/gpio.c
+++ b/drivers/mtd/nand/gpio.c
@@ -87,14 +87,14 @@ static void gpio_nand_writebuf(struct mtd_info *mtd, const
u_char *buf, int len)
 {
 	struct nand_chip *this = mtd->priv;

-	writesb(this->IO_ADDR_W, buf, len);
+	iowrite8_rep(this->IO_ADDR_W, buf, len);
 }

 static void gpio_nand_readbuf(struct mtd_info *mtd, u_char *buf, int len)
 {
 	struct nand_chip *this = mtd->priv;

-	readsb(this->IO_ADDR_R, buf, len);
+	ioread8_rep(this->IO_ADDR_R, buf, len);
 }

 static int gpio_nand_verifybuf(struct mtd_info *mtd, const u_char *buf, int len)
@@ -104,7 +104,7 @@ static int gpio_nand_verifybuf(struct mtd_info *mtd, const
u_char *buf, int len)
 	int i, err = 0;

 	for (i = 0; i < len; i++) {
-		read = readb(this->IO_ADDR_R);
+		read = ioread8(this->IO_ADDR_R);
 		if (read != p[i]) {
 			pr_debug("%s: err at %d (read %04x vs %04x)\n",
 			       __func__, i, read, p[i]);
@@ -120,13 +120,13 @@ static void gpio_nand_writebuf16(struct mtd_info *mtd,
const u_char *buf,
 	struct nand_chip *this = mtd->priv;

 	if (IS_ALIGNED((unsigned long)buf, 2)) {
-		writesw(this->IO_ADDR_W, buf, len>>1);
+		iowrite16_rep(this->IO_ADDR_W, buf, len>>1);
 	} else {
 		int i;
 		unsigned short *ptr = (unsigned short *)buf;

 		for (i = 0; i < len; i += 2, ptr++)
-			writew(*ptr, this->IO_ADDR_W);
+			iowrite16(*ptr, this->IO_ADDR_W);
 	}
 }

@@ -135,13 +135,13 @@ static void gpio_nand_readbuf16(struct mtd_info *mtd,
u_char *buf, int len)
 	struct nand_chip *this = mtd->priv;

 	if (IS_ALIGNED((unsigned long)buf, 2)) {
-		readsw(this->IO_ADDR_R, buf, len>>1);
+		ioread16_rep(this->IO_ADDR_R, buf, len>>1);
 	} else {
 		int i;
 		unsigned short *ptr = (unsigned short *)buf;

 		for (i = 0; i < len; i += 2, ptr++)
-			*ptr = readw(this->IO_ADDR_R);
+			*ptr = ioread16(this->IO_ADDR_R);
 	}
 }

@@ -154,7 +154,7 @@ static int gpio_nand_verifybuf16(struct mtd_info *mtd, const
u_char *buf,
 	len >>= 1;

 	for (i = 0; i < len; i++) {
-		read = readw(this->IO_ADDR_R);
+		read = ioread16(this->IO_ADDR_R);
 		if (read != p[i]) {
 			pr_debug("%s: err at %d (read %04x vs %04x)\n",
 			       __func__, i, read, p[i]);
-- 
1.5.6.4




-- 
Sincerely yours,
Mike.

linux-next: mtd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 912 bytes --]

Hi David,

Today's linux-next build (x86_64 allmodconfig) failed like this:

drivers/mtd/maps/dilnetpc.c: In function 'init_dnpc':
drivers/mtd/maps/dilnetpc.c:448: error: 'struct mtd_partition' has no member named 'mtdp'
drivers/mtd/maps/dilnetpc.c:449: error: 'struct mtd_partition' has no member named 'mtdp'
drivers/mtd/maps/dilnetpc.c:450: error: 'struct mtd_partition' has no member named 'mtdp'
drivers/mtd/maps/dilnetpc.c:451: error: 'struct mtd_partition' has no member named 'mtdp'

Caused by commit b90cf6681f4f6263920616e7ca2fd09130e4143a ("[MTD] Remove
option for add_mtd_partitions() to not register partitions.").  The
comment in that commit even says that it will cause this breakage, so why
hasn't this been fixed?

I have used the mtd tree from next-20090526 for today.  
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

linux-next: mtd tree build failure

[Stephen Rothwell]

Hi Dave,

Today's linux-next build (x86_64 allmodconfig) failed like this:

ERROR: "qry_present" [drivers/mtd/chips/cfi_probe.ko] undefined!
ERROR: "qry_mode_off" [drivers/mtd/chips/cfi_probe.ko] undefined!
ERROR: "qry_mode_on" [drivers/mtd/chips/cfi_probe.ko] undefined!

Caused by commit 2e489e077a6ad118c4f247faedf330117b107cce ("[MTD] [NOR]
Add qry_mode_on()/qry_omde_off() to deal with odd chips").

I added the following patch.

Those names are very generic for kernel global symbols ...
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

>From 6e7b82917d1b8621c6520581c3e036956d16f8ab Mon Sep 17 00:00:00 2001
From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Thu, 7 Aug 2008 13:28:04 +1000
Subject: [PATCH] mtd: export qry_{preset,mode_on,mode_off}

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 drivers/mtd/chips/cfi_util.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/drivers/mtd/chips/cfi_util.c b/drivers/mtd/chips/cfi_util.c
index 8d75536..4a60408 100644
--- a/drivers/mtd/chips/cfi_util.c
+++ b/drivers/mtd/chips/cfi_util.c
@@ -50,6 +50,7 @@ int __xipram qry_present(struct map_info *map, __u32 base,
 
 	return 1; 	/* "QRY" found */
 }
+EXPORT_SYMBOL_GPL(qry_present);
 
 int __xipram qry_mode_on(uint32_t base, struct map_info *map,
 				struct cfi_private *cfi)
@@ -73,12 +74,15 @@ int __xipram qry_mode_on(uint32_t base, struct map_info *map,
 	/* QRY not found */
 	return 0;
 }
+EXPORT_SYMBOL_GPL(qry_mode_on);
+
 void __xipram qry_mode_off(uint32_t base, struct map_info *map,
 				struct cfi_private *cfi)
 {
 	cfi_send_gen_cmd(0xF0, 0, base, map, cfi, cfi->device_type, NULL);
 	cfi_send_gen_cmd(0xFF, 0, base, map, cfi, cfi->device_type, NULL);
 }
+EXPORT_SYMBOL_GPL(qry_mode_off);
 
 struct cfi_extquery *
 __xipram cfi_read_pri(struct map_info *map, __u16 adr, __u16 size, const char* name)
-- 
1.5.6.3

Re: linux-next: mtd tree build failure

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1347 bytes --]

Hi Dave,

On Thu, 7 Aug 2008 13:30:05 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Today's linux-next build (x86_64 allmodconfig) failed like this:
> 
> ERROR: "qry_present" [drivers/mtd/chips/cfi_probe.ko] undefined!
> ERROR: "qry_mode_off" [drivers/mtd/chips/cfi_probe.ko] undefined!
> ERROR: "qry_mode_on" [drivers/mtd/chips/cfi_probe.ko] undefined!
> 
> Caused by commit 2e489e077a6ad118c4f247faedf330117b107cce ("[MTD] [NOR]
> Add qry_mode_on()/qry_omde_off() to deal with odd chips").

Also from an arm cam60_defconfig build:

drivers/built-in.o: In function `cfi_probe_chip':
i2c-dev.c:(.text+0x497a4): undefined reference to `qry_mode_on'
i2c-dev.c:(.text+0x49810): undefined reference to `qry_present'
i2c-dev.c:(.text+0x49850): undefined reference to `qry_mode_off'
i2c-dev.c:(.text+0x498a0): undefined reference to `qry_mode_off'
i2c-dev.c:(.text+0x498b0): undefined reference to `qry_present'
i2c-dev.c:(.text+0x498c8): undefined reference to `qry_mode_off'
i2c-dev.c:(.text+0x498d8): undefined reference to `qry_present'
i2c-dev.c:(.text+0x49ca4): undefined reference to `qry_mode_off'

This is today's linux-next with these symbols exported (which obviously
makes no difference here).

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: mtd tree build failure

[Alexey Korolev]

Hi Stephen,

Thanks for catching export symbol problem.  
Hmm, it seems very strange why i2c_dev.c can't reference the exported symbols. 
We are looking after of the problem. I believe it will be fixed ASAP.

Thanks,
Alexey

> Hi Dave,
> 
> On Thu, 7 Aug 2008 13:30:05 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> >
> > Today's linux-next build (x86_64 allmodconfig) failed like this:
> > 
> > ERROR: "qry_present" [drivers/mtd/chips/cfi_probe.ko] undefined!
> > ERROR: "qry_mode_off" [drivers/mtd/chips/cfi_probe.ko] undefined!
> > ERROR: "qry_mode_on" [drivers/mtd/chips/cfi_probe.ko] undefined!
> > 
> > Caused by commit 2e489e077a6ad118c4f247faedf330117b107cce ("[MTD] [NOR]
> > Add qry_mode_on()/qry_omde_off() to deal with odd chips").
> 
> Also from an arm cam60_defconfig build:
> 
> drivers/built-in.o: In function `cfi_probe_chip':
> i2c-dev.c:(.text+0x497a4): undefined reference to `qry_mode_on'
> i2c-dev.c:(.text+0x49810): undefined reference to `qry_present'
> i2c-dev.c:(.text+0x49850): undefined reference to `qry_mode_off'
> i2c-dev.c:(.text+0x498a0): undefined reference to `qry_mode_off'
> i2c-dev.c:(.text+0x498b0): undefined reference to `qry_present'
> i2c-dev.c:(.text+0x498c8): undefined reference to `qry_mode_off'
> i2c-dev.c:(.text+0x498d8): undefined reference to `qry_present'
> i2c-dev.c:(.text+0x49ca4): undefined reference to `qry_mode_off'
> 
> This is today's linux-next with these symbols exported (which obviously
> makes no difference here).
> 
> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/
> 

Re: linux-next: mtd tree build failure

[David Woodhouse]

On Sat, 2008-08-09 at 21:41 +0100, Alexey Korolev wrote:
> Hi Stephen,
> 
> Thanks for catching export symbol problem.  
> Hmm, it seems very strange why i2c_dev.c can't reference the exported symbols. 
> We are looking after of the problem. I believe it will be fixed ASAP.

http://git.infradead.org/mtd-2.6.git?a=commitdiff;h=c314dfdc

-- 
dwmw2

Re: linux-next: mtd tree build failure

[Alexey Korolev]

Hi David,

> > 
> > Thanks for catching export symbol problem.  
> > Hmm, it seems very strange why i2c_dev.c can't reference the exported symbols. 
> > We are looking after of the problem. I believe it will be fixed ASAP.
> 
> http://git.infradead.org/mtd-2.6.git?a=commitdiff;h=c314dfdc
>
Oh Great! Thank you very much! 

Alexey