mtd->size overflow

mtd->size overflow

[He Yong]

Hi,
  I'm using two 2G Bbyte nand flash chip (totally 4GB) under kernel 2.6.20.4,
the   chip->chipsize=0x80000000, and when decide the mtd->size = 2 *
0x80000000 = 0x00000000 (overflowed).
is there an official way to solve this problem?

Thanks!!

-- 
Best Regards!

He Yong

School of Information Security,
Shanghai Jiaotong University,
Dong chuan Road #800,
Minhang, Shanghai, P.R.China

Re: mtd->size overflow

[He Yong]

hi,Vinit
  can you give a patch fro that?  Thanks a lot!

-- He Yong

2007/8/29, Vinit Agnihotri <vinit.agnihotri@gmail.com>:
> I had same problem,
> I converted entire mtd sunsytem to work with 64-bit compatibility.
> You have make fields from mtd_info, erase_info & 2 more structures to 64-bits
> these fields are size, offset & similar ones...
>
> --vinit
>
> On 8/29/07, He Yong <hoffer1127@gmail.com> wrote:
> > Hi,
> >   I'm using two 2G Bbyte nand flash chip (totally 4GB) under kernel 2.6.20.4,
> > the   chip->chipsize=0x80000000, and when decide the mtd->size = 2 *
> > 0x80000000 = 0x00000000 (overflowed).
> > is there an official way to solve this problem?
> >
> > Thanks!!
> >
> > --
> > Best Regards!
> >
> > He Yong
> >
> > School of Information Security,
> > Shanghai Jiaotong University,
> > Dong chuan Road #800,
> > Minhang, Shanghai, P.R.China
> >
> > ______________________________________________________
> > Linux MTD discussion mailing list
> > http://lists.infradead.org/mailman/listinfo/linux-mtd/
> >
>
>
> --
> I feel free now
>


-- 
Best Regards!

He Yong

School of Information Security,
Shanghai Jiaotong University,
Dong chuan Road #800,
Minhang, Shanghai, P.R.China

Re: mtd->size overflow

[Vinit Agnihotri]

I dont mind giving out patch
but its about making mtd subsytem 64-bit,
so mtd-maintainers must allow that patch. As its kernel patch.

--vinit

On 8/29/07, He Yong <hoffer1127@gmail.com> wrote:
> hi,Vinit
>   can you give a patch fro that?  Thanks a lot!
>
> -- He Yong
>
> 2007/8/29, Vinit Agnihotri <vinit.agnihotri@gmail.com>:
> > I had same problem,
> > I converted entire mtd sunsytem to work with 64-bit compatibility.
> > You have make fields from mtd_info, erase_info & 2 more structures to 64-bits
> > these fields are size, offset & similar ones...
> >
> > --vinit
> >
> > On 8/29/07, He Yong <hoffer1127@gmail.com> wrote:
> > > Hi,
> > >   I'm using two 2G Bbyte nand flash chip (totally 4GB) under kernel 2.6.20.4,
> > > the   chip->chipsize=0x80000000, and when decide the mtd->size = 2 *
> > > 0x80000000 = 0x00000000 (overflowed).
> > > is there an official way to solve this problem?
> > >
> > > Thanks!!
> > >
> > > --
> > > Best Regards!
> > >
> > > He Yong
> > >
> > > School of Information Security,
> > > Shanghai Jiaotong University,
> > > Dong chuan Road #800,
> > > Minhang, Shanghai, P.R.China
> > >
> > > ______________________________________________________
> > > Linux MTD discussion mailing list
> > > http://lists.infradead.org/mailman/listinfo/linux-mtd/
> > >
> >
> >
> > --
> > I feel free now
> >
>
>
> --
> Best Regards!
>
> He Yong
>
> School of Information Security,
> Shanghai Jiaotong University,
> Dong chuan Road #800,
> Minhang, Shanghai, P.R.China
>


-- 
I feel free now

Re: mtd->size overflow

[He Yong]

oh..  I'll try to figure it out myself then.
Thanks anyway!

-- heyong

2007/8/29, Vinit Agnihotri <vinit.agnihotri@gmail.com>:
> I dont mind giving out patch
> but its about making mtd subsytem 64-bit,
> so mtd-maintainers must allow that patch. As its kernel patch.
>
> --vinit
>
> On 8/29/07, He Yong <hoffer1127@gmail.com> wrote:
> > hi,Vinit
> >   can you give a patch fro that?  Thanks a lot!
> >
> > -- He Yong
> >
> > 2007/8/29, Vinit Agnihotri <vinit.agnihotri@gmail.com>:
> > > I had same problem,
> > > I converted entire mtd sunsytem to work with 64-bit compatibility.
> > > You have make fields from mtd_info, erase_info & 2 more structures to 64-bits
> > > these fields are size, offset & similar ones...
> > >
> > > --vinit
> > >
> > > On 8/29/07, He Yong <hoffer1127@gmail.com> wrote:
> > > > Hi,
> > > >   I'm using two 2G Bbyte nand flash chip (totally 4GB) under kernel 2.6.20.4,
> > > > the   chip->chipsize=0x80000000, and when decide the mtd->size = 2 *
> > > > 0x80000000 = 0x00000000 (overflowed).
> > > > is there an official way to solve this problem?
> > > >
> > > > Thanks!!
> > > >
> > > > --
> > > > Best Regards!
> > > >
> > > > He Yong
> > > >
> > > > School of Information Security,
> > > > Shanghai Jiaotong University,
> > > > Dong chuan Road #800,
> > > > Minhang, Shanghai, P.R.China
> > > >
> > > > ______________________________________________________
> > > > Linux MTD discussion mailing list
> > > > http://lists.infradead.org/mailman/listinfo/linux-mtd/
> > > >
> > >
> > >
> > > --
> > > I feel free now
> > >
> >
> >
> > --
> > Best Regards!
> >
> > He Yong
> >
> > School of Information Security,
> > Shanghai Jiaotong University,
> > Dong chuan Road #800,
> > Minhang, Shanghai, P.R.China
> >
>
>
> --
> I feel free now
>


-- 
Best Regards!

He Yong

School of Information Security,
Shanghai Jiaotong University,
Dong chuan Road #800,
Minhang, Shanghai, P.R.China

Re: mtd->size overflow

[Jörn Engel]

On Wed, 29 August 2007 13:10:15 +0530, Vinit Agnihotri wrote:
> 
> I dont mind giving out patch
> but its about making mtd subsytem 64-bit,
> so mtd-maintainers must allow that patch. As its kernel patch.

Just send the patch.  Even if it doesn't get merged, reading the patch
will be useful to whoever does the actual 64bit transition.

Jörn

-- 
Joern's library part 14:
http://www.sandpile.org/

Re: mtd->size overflow

[He Yong]

Hi,all
I've done the work, and have it tested.
here is the patch:

Index: include/linux/mtd/nand.h
===================================================================
--- include/linux/mtd/nand.h	（修订版 166）
+++ include/linux/mtd/nand.h	（修订版 167）
@@ -392,7 +392,7 @@
 	int		bbt_erase_shift;
 	int		chip_shift;
 	int		numchips;
-	unsigned long	chipsize;
+	u_int64_t	chipsize;
 	int		pagemask;
 	int		pagebuf;
 	int		subpagesize;
@@ -491,8 +491,8 @@
 struct nand_bbt_descr {
 	int	options;
 	int	pages[NAND_MAX_CHIPS];
-	int	offs;
-	int	veroffs;
+	u_int64_t	offs;
+	u_int64_t	veroffs;
 	uint8_t	version[NAND_MAX_CHIPS];
 	int	len;
 	int	maxblocks;
Index: include/linux/mtd/mtd.h
===================================================================
--- include/linux/mtd/mtd.h	（修订版 166）
+++ include/linux/mtd/mtd.h	（修订版 167）
@@ -36,9 +36,9 @@
    specific to any particular block. */
 struct erase_info {
 	struct mtd_info *mtd;
-	u_int32_t addr;
-	u_int32_t len;
-	u_int32_t fail_addr;
+	u_int64_t addr;
+	u_int64_t len;
+	u_int64_t fail_addr;
 	u_long time;
 	u_long retries;
 	u_int dev;
@@ -50,7 +50,7 @@
 };

 struct mtd_erase_region_info {
-	u_int32_t offset;			/* At which this region starts, from the
beginning of the MTD */
+	u_int64_t offset;			/* At which this region starts, from the
beginning of the MTD */
 	u_int32_t erasesize;		/* For this region */
 	u_int32_t numblocks;		/* Number of blocks of erasesize in this region */
 };
@@ -92,7 +92,7 @@
 	size_t		retlen;
 	size_t		ooblen;
 	size_t		oobretlen;
-	uint32_t	ooboffs;
+	u_int64_t	ooboffs;
 	uint8_t		*datbuf;
 	uint8_t		*oobbuf;
 };
@@ -100,7 +100,7 @@
 struct mtd_info {
 	u_char type;
 	u_int32_t flags;
-	u_int32_t size;	 // Total size of the MTD
+	u_int64_t size;	 // Total size of the MTD -- by hoffer

 	/* "Major" erase size for the device. Naïve users may take this
 	 * to be the only erase size available, or may use the more detailed
Index: drivers/mtd/nand/nand_base.c
===================================================================
--- drivers/mtd/nand/nand_base.c	（修订版 166）
+++ drivers/mtd/nand/nand_base.c	（修订版 167）
@@ -2328,8 +2328,8 @@
 	       " 0x%02x, Chip ID: 0x%02x (%s %s)\n", *maf_id, dev_id,
 	       nand_manuf_ids[maf_idx].name, type->name);

-	printk(KERN_INFO "BlockSize: 0x%x, PageSize: 0x%x, OobSize: 0x%x,
BusWith:%d\n",
-           mtd->erasesize,mtd->writesize,mtd->oobsize,(chip->options
& NAND_BUSWIDTH_16) ? 16 : 8 );
+	printk(KERN_INFO "ChipSize: 0x%x, BlockSize: 0x%x, PageSize: 0x%x,
OobSize: 0x%x, BusWith:%d\n",
+           chip->chipsize,mtd->erasesize,mtd->writesize,mtd->oobsize,(chip->options
& NAND_BUSWIDTH_16) ? 16 : 8 );

 	return type;
 }
Index: drivers/mtd/mtdpart.c
===================================================================
--- drivers/mtd/mtdpart.c	（修订版 166）
+++ drivers/mtd/mtdpart.c	（修订版 167）
@@ -28,7 +28,7 @@
 struct mtd_part {
 	struct mtd_info mtd;
 	struct mtd_info *master;
-	u_int32_t offset;
+	u_int64_t offset;
 	int index;
 	struct list_head list;
 	int registered;
@@ -400,15 +400,15 @@
 				slave->offset = ((cur_offset / master->erasesize) + 1) * master->erasesize;
 				printk(KERN_NOTICE "Moving partition %d: "
 				       "0x%08x -> 0x%08x\n", i,
-				       cur_offset, slave->offset);
+				       cur_offset, (u_int32_t)slave->offset);
 			}
 		}
 		if (slave->mtd.size == MTDPART_SIZ_FULL)
 			slave->mtd.size = master->size - slave->offset;
 		cur_offset = slave->offset + slave->mtd.size;

-		printk (KERN_NOTICE "0x%08x-0x%08x : \"%s\"\n", slave->offset,
-			slave->offset + slave->mtd.size, slave->mtd.name);
+		printk (KERN_NOTICE "0x%08x-0x%08x : \"%s\"\n", (u_int32_t)slave->offset,
+			(u_int32_t)(slave->offset + slave->mtd.size), slave->mtd.name);

 		/* let's do some sanity checks */
 		if (slave->offset >= master->size) {
@@ -421,7 +421,7 @@
 		if (slave->offset + slave->mtd.size > master->size) {
 			slave->mtd.size = master->size - slave->offset;
 			printk ("mtd: partition \"%s\" extends beyond the end of device
\"%s\" -- size truncated to %#x\n",
-				parts[i].name, master->name, slave->mtd.size);
+				parts[i].name, master->name, (u_int32_t)slave->mtd.size);
 		}
 		if (master->numeraseregions>1) {
 			/* Deal with variable erase size stuff */
@@ -443,7 +443,7 @@
 		}

 		if ((slave->mtd.flags & MTD_WRITEABLE) &&
-		    (slave->offset % slave->mtd.erasesize)) {
+		    (((u_int32_t)slave->offset) % slave->mtd.erasesize  /*hoffer*/ )) {
 			/* Doesn't start on a boundary of major erase size */
 			/* FIXME: Let it be writable if it is on a boundary of _minor_
erase size though */
 			slave->mtd.flags &= ~MTD_WRITEABLE;
@@ -451,7 +451,7 @@
 				parts[i].name);
 		}
 		if ((slave->mtd.flags & MTD_WRITEABLE) &&
-		    (slave->mtd.size % slave->mtd.erasesize)) {
+		    (((u_int32_t)slave->mtd.size) % slave->mtd.erasesize /*hoffer*/ ) ) {
 			slave->mtd.flags &= ~MTD_WRITEABLE;
 			printk ("mtd: partition \"%s\" doesn't end on an erase block --
force read-only\n",
 				parts[i].name);
Index: drivers/mtd/mtdcore.c
===================================================================
--- drivers/mtd/mtdcore.c	（修订版 166）
+++ drivers/mtd/mtdcore.c	（修订版 167）
@@ -349,7 +349,7 @@
 	if (!this)
 		return 0;

-	return sprintf(buf, "mtd%d: %8.8x %8.8x \"%s\"\n", i, this->size,
+	return sprintf(buf, "mtd%d: %8.8x %8.8x \"%s\"\n", i, (u_int32_t)this->size,
 		       this->erasesize, this->name);
 }



2007/8/30, Jörn Engel <joern@logfs.org>:
> On Wed, 29 August 2007 13:10:15 +0530, Vinit Agnihotri wrote:
> >
> > I dont mind giving out patch
> > but its about making mtd subsytem 64-bit,
> > so mtd-maintainers must allow that patch. As its kernel patch.
>
> Just send the patch.  Even if it doesn't get merged, reading the patch
> will be useful to whoever does the actual 64bit transition.
>
> Jörn
>
> --
> Joern's library part 14:
> http://www.sandpile.org/
>


-- 
Best Regards!

He Yong

School of Information Security,
Shanghai Jiaotong University,
Dong chuan Road #800,
Minhang, Shanghai, P.R.China

Re: mtd->size overflow

[Vinit Agnihotri]

Its good work
however few things are missing
1> You need to make erasesize(struct mtd_info & struct
mtd_erase_region_info) field 64-bit
2> printk (KERN_NOTICE "0x%08x-0x%08x : \"%s\"\n", (u_int32_t)slave->offset,
			(u_int32_t)(slave->offset + slave->mtd.size), slave->mtd.name);
use %lx or 0x%16x to print 64-bit values, or else printk will print
only 32-bit value.
3> same thing in sysfs size variable(mtdcore.c).
4> You need to make changes in include/mtd/mtd-abi.h.i.e. for
         a> struct erase_info_user
         b> struct mtd_oob_buf
         c> struct mtd_info_user
         d> struct region_info_user
         e> struct otp_info
         f> struct nand_oobfree

   5> You need to make changes in include/linux/mtd/partitions.h
     a> struct mtd_partition


-- Vinit.




On 8/30/07, He Yong <hoffer1127@gmail.com> wrote:
> Hi,all
> I've done the work, and have it tested.
> here is the patch:
>
> Index: include/linux/mtd/nand.h
> ===================================================================
> --- include/linux/mtd/nand.h    （修订版 166）
> +++ include/linux/mtd/nand.h    （修订版 167）
> @@ -392,7 +392,7 @@
>         int             bbt_erase_shift;
>         int             chip_shift;
>         int             numchips;
> -       unsigned long   chipsize;
> +       u_int64_t       chipsize;
>         int             pagemask;
>         int             pagebuf;
>         int             subpagesize;
> @@ -491,8 +491,8 @@
>  struct nand_bbt_descr {
>         int     options;
>         int     pages[NAND_MAX_CHIPS];
> -       int     offs;
> -       int     veroffs;
> +       u_int64_t       offs;
> +       u_int64_t       veroffs;
>         uint8_t version[NAND_MAX_CHIPS];
>         int     len;
>         int     maxblocks;
> Index: include/linux/mtd/mtd.h
> ===================================================================
> --- include/linux/mtd/mtd.h     （修订版 166）
> +++ include/linux/mtd/mtd.h     （修订版 167）
> @@ -36,9 +36,9 @@
>     specific to any particular block. */
>  struct erase_info {
>         struct mtd_info *mtd;
> -       u_int32_t addr;
> -       u_int32_t len;
> -       u_int32_t fail_addr;
> +       u_int64_t addr;
> +       u_int64_t len;
> +       u_int64_t fail_addr;
>         u_long time;
>         u_long retries;
>         u_int dev;
> @@ -50,7 +50,7 @@
>  };
>
>  struct mtd_erase_region_info {
> -       u_int32_t offset;                       /* At which this region starts, from the
> beginning of the MTD */
> +       u_int64_t offset;                       /* At which this region starts, from the
> beginning of the MTD */
>         u_int32_t erasesize;            /* For this region */
>         u_int32_t numblocks;            /* Number of blocks of erasesize in this region */
>  };
> @@ -92,7 +92,7 @@
>         size_t          retlen;
>         size_t          ooblen;
>         size_t          oobretlen;
> -       uint32_t        ooboffs;
> +       u_int64_t       ooboffs;
>         uint8_t         *datbuf;
>         uint8_t         *oobbuf;
>  };
> @@ -100,7 +100,7 @@
>  struct mtd_info {
>         u_char type;
>         u_int32_t flags;
> -       u_int32_t size;  // Total size of the MTD
> +       u_int64_t size;  // Total size of the MTD -- by hoffer
>
>         /* "Major" erase size for the device. Naïve users may take this
>          * to be the only erase size available, or may use the more detailed
> Index: drivers/mtd/nand/nand_base.c
> ===================================================================
> --- drivers/mtd/nand/nand_base.c        （修订版 166）
> +++ drivers/mtd/nand/nand_base.c        （修订版 167）
> @@ -2328,8 +2328,8 @@
>                " 0x%02x, Chip ID: 0x%02x (%s %s)\n", *maf_id, dev_id,
>                nand_manuf_ids[maf_idx].name, type->name);
>
> -       printk(KERN_INFO "BlockSize: 0x%x, PageSize: 0x%x, OobSize: 0x%x,
> BusWith:%d\n",
> -           mtd->erasesize,mtd->writesize,mtd->oobsize,(chip->options
> & NAND_BUSWIDTH_16) ? 16 : 8 );
> +       printk(KERN_INFO "ChipSize: 0x%x, BlockSize: 0x%x, PageSize: 0x%x,
> OobSize: 0x%x, BusWith:%d\n",
> +           chip->chipsize,mtd->erasesize,mtd->writesize,mtd->oobsize,(chip->options
> & NAND_BUSWIDTH_16) ? 16 : 8 );
>
>         return type;
>  }
> Index: drivers/mtd/mtdpart.c
> ===================================================================
> --- drivers/mtd/mtdpart.c       （修订版 166）
> +++ drivers/mtd/mtdpart.c       （修订版 167）
> @@ -28,7 +28,7 @@
>  struct mtd_part {
>         struct mtd_info mtd;
>         struct mtd_info *master;
> -       u_int32_t offset;
> +       u_int64_t offset;
>         int index;
>         struct list_head list;
>         int registered;
> @@ -400,15 +400,15 @@
>                                 slave->offset = ((cur_offset / master->erasesize) + 1) * master->erasesize;
>                                 printk(KERN_NOTICE "Moving partition %d: "
>                                        "0x%08x -> 0x%08x\n", i,
> -                                      cur_offset, slave->offset);
> +                                      cur_offset, (u_int32_t)slave->offset);
>                         }
>                 }
>                 if (slave->mtd.size == MTDPART_SIZ_FULL)
>                         slave->mtd.size = master->size - slave->offset;
>                 cur_offset = slave->offset + slave->mtd.size;
>
> -               printk (KERN_NOTICE "0x%08x-0x%08x : \"%s\"\n", slave->offset,
> -                       slave->offset + slave->mtd.size, slave->mtd.name);
> +               printk (KERN_NOTICE "0x%08x-0x%08x : \"%s\"\n", (u_int32_t)slave->offset,
> +                       (u_int32_t)(slave->offset + slave->mtd.size), slave->mtd.name);
>
>                 /* let's do some sanity checks */
>                 if (slave->offset >= master->size) {
> @@ -421,7 +421,7 @@
>                 if (slave->offset + slave->mtd.size > master->size) {
>                         slave->mtd.size = master->size - slave->offset;
>                         printk ("mtd: partition \"%s\" extends beyond the end of device
> \"%s\" -- size truncated to %#x\n",
> -                               parts[i].name, master->name, slave->mtd.size);
> +                               parts[i].name, master->name, (u_int32_t)slave->mtd.size);
>                 }
>                 if (master->numeraseregions>1) {
>                         /* Deal with variable erase size stuff */
> @@ -443,7 +443,7 @@
>                 }
>
>                 if ((slave->mtd.flags & MTD_WRITEABLE) &&
> -                   (slave->offset % slave->mtd.erasesize)) {
> +                   (((u_int32_t)slave->offset) % slave->mtd.erasesize  /*hoffer*/ )) {
>                         /* Doesn't start on a boundary of major erase size */
>                         /* FIXME: Let it be writable if it is on a boundary of _minor_
> erase size though */
>                         slave->mtd.flags &= ~MTD_WRITEABLE;
> @@ -451,7 +451,7 @@
>                                 parts[i].name);
>                 }
>                 if ((slave->mtd.flags & MTD_WRITEABLE) &&
> -                   (slave->mtd.size % slave->mtd.erasesize)) {
> +                   (((u_int32_t)slave->mtd.size) % slave->mtd.erasesize /*hoffer*/ ) ) {
>                         slave->mtd.flags &= ~MTD_WRITEABLE;
>                         printk ("mtd: partition \"%s\" doesn't end on an erase block --
> force read-only\n",
>                                 parts[i].name);
> Index: drivers/mtd/mtdcore.c
> ===================================================================
> --- drivers/mtd/mtdcore.c       （修订版 166）
> +++ drivers/mtd/mtdcore.c       （修订版 167）
> @@ -349,7 +349,7 @@
>         if (!this)
>                 return 0;
>
> -       return sprintf(buf, "mtd%d: %8.8x %8.8x \"%s\"\n", i, this->size,
> +       return sprintf(buf, "mtd%d: %8.8x %8.8x \"%s\"\n", i, (u_int32_t)this->size,
>                        this->erasesize, this->name);
>  }
>
>
>
> 2007/8/30, Jörn Engel <joern@logfs.org>:
> > On Wed, 29 August 2007 13:10:15 +0530, Vinit Agnihotri wrote:
> > >
> > > I dont mind giving out patch
> > > but its about making mtd subsytem 64-bit,
> > > so mtd-maintainers must allow that patch. As its kernel patch.
> >
> > Just send the patch.  Even if it doesn't get merged, reading the patch
> > will be useful to whoever does the actual 64bit transition.
> >
> > Jörn
> >
> > --
> > Joern's library part 14:
> > http://www.sandpile.org/
> >
>
>
> --
> Best Regards!
>
> He Yong
>
> School of Information Security,
> Shanghai Jiaotong University,
> Dong chuan Road #800,
> Minhang, Shanghai, P.R.China
>


-- 
I feel free now

Re: mtd->size overflow

[Matthieu CASTET]

> 2007/8/29, Vinit Agnihotri <vinit.agnihotri@gmail.com>:
>> I had same problem,
>> I converted entire mtd sunsytem to work with 64-bit compatibility.
>> You have make fields from mtd_info, erase_info & 2 more structures to 64-bits
>> these fields are size, offset & similar ones...
Or you could sift the size field, because I bet you never have 1B flash...


Matthieu