[U-Boot] uboot kernel rejects filesystem

[U-Boot] uboot kernel rejects filesystem

[Bharat Bhushan]

Hi All,

I am very new to u-boot.

I have kernel image, which I loaded to NOR of my board and with bootm,
control passes to kernel and it starts booting.
There is panic in kernel: since it can not find the file-system. I am
afraid, I did make trivial mistake.

I have flashed squashfs filesystem image to NOR @ address NOR1 <I did not
define any partition>.
I do not have MTD support.

Now I tried booting like after setting


setenv kernel_addr 0xBF060000        <<Note this has u-boot hdr generated
form mkimage>>
setenv ramdisk_addr 0xBF260000    <<Note this does not have u-boot hdr
generated form mkimage>>
setenv bootargs 'filesystem=squashfs'
I do bootm ${kernel_addr} ${ramdisk_addr}

I do get error from U_BOOT that :
*Wrong Ramdisk Image Format
Ramdisk image is corrupt or invalid


*I am sure that I have done something wrong in this process:

Few Questions which I could not find the answer:
1. Can I nor-flash the image without u-boot hdr.
2. Is it necessary to have MTD partition defined.
3. Do we have any document which specify the boot arguments in context of
U_BOOT. I tried denx website. I could not figure out much. May be I do not
what to search for.


Any help would be highly appreciated.

Regards,
Bharat

[U-Boot] uboot kernel rejects filesystem

[Detlev Zundel]

Hi Bharat,

> I am very new to u-boot.

Welcome aboard!

> I have kernel image, which I loaded to NOR of my board and with bootm,
> control passes to kernel and it starts booting.
> There is panic in kernel: since it can not find the file-system. I am
> afraid, I did make trivial mistake.
>
> I have flashed squashfs filesystem image to NOR @ address NOR1 <I did not
> define any partition>.
> I do not have MTD support.
>
> Now I tried booting like after setting
>
>
> setenv kernel_addr 0xBF060000        <<Note this has u-boot hdr generated
> form mkimage>>
> setenv ramdisk_addr 0xBF260000    <<Note this does not have u-boot hdr
> generated form mkimage>>
> setenv bootargs 'filesystem=squashfs'
> I do bootm ${kernel_addr} ${ramdisk_addr}

This is not correct.  The second argument for bootm is a "real ramdisk"
or "initrd" in Linux speak.  The squashfs filesystem that you use needs
a MTD device and thus can not be passed in this way.

> I do get error from U_BOOT that :
> *Wrong Ramdisk Image Format
> Ramdisk image is corrupt or invalid

That's to be expected, because U-Boot wants the initrd to be wrapped
with mkimage also - which is irrelevant in your case though.

> *I am sure that I have done something wrong in this process:
>
> Few Questions which I could not find the answer:
> 1. Can I nor-flash the image without u-boot hdr.

Sure you can.  You should flash it to the address of a partition defined
inside of the Linux MTD subsystem.

> 2. Is it necessary to have MTD partition defined.

You will of course need to define MTD partitions inside of Linux -
whereas in U-Boot this is optional.

> 3. Do we have any document which specify the boot arguments in context of
> U_BOOT. I tried denx website. I could not figure out much. May be I do not
> what to search for.

What exactly do you mean by "boot arguments"?

> Any help would be highly appreciated.

To solve your problem, do the following:

1. Find out which MTD device (in Linux) you want to use to hold the
   squashfs image.  I'll assume mtd0 below.
2. If you do not want to use partitioning in U-Boot then determine the
   physical address that this device maps to
3. Flash the squasfs image to that address
4. Be sure to include "root=/dev/mtdblock0" on the Linux command line so
   Linux knows where it finds its rootfs.

Cheers
  Detlev

-- 
We support democracy, but that doesn't mean we have to support
governments that get elected as a result of democracy.
                                     -- George W. Bush - March 30, 2006
--
DENX Software Engineering GmbH,      MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich,  Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu at denx.de

[U-Boot] uboot kernel rejects filesystem

[Detlev Zundel]

Hi Bharat,

[re-adding the mailing list as others may also profit from the
discussion]

> Thank you for the prompt reply.
>
> Unfortunately in linux tree, my board specific MTD partition info is not
> preset.
>
> Is there any quick way round.

Yep - thinking some more, I see two possible (quick) solutions.

> M/M Layout:
> 0xBF000000 : U-BOOT
> 0xBF040000 : u_boot_config_area
> 0xBF060000: vm_linux_image (with u-boot header : created by mkimage and LZMA)
> 0xBF260000: root_fs area... Here I have flashed squashfs (without u-boot hdr)
> Should I?
> 0xBFFC0000: root_fs_data .. JFFS RW filesystem

Just as a sidenote - are you able to use the JFFS (really JFFS, not
JFFS2?) image?  If so, then you _do_ have an MTD partition in Linux
which you only need to tweak.

> I do following on u-boot prompt
>
> board>bootm 0xBF060000 0xBF260000

This command only makes sense, if the image at 0xbf260000 is wrapped
with mkimage.  This is exactly the first solution that you can try,
i.e. use mkimage with -T ramdisk to wrap your image and use the it with
the two argument bootm version like you tried.  Maybe you will need to
specify "rootfstype=squashfs" on the command line also, you will have to
check this.

Having said this, if you are on an ARM platform, the above will very
likely not work, so you will have to turn to the other solution I have
in mind:

> NO SUCCESS
> again I try:
> with modification to
> include/configs/my_board.h file as
>
> #define MTDIDS_DEFAULT "nor0=app"
> #define MTDPARTS_DEFAULT \
> ??? "mtdparts=app:64k(u-boot),64k(boot_config),64k(board_config),2048k
> (kernel),13696(rootfs),256kb(roots_data)"
> #define CONFIG_BOOTARGS "root=/dev/mtdblock4 rootfstype=squashfs console=
> ttyS1,115200"
>
> board>bootm 0xbf060000

The "mtdparts" variable in U-Boot has no automagic connection to the
Linux kernel.  It was designed however with connecting in Linux by using
the following Linux config:

drivers/mtd/Kconfig:106
 "Device Drivers"
config MTD_CMDLINE_PARTS
        bool "Command line partition table parsing"
        depends on MTD_PARTITIONS = "y" && MTD = "y"
        ---help---
          Allow generic configuration of the MTD partition tables via the kernel
          command line. Multiple flash resources are supported for hardware where
          different kinds of flash memory are available.

So if you have this support in your Linux configuration, you can pass
the whole "mtdpart=.." through the linux commandline.  A nasty part here
is the beginning (you chose "app") of the configuration - as you will
need to find what your kernel thinks the name of the flash chip is.
 
> I think in second case it does make sense to pass mtdblock4 since my kernel
> (;;) does not recognize it.
>
> Is there any work around like adding mtd table in linux kernel from uboot via
> bootargs.

Yes, like I stated above.

Please try either method - if you are trying via MTD, be sure to include
the contents of /proc/mtd and the relevant section of the Linux bootlog
where the MTD layer registers in your next mail.

Thanks
  Detlev

PS: And please stop quoting the full text of the whole e-mail thread as
    this is considered to be impolite on mailing lists.

-- 
Each language has its purpose, however humble. Each language expresses
the Yin and Yang of software. Each language has its place within the Tao.

But do not program in COBOL if you can avoid it.
                           -- The Tao of Programming
--
DENX Software Engineering GmbH,      MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich,  Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu at denx.de

[U-Boot] uboot kernel rejects filesystem

[Bharat Bhushan]

Hi Detlev,

On Tue, May 12, 2009 at 9:10 PM, Detlev Zundel <dzu@denx.de> wrote:

> Hi Bharat,
>
> [re-adding the mailing list as others may also profit from the
> discussion]
>
> > Thank you for the prompt reply.
> >
> > Unfortunately in linux tree, my board specific MTD partition info is not
> > preset.
> >
> > Is there any quick way round.
>
> Yep - thinking some more, I see two possible (quick) solutions.
>
> > M/M Layout:
> > 0xBF000000 : U-BOOT
> > 0xBF040000 : u_boot_config_area
> > 0xBF060000: vm_linux_image (with u-boot header : created by mkimage and
> LZMA)
> > 0xBF260000: root_fs area... Here I have flashed squashfs (without u-boot
> hdr)
> > Should I?
> > 0xBFFC0000: root_fs_data .. JFFS RW filesystem
>
> Just as a sidenote - are you able to use the JFFS (really JFFS, not
> JFFS2?) image?  If so, then you _do_ have an MTD partition in Linux
> which you only need to tweak.
>
> > I do following on u-boot prompt
> >
> > board>bootm 0xBF060000 0xBF260000
>
> This command only makes sense, if the image at 0xbf260000 is wrapped
> with mkimage.  This is exactly the first solution that you can try,
> i.e. use mkimage with -T ramdisk to wrap your image and use the it with
> the two argument bootm version like you tried.  Maybe you will need to
> specify "rootfstype=squashfs" on the command line also, you will have to
> check this.
>
> Having said this, if you are on an ARM platform, the above will very
> likely not work, so you will have to turn to the other solution I have
> in mind:
>
> > NO SUCCESS
> > again I try:
> > with modification to
> > include/configs/my_board.h file as
> >
> > #define MTDIDS_DEFAULT "nor0=app"
> > #define MTDPARTS_DEFAULT \
> >     "mtdparts=app:64k(u-boot),64k(boot_config),64k(board_config),2048k
> > (kernel),13696(rootfs),256kb(roots_data)"
> > #define CONFIG_BOOTARGS "root=/dev/mtdblock4 rootfstype=squashfs console=
> > ttyS1,115200"
> >
> > board>bootm 0xbf060000
>
I have modified in kernel for mtd support. in U_BOOT following is the info:

u-b>flinfo
Bank # 1: 256x64K=16MB, Spansion SL25FL128P (0x012018)
u-b>mtdparts
device nor0 <m25p80>, # parts = 5
 #: name                        size            offset          mask_flags
 0: boot                0x00040000      0x00000000      0
 1: boot_config         0x00010000      0x00040000      0
 2: board_config        0x00010000      0x00050000      0
 3: kernel              0x00200000      0x00060000      0
 4: filesystem          0x00400000      0x00260000      0

active partition: nor0,0 - (boot) 0x00040000 @ 0x00000000

defaults:
mtdids  : nor0=m25p80
mtdparts:
mtdparts=m25p80:256k(boot),64k(boot_config),64k(board_config),2048k(kernel),4m(filesystem)

u-b>bootm 0xbf06000
Log messages from kernel:


## Booting kernel from Legacy Image at bf060000 ...
   Image Name:   MIPS OpenWrt Linux-2.6.28.9
   Image Type:   MIPS Linux Kernel Image (lzma compressed)
   Data Size:    767717 Bytes = 749.7 kB
   Load Address: 80060000
   Entry Point:  80060000
   Verifying Checksum ... OK
## No init Ramdisk
   ramdisk start = 0x00000000, ramdisk end = 0x00000000
   Uncompressing Kernel Image ... OK
linux_argv[0]:<NULL>
linux_env a3f9c3c0
theKernel 80060000

Starting kernel ...

Linux version 2.6.28.9 (bharat at debian) (gcc version 4.1.2) #5 Thu May 14
22:15:40 IST 2009
console [early0] enabled
CPU revision is: 00019374 (MIPS 24Kc)
Atheros AR7161 rev 2 (id:0xaa), CPU:680.000 MHz, AHB:170.000 MHz,
DDR:340.000 MHz
Determined physical RAM map:
 memory: 04000000 @ 00000000 (usable)
Initrd not found or empty - disabling initrd
Zone PFN ranges:
  Normal   0x00000000 -> 0x00004000
Movable zone start PFN for each node
early_node_map[1] active PFN ranges
    0: 0x00000000 -> 0x00004000
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 16256
Kernel command line: root=/dev/mtdpart4 rootfstype=squashfs,yaffs,jffs2
noinitrd console=ttyS0,115200 init=/etc/preinit
Primary instruction cache 64kB, VIPT, 4-way, linesize 32 bytes.
Primary data cache 32kB, 4-way, VIPT, cache aliases, linesize 32 bytes
Writing ErrCtl register=00001c10
Readback ErrCtl register=00001c10
PID hash table entries: 256 (order: 8, 1024 bytes)
Dentry cache hash table entries: 8192 (order: 3, 32768 bytes)
Inode-cache hash table entries: 4096 (order: 2, 16384 bytes)
Memory: 62120k/65536k available (1820k kernel code, 3340k reserved, 330k
data, 132k init, 0k highmem)
SLUB: Genslabs=6, HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Calibrating delay loop... 452.19 BogoMIPS (lpj=2260992)
Mount-cache hash table entries: 512
net_namespace: 480 bytes
NET: Registered protocol family 16
MIPS: machine is Generic AR71xx board
NET: Registered protocol family 2
IP route cache hash table entries: 1024 (order: 0, 4096 bytes)
TCP established hash table entries: 2048 (order: 2, 16384 bytes)
TCP bind hash table entries: 2048 (order: 1, 8192 bytes)
TCP: Hash tables configured (established 2048 bind 2048)
TCP reno registered
NET: Registered protocol family 1
squashfs: version 3.0 (2006/03/15) Phillip Lougher
Registering mini_fo version $Id$
JFFS2 version 2.2. (NAND) (SUMMARY)  ??? 2001-2006 Red Hat, Inc.
yaffs May 13 2009 12:26:38 Installing.
msgmni has been set to 121
io scheduler noop registered
io scheduler deadline registered (default)
Serial: 8250/16550 driver1 ports, IRQ sharing disabled
serial8250.0: ttyS0 at MMIO 0x18020000 (irq = 11) is a 16550A
console handover: boot [early0] -> real [ttyS0]
Atheros AR71xx SPI Controller driver version 0.2.2
Atheros AR71xx hardware watchdog driver version 0.1.0
TCP vegas registered
NET: Registered protocol family 17
802.1Q VLAN Support v1.8 Ben Greear <greearb@candelatech.com>
*All bugs added by David S. Miller <davem@redhat.com>
VFS: Cannot open root device "mtdpart4" or unknown-block(0,0)
Please append a correct "root=" boot option; here are the available
partitions:
Kernel panic - not syncing: VFS: Unable to mount root fs on
unknown-block(0,0)
*


I am sure I am doing something wrong here.....? It is MIPS Architecture +
Linux 2.6.28...





>
>
> The "mtdparts" variable in U-Boot has no automagic connection to the
> Linux kernel.  It was designed however with connecting in Linux by using
> the following Linux config:
>
> drivers/mtd/Kconfig:106
>  "Device Drivers"
> config MTD_CMDLINE_PARTS
>        bool "Command line partition table parsing"
>        depends on MTD_PARTITIONS = "y" && MTD = "y"
>        ---help---
>          Allow generic configuration of the MTD partition tables via the
> kernel
>          command line. Multiple flash resources are supported for hardware
> where
>          different kinds of flash memory are available.
>
> So if you have this support in your Linux configuration, you can pass
> the whole "mtdpart=.." through the linux commandline.  A nasty part here
> is the beginning (you chose "app") of the configuration - as you will
> need to find what your kernel thinks the name of the flash chip is.
>
> > I think in second case it does make sense to pass mtdblock4 since my
> kernel
> > (;;) does not recognize it.
> >
> > Is there any work around like adding mtd table in linux kernel from uboot
> via
> > bootargs.
>
> Yes, like I stated above.
>
> Please try either method - if you are trying via MTD, be sure to include
> the contents of /proc/mtd and the relevant section of the Linux bootlog
> where the MTD layer registers in your next mail.
>
> Thanks
>  Detlev
>
> PS: And please stop quoting the full text of the whole e-mail thread as
>    this is considered to be impolite on mailing lists.
>
> --
> Each language has its purpose, however humble. Each language expresses
> the Yin and Yang of software. Each language has its place within the Tao.
>
> But do not program in COBOL if you can avoid it.
>                           -- The Tao of Programming
> --
> DENX Software Engineering GmbH,      MD: Wolfgang Denk & Detlev Zundel
> HRB 165235 Munich,  Office: Kirchenstr.5, D-82194 Groebenzell, Germany
> Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu at denx.de
>



-- 
Bharat

[U-Boot] uboot kernel rejects filesystem

[Wolfgang Denk]

Dear Bharat Bhushan,

In message <a29be8140905141245r10b62c47u3a27a28b93bdba19@mail.gmail.com> you wrote:
> 
> u-b>mtdparts
> device nor0 <m25p80>, # parts =3D 5
>  #: name                        size            offset          mask_flags
>  0: boot                0x00040000      0x00000000      0
>  1: boot_config         0x00010000      0x00040000      0
>  2: board_config        0x00010000      0x00050000      0
>  3: kernel              0x00200000      0x00060000      0
>  4: filesystem          0x00400000      0x00260000      0
> 
> active partition: nor0,0 - (boot) 0x00040000 @ 0x00000000
> 
> defaults:
> mtdids  : nor0=3Dm25p80
> mtdparts:
> mtdparts=m25p80:256k(boot),64k(boot_config),64k(board_config),2048k(kernel),4m(filesystem)

If this is what you want, you should let the kenrel know about that...

> Log messages from kernel:
...
> Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 16256
> Kernel command line: root=/dev/mtdpart4 rootfstype=squashfs,yaffs,jffs2
> noinitrd console=ttyS0,115200 init=/etc/preinit

... but your kernel's boot arguments show no trace of any mtdparts=
argument.

> I am sure I am doing something wrong here.....? It is MIPS Architecture +
> Linux 2.6.28...

Indeed. You did not read what Detlev explained to you:

> > So if you have this support in your Linux configuration, you can pass
-------------------------------------------------------------^^^^^^^^^^^^
> > the whole "mtdpart=.." through the linux commandline.  A nasty part here
----^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


Seems you also didn't read Detlev's PS:


> > PS: And please stop quoting the full text of the whole e-mail thread as
> >    this is considered to be impolite on mailing lists.


:-(

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Anyone who isn't confused here doesn't really know what's going on.

[U-Boot] uboot kernel rejects filesystem

[Detlev Zundel]

Hi Bharat,

> I have modified in kernel for mtd support. in U_BOOT following is the info:

Apart from the things which Wolfgang pointed out - I do *not* see any
MTD related info in your Linux kernel log.  Are you sure yoiu have mtd
support?  For example when I boot a sequoia (440epx amcc) board, I see
this:

1fc000000.nor_flash: Found 1 x16 devices at 0x0 in 16-bit bank
 Amd/Fujitsu Extended Query Table at 0x0040
1fc000000.nor_flash: CFI does not contain boot bank location. Assuming top.
number of CFI chips: 1
cfi_cmdset_0002: Disabling erase-suspend-program due to code brokenness.
mtd: no closing ) found in partition name
RedBoot partition parsing not available
Creating 5 MTD partitions on "1fc000000.nor_flash":
0x000000000000-0x000000180000 : "kernel"
0x000000180000-0x000000380000 : "root"
0x000000380000-0x000003f20000 : "user"
0x000003f60000-0x000003fa0000 : "env"
0x000003fa0000-0x000004000000 : "u-boot"

So recheck your kernel configuration.

Cheers
  Detlev

-- 
Less talking -- more hacking
                                        -- Olin Shivers
--
DENX Software Engineering GmbH,      MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich,  Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu at denx.de

[U-Boot] uboot kernel rejects filesystem

[Bharat Bhushan]

Hi Detlev,

I have gone thru the kernel code. I could only figure out that m25p80
is spi driver which is register as spi-driver. Registration is done,
which I could see from early printk's. But this module is never
probed. Next, kernel iterates through the list of devices and tries to
find out mtdpart, which it can not see. Since your mail gave very good
pointer where to look for with ref to another nor flash.


Regards,
Bharat

Seems I quoted text in previous mail which annoyed you. Sorry for
being naive. I hope I will not repeat the same.



-- 
Bharat

[U-Boot] uboot kernel rejects filesystem

[Detlev Zundel]

Hi Bharat,

> I have gone thru the kernel code. I could only figure out that m25p80
> is spi driver which is register as spi-driver. 

Oh wow, an SPI NOR flash.  In my opinion, this is a rare constellation,
so you will probably need to debug MTD in Linux.

> Registration is done, which I could see from early printk's. But this
> module is never probed. Next, kernel iterates through the list of
> devices and tries to find out mtdpart, which it can not see. Since
> your mail gave very good pointer where to look for with ref to another
> nor flash.

Ok, so the problem is really that the MTD layer does not know how to
talk to your spi flash.  I will not be of much help here, I'm afraid, so
I'll advise to find a mailing list closer to the problem...

> Seems I quoted text in previous mail which annoyed you. Sorry for
> being naive. I hope I will not repeat the same.

Well if one reads lots and lots of messages, it is basically important that
everybody uses an e-mail style allowing efficient processing.  There is
a good FAQ on this:

http://www.netmeister.org/news/learn2quote.html

Cheers
  Detlev

-- 
A statistician can have his head in an oven and his feet in ice, and
he will say that on the average he feels fine.
--
DENX Software Engineering GmbH,      MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich,  Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu at denx.de

[U-Boot] uboot kernel rejects filesystem

[Bharat Bhushan]

Hi Detlev,

Seems there was an issue with my kernel code that you correctly
pointed out in previous mail.

Now  I could pass bootarg board='myboard'. For myboard, I have done
relevent changes in kernel code, so kernel does init for my hardware.
Earlier while choosing mips machine, it was choosing generic mips
machine which did not mtdpartitions defined. Now with added support of
myboard, everything seems to be good.

Thanks for the promptly helps you have given.

Regards,
Bharat

[U-Boot] uboot kernel rejects filesystem

[Detlev Zundel]

Hi Bharat,

> Seems there was an issue with my kernel code that you correctly
> pointed out in previous mail.
>
> Now  I could pass bootarg board='myboard'. For myboard, I have done
> relevent changes in kernel code, so kernel does init for my hardware.
> Earlier while choosing mips machine, it was choosing generic mips
> machine which did not mtdpartitions defined. Now with added support of
> myboard, everything seems to be good.

Excellent - congratulations.

> Thanks for the promptly helps you have given.

One problem down, so many to go ;)

Cheers
  Detlev

-- 
The latest code  looks a bit similar to the old  [linux] big-reader-locks  hack
(which got dropped for good many eons ago and with which i deny any involvement
with, such as having authored it. [oh, did i say that out loud? crap.]), imple-
mented cleanly and properly.     -- Ingo Molnar <20090428124033.GA1655@elte.hu>
--
DENX Software Engineering GmbH,      MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich,  Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu at denx.de