Skip to contentSkip to conversation

New: CONFIG_CMDLINE doesn't pass to kernel

Bugme-new Andrew Morton 6 participants 9 messages Oct 25, 2007 – Oct 28, 2007 1 message not in the archive

lore
Share & tools
lore
#1 lore raw
In reply to a message not in the archive
Quoted 17 lines · On Wed, 24 Oct 2007 05:32:14 -0700 (PDT) [email protected] wrote:

http://bugzilla.kernel.org/show_bug.cgi?id=9217

           Summary: CONFIG_CMDLINE doesn't pass to kernel
           Product: Other
           Version: 2.5
     KernelVersion: 2.6.23
          Platform: All
        OS/Version: Linux
              Tree: Mainline
            Status: NEW
          Severity: high
          Priority: P1
         Component: Other
        AssignedTo: [email protected]
        ReportedBy: [email protected]
Quoted 15 lines

Most recent kernel where this bug did not occur:2.6.24-rc1
Distribution: Debian Armel (arm with EABI)
Hardware Environment: Sharp Zaurus 750 (ARM pxa255)
Software Environment: Debian
Problem Description:
I try to upgrade kernel from 2.6.22.9 to 2.6.23.
2.6.22.9 works correct. Kernel 2.6.23 doesn't use CONFIG_CMDLINE and doesn't pass bootparams:
CONFIG_CMDLINE="console=ttyS0,115200n8 console=tty1 noinitrd root=/dev/mmcblk0p1 rootfstype=ext3 rw rootdelay=5"
But kernel still try to mount /dev/mtdblock2 using jffs2.
Steps to reproduce:
Compile kernel with CONFIG_CMDLINE and try to boot.

Something broke CONFIG_CMDLINE of ARM (at least) between 2.6.22 and 2.6.23.

I don't know whether it was an ARM patch one of those kernel-wide changes. We have futzed with the command-line parsing a bit recently, but the 2.6.23 changelog doesn't suggest anything obvious.

#2 lore raw
Quoted 5 lines · Andrew Morton

Something broke CONFIG_CMDLINE of ARM (at least) between 2.6.22 and 2.6.23.

I don't know whether it was an ARM patch one of those kernel-wide changes. We have futzed with the command-line parsing a bit recently, but the 2.6.23 changelog doesn't suggest anything obvious.

What does the affected system's bootloader pass in r2? If it's nonzero, ARM's 2.6.23 may interpret it as being an ATAGS pointer. And when that happens, the system prefers the ATAGS over CONFIG_CMDLINE.

There's sanity checking in __vet_atags, but maybe it isn't enough. Other than that, I can't see anything yet.

b.g.

Signature · 2 lines
Bill Gatliff
[email protected]
#9 lore raw
In reply to Andrew Morton
Quoted 35 lines · On Wed, 24 Oct 2007 05:32:14 -0700 (PDT) [email protected] wrote:
Quoted 18 lines

http://bugzilla.kernel.org/show_bug.cgi?id=9217

           Summary: CONFIG_CMDLINE doesn't pass to kernel            Product: Other
           Version: 2.5
     KernelVersion: 2.6.23
          Platform: All
        OS/Version: Linux
              Tree: Mainline
            Status: NEW
          Severity: high
          Priority: P1
         Component: Other
        AssignedTo: [email protected]         ReportedBy: [email protected]

Quoted 15 lines

Most recent kernel where this bug did not occur:2.6.24-rc1
Distribution: Debian Armel (arm with EABI)
Hardware Environment: Sharp Zaurus 750 (ARM pxa255)
Software Environment: Debian
Problem Description:
I try to upgrade kernel from 2.6.22.9 to 2.6.23.
2.6.22.9 works correct. Kernel 2.6.23 doesn't use  CONFIG_CMDLINE and doesn't pass bootparams:
CONFIG_CMDLINE="console=ttyS0,115200n8 console=tty1 noinitrd root=/dev/mmcblk0p1 rootfstype=ext3 rw rootdelay=5"
But kernel still try to mount /dev/mtdblock2 using jffs2.
Steps to reproduce:
Compile kernel with CONFIG_CMDLINE and try to boot.

Quoted 4 lines

Something broke CONFIG_CMDLINE of ARM (at least) between 2.6.22 and 2.6.23.

I don't know whether it was an ARM patch one of those kernel-wide changes. We have futzed with the command-line parsing a bit recently, but the 2.6.23 changelog doesn't suggest anything obvious.

I think the zaurus kernel deliberately ignored the bootparams passed in by the device's bootloader. The 2.6.23 kernel now honors the r2 setting from the bootloader, which is most likely pointing to those ATAGs that were being ignored earlier.

I suppose you can hack up the logic in setup.c where the ATAGs are parsed, but it seems to me that it would be better to work out some logic, perhaps in the fixup section for that machine, to do whatever is necessary to override or otherwise get along with the bootloader.

Mike (mwester)

#3 lore raw
In reply to Bill Gatliff
Quoted 6 lines · Bill Gatliff
Quoted 5 lines

Something broke CONFIG_CMDLINE of ARM (at least) between 2.6.22 and 2.6.23.

I don't know whether it was an ARM patch one of those kernel-wide changes. We have futzed with the command-line parsing a bit recently, but the 2.6.23 changelog doesn't suggest anything obvious.

Quoted 6 lines

What does the affected system's bootloader pass in r2? If it's nonzero, ARM's 2.6.23 may interpret it as being an ATAGS pointer. And when that happens, the system prefers the ATAGS over CONFIG_CMDLINE.

There's sanity checking in __vet_atags, but maybe it isn't enough. Other than that, I can't see anything yet.

If this is the SHARP Shepherd (which seems to be the Zaurus SL-C750, which looks like it's the one the submitter is using), it's not setting boot_params at all:

MACHINE_START(SHEPHERD, "SHARP Shepherd")
.phys_io = 0x40000000,
.io_pg_offst = (io_p2v(0x40000000) >> 18) & 0xfffc,
.fixup = fixup_corgi,
.map_io = pxa_map_io,
.init_irq = pxa25x_init_irq,
.init_machine = corgi_init,
.timer = &pxa_timer,
MACHINE_END

#4 lore raw
Quoted 19 lines · Lennert Buytenhek
Quoted 6 lines · Bill Gatliff
Quoted 5 lines

Something broke CONFIG_CMDLINE of ARM (at least) between 2.6.22 and 2.6.23.

I don't know whether it was an ARM patch one of those kernel-wide changes. We have futzed with the command-line parsing a bit recently, but the 2.6.23 changelog doesn't suggest anything obvious.

Quoted 6 lines

What does the affected system's bootloader pass in r2? If it's nonzero, ARM's 2.6.23 may interpret it as being an ATAGS pointer. And when that happens, the system prefers the ATAGS over CONFIG_CMDLINE.

There's sanity checking in __vet_atags, but maybe it isn't enough. Other than that, I can't see anything yet.

If this is the SHARP Shepherd (which seems to be the Zaurus SL-C750, which looks like it's the one the submitter is using), it's not setting boot_params at all:

Quoted 9 lines

MACHINE_START(SHEPHERD, "SHARP Shepherd")
.phys_io = 0x40000000,
.io_pg_offst = (io_p2v(0x40000000) >> 18) & 0xfffc,
.fixup = fixup_corgi,
.map_io = pxa_map_io,
.init_irq = pxa25x_init_irq,
.init_machine = corgi_init,
.timer = &pxa_timer,
MACHINE_END

The Zaurus has an insane bootloader which doesn't pass any sane values, it doesn't support the right (well, any) machine number and the commandline it passes is unchangeable and bogus too. The approach has always therefore been to pretend it doesn't exist and used a commandline compiled into the kernel.

Could the original bug reporter please report what commandline the kernel actually uses please? In theory it can only be either:

* the one being compiled in with CONFIG_CMDLINE
* the one hardcoded into the bootloader

but I'd be surprised it boots at all with the latter!

Cheers,

Richard

#5 lore raw
Quoted 36 lines · On Thu, 25 Oct 2007 14:43:36 +0100 Richard Purdie <[email protected]> wrote:
Quoted 19 lines · Lennert Buytenhek
Quoted 6 lines · Bill Gatliff
Quoted 5 lines

Something broke CONFIG_CMDLINE of ARM (at least) between 2.6.22 and 2.6.23.

I don't know whether it was an ARM patch one of those kernel-wide changes. We have futzed with the command-line parsing a bit recently, but the 2.6.23 changelog doesn't suggest anything obvious.

Quoted 6 lines

What does the affected system's bootloader pass in r2? If it's nonzero, ARM's 2.6.23 may interpret it as being an ATAGS pointer. And when that happens, the system prefers the ATAGS over CONFIG_CMDLINE.

There's sanity checking in __vet_atags, but maybe it isn't enough. Other than that, I can't see anything yet.

If this is the SHARP Shepherd (which seems to be the Zaurus SL-C750, which looks like it's the one the submitter is using), it's not setting boot_params at all:

Quoted 9 lines

MACHINE_START(SHEPHERD, "SHARP Shepherd")
.phys_io = 0x40000000,
.io_pg_offst = (io_p2v(0x40000000) >> 18) & 0xfffc,
.fixup = fixup_corgi,
.map_io = pxa_map_io,
.init_irq = pxa25x_init_irq,
.init_machine = corgi_init,
.timer = &pxa_timer,
MACHINE_END

The Zaurus has an insane bootloader which doesn't pass any sane values, it doesn't support the right (well, any) machine number and the commandline it passes is unchangeable and bogus too. The approach has always therefore been to pretend it doesn't exist and used a commandline compiled into the kernel.

Quoted 8 lines

Could the original bug reporter please report what commandline the kernel actually uses please? In theory it can only be either:

* the one being compiled in with CONFIG_CMDLINE
* the one hardcoded into the bootloader

but I'd be surprised it boots at all with the latter!

It was in the inital report, at
http://bugzilla.kernel.org/show_bug.cgi?id=9217 :

: I try to upgrade kernel from 2.6.22.9 to 2.6.23.
: 2.6.22.9 works correct. Kernel 2.6.23 doesn't use CONFIG_CMDLINE and doesn't : pass bootparams:
: CONFIG_CMDLINE="console=ttyS0,115200n8 console=tty1 noinitrd
: root=/dev/mmcblk0p1 rootfstype=ext3 rw rootdelay=5"
: But kernel still try to mount /dev/mtdblock2 using jffs2.
: Steps to reproduce:
: Compile kernel with CONFIG_CMDLINE and try to boot.

#6 lore raw
Quoted 5 lines · Andrew Morton
Quoted 3 lines · On Thu, 25 Oct 2007 14:43:36 +0100 Richard Purdie <[email protected]> wrote:

Could the original bug reporter please report what commandline the kernel actually uses please? In theory it can only be either:

Quoted 7 lines
Quoted 4 lines

* the one being compiled in with CONFIG_CMDLINE
* the one hardcoded into the bootloader

but I'd be surprised it boots at all with the latter!

It was in the inital report, at
http://bugzilla.kernel.org/show_bug.cgi?id=9217 :

No it wasn't, I'd looked.

Quoted 8 lines

: I try to upgrade kernel from 2.6.22.9 to 2.6.23.
: 2.6.22.9 works correct. Kernel 2.6.23 doesn't use CONFIG_CMDLINE and doesn't : pass bootparams:
: CONFIG_CMDLINE="console=ttyS0,115200n8 console=tty1 noinitrd
: root=/dev/mmcblk0p1 rootfstype=ext3 rw rootdelay=5"
: But kernel still try to mount /dev/mtdblock2 using jffs2.
: Steps to reproduce:
: Compile kernel with CONFIG_CMDLINE and try to boot.

This is the commandline they wanted to use, not the one that was actually used. The one that was used should appear on the console, on the serial console if they have the lead or in dmesg when the device boots. We know it probably contained root=/dev/mtdblock2 but thats it.

I'd like to confirm which commandline is appearing since if we know which one it is we might stand a chance of knowing where it came from, until then this is just a guessing game.

Cheers,

Richard

#7 lore raw
Quoted 4 lines · Richard Purdie
Quoted 2 lines · Andrew Morton

It was in the inital report, at
http://bugzilla.kernel.org/show_bug.cgi?id=9217 :

Quoted 8 lines

This is the commandline they wanted to use, not the one that was actually used. The one that was used should appear on the console, on the serial console if they have the lead or in dmesg when the device boots. We know it probably contained root=/dev/mtdblock2 but thats it.

I'd like to confirm which commandline is appearing since if we know which one it is we might stand a chance of knowing where it came from, until then this is just a guessing game.

There was a followup in the bugzilla:

Quoted 7 lines

I use u-boot to boot kernel.
When kernel >=2.6.23 boot it use u-boot default bootparam:
console=ttyS0,115200 console=tty1 root=/dev/mtdblock2 rootfstype=jffs2 Now I try to change it in u-boot source.

kernel 2.6.22.9 change default u-boot boot params to what I set in CONFIG_CMDLINE But 2.6.23 and later doesn't

So its not the standard bootloader, its one which passes an ATAG commandline and 2.6.23 onwards uses an ATAG commandline if present over the compiled in commandline.

My opinion is that this is a feature. Recompile the bootloader without the commandline if you want to use the one compiled into the kernel.

This is going to be useful in the future when booting Zaurus kernels with kexec since it means we can customise the commandline through the ATAGs.

Cheers,

Richard

#8 lore raw
Quoted 21 lines · Richard Purdie
Quoted 4 lines · Richard Purdie
Quoted 2 lines · Andrew Morton

It was in the inital report, at
http://bugzilla.kernel.org/show_bug.cgi?id=9217 :

Quoted 8 lines

This is the commandline they wanted to use, not the one that was actually used. The one that was used should appear on the console, on the serial console if they have the lead or in dmesg when the device boots. We know it probably contained root=/dev/mtdblock2 but thats it.

I'd like to confirm which commandline is appearing since if we know which one it is we might stand a chance of knowing where it came from, until then this is just a guessing game.

There was a followup in the bugzilla:

Quoted 5 lines

I use u-boot to boot kernel.
When kernel >=2.6.23 boot it use u-boot default bootparam:
console=ttyS0,115200 console=tty1 root=/dev/mtdblock2 rootfstype=jffs2 Now I try to change it in u-boot source.

Quoted 6 lines
Quoted 2 lines

kernel 2.6.22.9 change default u-boot boot params to what I set in CONFIG_CMDLINE But 2.6.23 and later doesn't

So its not the standard bootloader, its one which passes an ATAG commandline and 2.6.23 onwards uses an ATAG commandline if present over the compiled in commandline.

That statement is misleading. The kernel has _always_ used the command line passed from the boot loader in preference to the built-in command line.

The thing that's changed is that the kernel how honors the value passed from the boot loader in R2 to find out where the ATAG list is - as documented since 2002 that it will eventually do. That's 5 years of warning.

Signature · 3 lines
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of: