[bug] 2.6.0 COMMAND_LINE_SIZE <160???

[bug] 2.6.0 COMMAND_LINE_SIZE <160???

[Jeff Chua]

I can't seems to pass more than 160 bytes on command line when booting
linux-2.6.0. 2.4.24 is ok.

Booting via loadlin, but I've tried linld, still the same problem.

It hangs after "Ok, booting the kernel".

Thanks,
Jeff
[ jchua@fedex.com ]

Re: [bug] 2.6.0 COMMAND_LINE_SIZE <160???

[Randy.Dunlap]

On Tue, 23 Dec 2003 21:07:45 +0800 (SGT) Jeff Chua <jeffchua@silk.corp.fedex.com> wrote:

| 
| I can't seems to pass more than 160 bytes on command line when booting
| linux-2.6.0. 2.4.24 is ok.
| 
| Booting via loadlin, but I've tried linld, still the same problem.
| 
| It hangs after "Ok, booting the kernel".

So 2.4.24-pre works OK for you.

2.6.0 works OK for me with a command line length of 219 bytes
(about half of it doesn't matter for booting, but it's all
there in /proc/cmdline).  Using lilo.

Same processor arch. type in both .config files?
Same compiler version building them?

--
~Randy
MOTD:  Always include version info.

Re: [bug] 2.6.0 COMMAND_LINE_SIZE <160???

[Jeff Chua]

On Tue, 23 Dec 2003, Randy.Dunlap wrote:

> On Tue, 23 Dec 2003 21:07:45 +0800 (SGT) Jeff Chua <jeffchua@silk.corp.fedex.com> wrote:
>
> Same processor arch. type in both .config files?
> Same compiler version building them?

Same for both.

gcc 2.95.3 20010315
glibc-2.2.5-34

perhaps, it's loadlin or linld problem, or fat32 problem. The difference
is lilo doesn't need _dos_ to boot.

I tried to replace ./arch/i386/boot/setup.S with the one from 2.4.24-pre1
and it seems to go further before it breaks.

What's next to debug this?

Thanks,
Jeff

Re: [bug] 2.6.0 COMMAND_LINE_SIZE <160???

[Randy.Dunlap]

On Wed, 24 Dec 2003 18:11:54 +0800 (SGT) Jeff Chua <jeffchua@silk.corp.fedex.com> wrote:

| 
| On Tue, 23 Dec 2003, Randy.Dunlap wrote:
| 
| > On Tue, 23 Dec 2003 21:07:45 +0800 (SGT) Jeff Chua <jeffchua@silk.corp.fedex.com> wrote:
| >
| > Same processor arch. type in both .config files?
| > Same compiler version building them?
| 
| Same for both.
| 
| gcc 2.95.3 20010315
| glibc-2.2.5-34
| 
| perhaps, it's loadlin or linld problem, or fat32 problem. The difference
| is lilo doesn't need _dos_ to boot.

Just curious, what kind of disk/OS config do you have where
using loadlin (or linld) is useful?  Does it have anything to do
with not writing (updating) the MBR on disk?
Are there some booting alternatives for you?

Is Linux on the same hard disk as Windows (or DOS)?

| I tried to replace ./arch/i386/boot/setup.S with the one from 2.4.24-pre1
| and it seems to go further before it breaks.

Yes, not much difference in those source files.

| What's next to debug this?

Good question.  I'm not sure.  Anyone else have suggestions?

--
~Randy