Re: [PATCH][TAKE 3] THE LINUX/I386 BOOT PROTOCOL - Breaking the 256 limit

["H. Peter Anvin" <hpa@zytor.com>]

Alon Bar-Lev wrote:
> 
> Hello,
> 
> The problem is that boot loader developers did not understand the old 
> statement: "A string that is too long will be automatically truncated by 
> the kernel, a boot loader may allow a longer command line to be passed 
> to permit future kernels to extend this limit."
> 
> Most of them handed the same buffer to < 2.02 protocols and >= 2.0.2 
> protocols. When I've opened bugs against that they claimed that they 
> follow instructions since the 256 limit was explicitly mentioned. I've 
> ended up in patching GRUB my-self to allow this.
> 
> I thought that this should be made clearer... But maybe I did not write 
> it too well.
> 
> I've removed the 255+1 limitation from the boot protocol main 
> description, so there will be no known limit there... And moved it to 
> the <2.02 section notes.
> 
> Can you please suggest a different phrasing? Or maybe you think that it 
> is not needed at all... But then I have a problem of making boot loader 
> fix their code.
> 

Well, obviously, since apparently LILO doesn't properly null-terminate 
long command line.

Thinking about it a bit, the way to deal with the LILO problem is 
probably to actually *usw* the boot loader ID byte we've had in there 
since the 2.00 protocol.  In other words, if the boot loader ID is 0x1X 
where X <= current version (I don't know how LILO manages this ID) then 
truncate the command line to 255 bytes; when this is fixed in LILO then 
LILO gets to bump its boot loader ID version number.

	-hpa