linux-kernel - Re: [PATCHv3 1/3] x86: use ELF format in compressed images.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sun, 06 Apr 2008 10:25:28 -0700
From:	"H. Peter Anvin" <hpa@...or.com>
To:	Yinghai Lu <yhlu.kernel@...il.com>
CC:	Ian Campbell <ijc@...lion.org.uk>, Ingo Molnar <mingo@...e.hu>,
	linux-kernel@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...hat.com>,
	Jeremy Fitzhardinge <jeremy@...p.org>,
	virtualization@...ts.linux-foundation.org
Subject: Re: [PATCHv3 1/3] x86: use ELF format in compressed images.

Yinghai Lu wrote:
> On Sun, Apr 6, 2008 at 9:38 AM, H. Peter Anvin <hpa@...or.com> wrote:
>> Yinghai Lu wrote:
>>
>>> so will cost every bzImage extra memory copy? that could be 18M or
>>> even more big.
>>>
>>  I wouldn't worry about that.  You will typically have several copies of the
>> images during the execution of the boot loader.
> 
> i put all drivers needed in kernel.
> 1. bootloader copy bzImage (6M) to memory
> 2. arch/x86/boot/compressed/head_32.S, will copy bzImage to end of
> buffer to do uncompress on possiton.
> 3. parse_elf will copy the vmlinux (the uncompressed, that is some big, 18M)
> 
> I suggest that could have special elf header, and will only have one
> PT_LOAD, and avoid the copy, and just offset start address of
> uncompressed kernel for jump later.

Once again, I think you will have a hard time measuring the time difference.

The start address of the uncompressed kernel is defined at compile time. 
  Typically it is set based on alignment constraints in the hardware; 
for x86-32 is is normally 1 MB but for reasonably large hardware 16 MB 
is better (in fact, I have proposed making 16 MB the default.)  Recovery 
kernels are frequently compiled with completely different addresses.

With your proposal, this would no longer be possible to be a 
configurable option, which would be a major loss -- a major loss of 
functionality for a small fraction of time at bootup time (which is 
almost certainly dwarfed by all the other copying and initialization 
that happens at boot time.)

	-hpa
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/