Bootloader question

Martin Langhoff martin.langhoff at
Thu Jun 4 04:24:47 EDT 2009

On Thu, Jun 4, 2009 at 5:55 AM, Mitch Bradley <wmb at> wrote:
>> 18 seconds to initrd load
> dominated by decompression time.   Eliminate the initrd ...

Or pare it down significantly. I suspect that

 - some bits of python libs are not used
 - a newer python that does -O the right way can help
 - replacing the initrd with a conventional non-bitfrosty python can
be an alternative

> Parallelization will not help at all if you are using JFFS2.  The low
> level NAND driver that JFFS2 uses busy waits for I/O, and then JFFS2 is
> CPU-bound on the decompression step, preventing any useful concurrency.
> The busy-wait could be changed to an interrupt - if only someone had
> time to do the work and test it extensively.

That is painfully true. Search the archives for Daniel Drake's example
patch to try to fix the busy wait and the associated discussion. There
were earlier discussions with David Woodhouse on the topic.

Very valuable info there for someone with kernel hacking chops to tackle.

> Easily?  Yeah, right.

Well, he can show us how it's done. I'll definitely be impressed :-)

 martin.langhoff at
 martin at -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first

More information about the Devel mailing list