gnewsense-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [gNewSense-users] KFV: check for non-free firmware


From: Paul O'Malley
Subject: Re: [gNewSense-users] KFV: check for non-free firmware
Date: Sun, 04 May 2008 10:31:13 +0100
User-agent: Thunderbird 2.0.0.12 (X11/20080227)

Christophe Jarry wrote:
I think I've found four non-free firmware in the kernel (still haven't browse 'drivers' section). Here they are:

'arch/powerpc/oprofile/op_model_cell.c' contains (line 703, I've removed some 
digits to ease the reading):

#define ENTRIES  303
#define MAXLFSR  0xFFFFFF

/* precomputed table of 24 bit LFSR values */
static int initial_lfsr[] = {
 8221349, [lots of digits], 8388607
};


'arch/powerpc/platforms/cell/spufs/spu_restore_dump.h_shipped' contains (line 
1, I've removed some digits to ease the reading
):

/*
 * spu_restore_dump.h: Copyright (C) 2005 IBM.
 * Hex-dump auto generated from spu_restore.c.
 * Do not edit!
 */
static unsigned int spu_restore_code[]  __attribute__((__aligned__(128))) = {
0x40800000,
[lots of digits],
0x00000000,
};

'arch/powerpc/platforms/cell/spufs/spu_save_dump.h_shipped' (almost the same as 
'arch/powerpc/platforms/cell/spufs/spu_restore_dump.h_shipped'
).

^^^^^^^^^^^^^^^^^^^^

Hi Christophe,

please see that - it is PPC, we dont do PPC at this stage if helps more if you stick to what we produce that is the X86 platform.

What we need is to do is clean drivers first then we can hit other things.

Each of those items would call for an examination of all the code around them, which is fine, but we only have one programmer that is not fair on him as we are not even compiling for release such a kernel.

Please stay with drivers on x86, the team starts there and then when someone builds multi arch then we can chase their drivers out out.

When this is done we can consider other parts of the kernel, but as a small group breaking up very big tasks into small bite (byte) size pieces and working with the team on those parts means we can measure each part of what is done, and actually not be randomly going in against millions of lines of code with no plan.

I hope you can see from the above that this is best done in stages, and the best way to cover ground is to cover it with others.


Regards,

Paul




reply via email to

[Prev in Thread] Current Thread [Next in Thread]