Need help with compiling kernel v2.4.29

[mostlyharmless]

I'm attempting to compile 2.4.29 on a 486 which refuses to quit, and I have two relentless errors I need some help in eliminating...

1.
drivers/char/char.o: In function `siig10x_init_fn':
drivers/char/char.o(.text.init+0x12cd): undefined reference to `pci_siig10x_fn'
drivers/char/char.o: In function `siig20x_init_fn':
drivers/char/char.o(.text.init+0x12ed): undefined reference to `pci_siig20x_fn'

In my exploring, one apparent fix for this one is to enable CONFIG_PCI. This however a) makes no sense on a 486 with only ISA and VLB, and b) triggers a plethora of other errors when I try it.


2.
crypto/crypto.o: In function `crypto_init_digest_ops':
crypto/crypto.o(.text+0xeec): undefined reference to `crypto_alloc_hmac_block'
crypto/crypto.o: In function `crypto_exit_digest_ops':
crypto/crypto.o(.text+0xf06): undefined reference to `crypto_free_hmac_block'

As far as I can tell this is simply because the cryptography libraries are not distributed with the kernel source. There does not appear however to be a crypto API kernel patch available for 2.4.29.

All dependancies are at the required level or higher as specified in the 2.4.29 documentation.

With respect to applying the crypto API kernel patches, the most recent 2.4.x crypto patch which I can find is 2.4.3, and it will not apply to the 2.4.29 source cleanly.


Feel free to push me gently or kick me violently in the appropriate direction... I can take it. :

-d.

 

[marsd]

Have you checked the 2.4.29-rc(x) stuff?
The crypto stuff should go away, and you can make
it go away by not including it of course. I get the
feeling though that you need it.

The char device stuff is going to be another molehill.
If this builds successfully on a 586+ I would post this
and your hardware info to the kernel buglist and see who bites.
You might want to take a look at the source in question
and see if you can find any leads on the problem to save
time and brow sweat.

Good Luck and way to keep with it on a 486!

 

[mostlyharmless]

Have you checked the 2.4.29-rc(x) stuff?

Do you mean for the kernel source itself, or the crypto API? I haven't encountered it for either, but I may be looking in the wrong places.


The crypto stuff should go away, and you can make
it go away by not including it of course. I get the
feeling though that you need it.

Actually I don't need it, I just like it. Blame college for that one. Though if I can't get anywhere with it with respect to this exercise, I may simply do without for now.


The char device stuff is going to be another molehill.
If this builds successfully on a 586+ I would post this
and your hardware info to the kernel buglist and see who bites.

I hadn't thought of that... thanks!

 

[xmb]

drivers/char/char.o: In function `siig10x_init_fn':
drivers/char/char.o(.text.init+0x12cd): undefined reference to `pci_siig10x_fn'

Disabling this sii-image may help, CONFIG_BLK_DEV_SIIMAGE=n

2.6 is nice btw.

. Mac for productivity
.. Linux for developement
... Windows for solitaire