Skip to contentSkip to conversation

2.4.14 loopback blk dev compilation trouble

DISCUSSION Ishan Oshadi Jayawardena 3 participants 3 messages Nov 21, 2001

lore
Share & tools
lore
#0 lore raw

Greetings.
I've seen that the compilation of off-the-server
2.4.14 tree fails at the end of 'make bzImage' because drivers/block/loop.c uses the deactivate_page() function, which seems to have been removed from the source tree.
By following the progress of the kernel through
2.4.12, 2.4.13, and 2.4.14 patches, I've seen that page_cache_release() does the same things as deactivate_page(). Both these functions are used in the together twice in drivers/block/loop.c. I compiled the 2.4.14 kernel by commenting out the references to deactivate_page() but leaving page_cache_release(), and loopback block devs work; but I do not have the resources to check it for memory leaks etc.
I _think_ I've done the right thing, but
would appreciate verification by a regular kernel- hacker :-)
I searched the net for some reference to this problem, but couldn't find anything so far.

(I've compiled loop blk-dev support in to the kernel.)

[email protected]
[email protected]

#1 lore raw
Quoted 4 lines · Ishan Oshadi Jayawardena

I've seen that the compilation of off-the-server 2.4.14 tree fails at the end of 'make bzImage' because drivers/block/loop.c uses the deactivate_page() function, which seems to have been removed from the source tree.

Remove the lines containing deactivate_page() and compilation should go smoothly. This problem was fixed in 2.4.15-pre1.

This was discussed on this list _many_ times. Wonder how you missed ;)

Cheers,

Anuradha

Signature · 5 lines
Debian GNU/Linux (kernel 2.4.13)

To fear love is to fear life, and those who fear life are already three
parts dead.
		-- Bertrand Russell
#2 lore raw
In reply to Ishan Oshadi Jayawardena
Quoted 20 lines · Ishan Oshadi Jayawardena

Greetings.
I've seen that the compilation of off-the-server
2.4.14 tree fails at the end of 'make bzImage' because drivers/block/loop.c uses the deactivate_page() function, which seems to have been removed from the source tree.
By following the progress of the kernel through
2.4.12, 2.4.13, and 2.4.14 patches, I've seen that page_cache_release() does the same things as deactivate_page(). Both these functions are used in the together twice in drivers/block/loop.c. I compiled the 2.4.14 kernel by commenting out the references to deactivate_page() but leaving page_cache_release(), and loopback block devs work; but I do not have the resources to check it for memory leaks etc.
I _think_ I've done the right thing, but
would appreciate verification by a regular kernel- hacker :-)
I searched the net for some reference to this problem, but couldn't find anything so far.

Quoted 1 line

(I've compiled loop blk-dev support in to the kernel.)

That's the correct way to correct the compilation problem. It has been reported 22987432 times already...

DK