Miscellaneous compile problems in 2.1.31 - repost and additions

Tom Zerucha (tz@execpc.com)
Sat, 5 Apr 1997 14:40:36 -0500

On Thu, 3 Apr 1997, Tom Zerucha wrote:

> I have a number of patches applied, so these may not apply to the actual
> distribution, but I think they do, and someone should take a look
> drivers/block: xd.c and xd.h disagree about xd_release, it is declared as
> returning void in xd.h and should be returning int.

enable xd (this case was as module) to find this.

> net/ipv4/ipmr.c Look inside for an impr (which should be ipmr). Either

The m and p are reversed. Enable multicast routing and compile

> my patch program or the patch suffers from dyslexia.
> drivers/scsi/g_NCR5380.c around line 399, there is a for( i=0..., but i is
> not declared. This only happens for port mode.

Enable port instead of memory, and ask for the driver (maybe as module) to
see this.

> Even with the new SMP stuff, in arch/i386/kernel/process.c, APM and SMP
> are mutually exclusive - it works, but the APM doesn't call the necessary
> routines, so things like power events aren't handled right.
> tz@execpc.com
> finger tz@execpc.com for PGP key

Enabling ltpc and shaper as modules has the following problem.

In drivers/net:

ltpc.c, at line 1261, it is missing the semicolon at the end

shaper.c, at line 617, (dev) should be (&dev_shaper)

ALSO, I think the default for the CONFIG_VT and CONFIG_VT_CONSOLE should
be 'Y' but aren't, so no console comes up if you simply let everything

finger tz@execpc.com for PGP key