Re: debugging a pager

[Thomas Bushnell, BSG]

Diego Roversi <diegor@tiscalinet.it> writes:

>   I need some help in debugging a pager. I'm trying to implement a cache in
> libstore and I modified the dev_read e dev_write function, but they are used
> to serve the page foults. The problem is that the translator died with a
> SIGSEGV the first time it tries to access to the file system (to the
> superblock). I can attach a gdb to a running translator, but I can't set
> breakpoints. So I started my translator using the information find in
> "Manually bootstrapping a Translator", but I found that I can set
> breakpoints everyware but they are just ignored if used by a pager :-(.

Well, I'm not sure just what you're doing.  Breakpoints in pagers do
work just fine, but something else may be going on.

> I've also thinked to refactor the ext2fs translator, so it doesn't
> explicitaly rely on memory mapping. The idea is to replace explicit access
> to mapped memory with something like a pseudo call to bread e bwrite.

This would be a step backwards.  The whole reason that memory mapping
is used is so that virtual memory is managed as a single large pool,
by a kernel that knows global usage patterns.  This is the Right
Thing.  bread and bwrite are awful.