svn commit: r304187 - in head: . share/man/man4 sys/conf sys/dev/mcd sys/modules sys/modules/mcd

[Bruce Evans]

On Fri, 19 Aug 2016, Alexey Dokuchaev wrote:

> On Fri, Aug 19, 2016 at 09:12:53PM +1000, Bruce Evans wrote:
>> I normally use my version of biosboot for boot2.  I improved its caching
>> just a couple of years ago.  It was using 9K buffer optimized for 1440K
>> floppies.  Now it uses a 32K buffer.  Booting a 5.5MB kernel takes a
>> fraction of a second.
>
> That looks like change that everyone would benefit from.  Consider posting
> a patch or committing it yourself. ;-)

Oops, it is actually loading that takes a fraction of a second.  Not
much different than with -current boot2 or old loader.  -current boot2
also works for me, but I managed to squeeze more of the features that
I want into biosboot (everything except ufs2).  Booting from the boot1
prompt takes 20-25 seconds here.

>> Of course I don't use modules, so not many seeks are needed.
>
> Why not, they're convenient (apart from "kernel version mismatch" crap
> that I keep hitting now and then)?  AFAIR they were slow to load at some
> point but that was (not so) recently fixed.

Same reason that I don't use shared libraries if possible - they are
larger, slower and more difficult to debug.  For kernel development,
the version control problem is large.  Modules are good for avoiding
rebooting when developing something in a single module, but I usually
work on either small changes that panic often or system wide-changes
that need recompiling everything.  Both require rebooting a lot to
test, and I can rebuild a kernel and without modules and reboot it
almost faster than I can remember where the modules directories are.

Bruce