On Mon, 19 Jan 2004, Muli Ben-Yehuda wrote:

> On Mon, Jan 19, 2004 at 02:37:28PM +0200, guy keren wrote:
>
> > > - the capitalization of Kernel, Driver and Hardware is a bit
> > > jarring. Same thing for all other slides.
> >
> > i want the terms discussed in the lecture to Stand Out, in the
> > least-intrusive possible manner.
>
> Then make them bold, or italic, or even _like this_. Capitalizing
> inappropriate Words Makes it Look silly.
>
> > if i get more ocmplaints about this, i'll
> > un-capitalize them next time.
>
> You can count my complaint as two, if you wish ;-)

ok, ok. next time.

> > > slide 2
> > >
> > > - the kernel can be compiled entirely monolithic (no modules)
> >
> > ok, but this is not what people normally do (and there are features which,
> > as far as i know, can only be compiled as modules).
>
> In the core kernel? few if any. I would consider any such a feature a
> bug. Pointers?

in redhat 7.3's 2.4.18-17.7.x kernel, 'cryptography support'/'crypto
devices' - can't make it allow me to choose 'y'. either 'm' (module) or
'n'(no). or in USb, 'usb audio support'. maybe there is a way to tell it
to let me compile it as a module - i don't see how.

> > i don't want to get
> > into that - this is not a lecture for people in the embedded systems area,
> > after all :)
>
> Compiling a monolithic kernel is the easiest thing to do when you
> compile on one machine and install on another - but it's your talk and
> your call.

and it's beyond the scope.

> > > - explain why we have major and minors?
> >
> > hmmm... good idea. i did write what they stand for - just not why they
> > exist. however, this explanation matters more for kernel programmers, and
> > for users, so i'm not sure if it fits into this lecture. won't it be
> > better to leave this for your 'device drivers' lecture? :)
>
> Possibly. Your call.

or yours :0

> > > - mention that for the user, the distrinction between character and
> > > block devices is irrelevant? it's only the driver that cares
> > > (different kernel APIs)
> >
> > ok. thought this is not entirely correct. it reflects on the nature of
> > data tranasfer to/from the device.
>
> And the user cares because?
>
> > after all, passing data in blocks is
> > normally faster then passing it in serial (ofcourse, passing in blocks ot
> > a floppy is much much slower then passing serially to a firewire
> > device...). hmmm.. ok, i'll mention this, after all.
>
> For the user, /dev/kmem and /dev/sda are accessed exactly the same
> (cat, dd, tail, echo, whatever). It's only for the programmer that the
> distinction matters.

as i said, just expected throughput...

> > > slide 11
> > >
> > > - 2.6 disables module unloading (by default?)
> >
> > i'm focusing on eisting kernels and existing distributions.
>
> Some distros already have 2.6 in their testing branches. Fedora
> "testing" already comes with a 2.6 kernel, AFAICR.

*shrug* we,, let them be doomed ;)

-- 
guy

"For world domination - press 1,
 or dial 0, and please hold, for the creator." -- nob o. dy

--------------------------------------------------------------------------
Haifa Linux Club Mailing List (http://www.haifux.org)
To unsub send an empty message to [EMAIL PROTECTED]


Reply via email to