Re: [PATCH 5/5] added an image to the hurd/what_is_the_gnu_hurd page.

[Samuel Thibault]

Hello,

jbranso@dismail.de, le jeu. 30 mai 2024 13:29:49 -0400, a ecrit:
> * hurd/what_is_the_gnu_hurd.mdwn: I added the image from the
> open_issues/anatomy_of_a_hurd_system.mdwn.  This page is probably
> something that is viewed a lot.  People like images.
> ---
>  hurd/what_is_the_gnu_hurd.mdwn | 9 ++++++++-
>  1 file changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/hurd/what_is_the_gnu_hurd.mdwn b/hurd/what_is_the_gnu_hurd.mdwn
> index 8315bfff..09c26ee7 100644
> --- a/hurd/what_is_the_gnu_hurd.mdwn
> +++ b/hurd/what_is_the_gnu_hurd.mdwn
> @@ -33,7 +33,14 @@ provides the same standard interfaces known from other 
> [[UNIX]]-like systems.
>  Thus, for a typical user, the Hurd is intended to silently work in the
>  background providing the services and infrastructure which the 
> [[microkernel]]
>  itself has no business implementing, but that are required for higher level
> -programs and libraries to operate.
> +programs and libraries to operate. Let's look at an example.
> +
> +[[!img open_issues/images/overview.svg]]
> +
> +Firefox invokes glibc's `send ()`, which in turn uses the pfinet (or
> +lwip) TCP/IP stack, which talk to our device drivers (rump or netdde),
> +which finally talk to GNU Mach.  Only GNU Mach runs in kernel space!
> +Everything else is userspace!

Actually netdde doesn't even talk to GNU Mach. It directly uses i/o with
the hardware. The interrupt notification does go through GNU Mach, but
that could also use Intel's userland interrupt support to avoid that
part. So I'd say direct the arrow directly to the network board rather
than to GNU Mach.

Samuel