Re: [PATCH 5/5] add setting gs/fsbase

[Sergey Bugaev]

Hi,

On Mon, Apr 24, 2023 at 11:02 PM Luca Dariz <luca@orpolo.org> wrote:
> Il 24/04/23 17:19, Sergey Bugaev ha scritto:
> > Resending without the attachment, because apparently the email did not
> > make it into the list archive so maybe it didn't get to you either;
> > and I'm too conscious about email just eating my letters without any
> > notice or indication since that one time. If you didn't get the last
> > one and still want the attachment, let me know and we'll work
> > something out. And please ack receiving this one in any case.
>
> I received the previous one, including the attachment, thanks!

Good -- and it did eventually make its way into the archive too.
Interestingly, the second one got into the archive first, and it was
being displayed as a reply to an unknown message for a while, and then
the first one showed up. I don't know what's going on, perhaps
*something* was taking its time scanning my 5.6 MB binary attachment
for malware?

> I can see issues also with simpler tests, if I use the graphical
> console; actually the whole bootstrap task doesn't start.

Yes, with the graphical console (called VGA text console I believe?)
the bootstrap crashes before it even first enters user mode. So please
debug that one too.

> Did you try to
> use the serial console? It has the drawback of reconfiguring the
> terminal every time, but it seems to work for both in/out.

...but yes, since you mentioned using -nographic & console=com0 last
time, this is what I've been using.

> I tried using
> the "com" mach device with the device_read/write() rpcs.

I'm trying the "console" device, and glibc uses the _inband versions
of the RPCs, I believe. Please try my bc build and see if you can
reproduce it.

Alternatively, if it is some fatal crash in userland, what should I
break on in gnumach to detect it? Somewhere in vm_fault.c? I can't
just break on all page faults since there are a lot of them and most
are benign. Maybe on i386_exception ()? Hm, maybe I should just -D
MACH_KDB.

Sergey