Re: Ruby redux

[Zoltan B. <zb...@fr...>]

Hi!

Aivils Stoss =EDrta:
> Hello Zoltan!
>=20
> Job completed. Now we have linux-ruby patch against -mm1
> http://www.ltn.lv/~aivils/files/test/ruby-2.6.11-rc4-mm1-2.diff.bz2
>=20
> Changes:
> - dummy console command line changed dumbcon=3D3:5 - where 3 count of d=
ummy
> consoles and 5 VC count per each dummy console.
> - fbcon does not manage VT. Instead be in use dummy console. On fbcon
> start, it takes over VGA console. Second fbdev consoles must add manual=
y
> with con2fb tool. fbcon setup parameter vc: is ignored. If dummy consol=
e
> have vc:17-19 ,then sysop must run con2fb three times:
> con2fb /dev/fb1 17;con2fb /dev/fb1 18;con2fb /dev/fb1 19;
> Unfortunately this feature is untested.
> - fbcon mode when single user uses multiple heads is not supported. In
> abstract it should work, but code do not suport one keyboard for two or
> more VT_STRUCT.
>=20
> Bugs:
> - at console start pop up 1 warning per each console.
> - more unknown.
>=20
> - PS/2 devices will not work under 2.6.11-rc4-mm1
>=20
> Aivils Stoss

Now I was able to compile and try it.

Your patch backs out the drivers/char/sysrq.c changes that is in
2.6.11-rc4-mm1. Main problem was that you haven't changed neither
the prototype of __handle_sysrq() in include/linux/sysrq.h nor
it's callers. No wonder it didn't compile. You don't select SysRQ
support in your kernels? Anyway, here's the patch that resets
sysrq.c back to the state as it was in 2.6.11-rc4-mm1. You have
to apply it over your ruby-2.6.11-rc4-mm1-2.diff.bz2 patch, your
changes from sysrq_handle_SAK() and sysrq_handle_unraw() don't
get lost. THEN it compiles.

And runs! Congratulations! :-)

Although I worried about a implicit declaration warning from
vt_ioctl.c, you commented out the correct prototype of update_attr()
in vt_kern.h.

Anyway, the problem with PS/2 keyboards has an effect on my machine,
too. Only one of my two PS/2 keyboards work, the one on the keyboard
port. The other one, plugged into the mouse port didn't.

Well, until I rebooted with "i8042.nopnp". That brought back alive both
my keyboards. :-) I guess this brings back your PS/2 keyboard, too.

I found only one problem in dmesg:

-------------------------------------------------------------
Badness in vc_resize at drivers/char/vt.c:683

Call Trace:<ffffffff8022740f>{vc_resize+79}=20
<ffffffff801a86f2>{create_proc_entry +146}
        <ffffffff80226000>{read_kbd_phys+0}=20
<ffffffff80227221>{visual_init+65}
        <ffffffff80228744>{vt_map_display+500} <ffffffff801e2910>{memset+=
32}
        <ffffffff801ef790>{dumbcon_add+80}=20
<ffffffff80467a75>{dumb_console_init+2 1}
        <ffffffff8046a79f>{vty_init+335} <ffffffff80469f9e>{tty_init+462}
        <ffffffff8010c0e9>{init+169} <ffffffff8010eccb>{child_rip+8}
        <ffffffff8010c040>{init+0} <ffffffff8010ecc3>{child_rip+0}

Console: mono dummy device 80x25 vc:17-17
-------------------------------------------------------------

Anyway, it does not seem to cause problems during runtime.
I used "dumbcon=3D1:1", I don't know what causes the above messages,
I haven't looked at your code yet.