Re: acer swift 7, atheros qca6174 wireless and intel hd 615 video

[Jonathan Gray]

On Sun, Dec 04, 2016 at 05:18:47PM -0600, Peter Miller wrote:
> On Sat, Dec 3, 2016 at 8:24 PM, Jonathan Gray  wrote:
> > On Sat, Dec 03, 2016 at 11:57:18AM -0600, Peter Miller wrote:
> >> On Mon, Nov 28, 2016 at 11:23 AM, Stefan Sperling  wrote:
> >> > On Mon, Nov 28, 2016 at 11:09:12AM -0600, Peter Miller wrote:
> >> >> As for the wifi, I don't see support for the atheros 6174 chipest in
> >> >> the man pages, and I don't know if anyone is working on it. As of now
> >> >> it is "unknown product" in the dmesg. If it is not soldered in, I
> >> >> would be willing to remove it and ship it to a developer, and just
> >> >> replace it with a card that is supported.
> >> >
> >> > Yes, swap it. AFAIK nobody is actively working on atheros drivers.
> >> >
> >> > Get a device supported by iwn(4) or iwm(4) instead.
> >> > Those are currently best for laptops.
> >> >
> >> > Beware of MiniPCIe vs. M.2 connectors (make sure you get a device
> >> > that will actually fit the slot on the motherboard). AFAIK the only
> >> > M.2 device we support is the iwm(4) 8260 chip generation.
> >>
> >> Thanks Stefan.
> >>
> >> I was hoping someone might chime in on whether X should be working for
> >> my setup or not.
> >>
> >> I have the Intel HD Graphics 615 (GT2) the device is 8086:591e
> >> from 
> >> https://cgit.freedesktop.org/xorg/driver/xf86-video-intel/tree/src/i915_pciids.h
> >> I can see the 591e id is listed, so it looks like the latest Xorg will
> >> recognize and support my card. It was added back in November of 2015.
> >> https://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/src/i915_pciids.h?id=c446a7ccc783e3ca00b4b15d017c6e3af66dc646
> >>
> >> from openbsd xenocara
> >> http://cvsweb.openbsd.org/cgi-bin/cvsweb/xenocara/driver/xf86-video-intel/src/i915_pciids.h?rev=1.2=text/x-cvsweb-markup
> >> it looks like 591e has not been added to OpenBSD yet.
> >>
> >> So I should not expect the 'intel' driver to work, and also from dmesg I 
> >> see
> >> vendor "Intel", unknown product 0x591e (class display subclass VGA,
> >> rev 0x02) at pci0 dev 2 function 0 not configured
> >
> > The intel xorg driver doesn't do much of anything now.  The inteldrm
> > driver in the kernel is what does the modesetting and memory management.
> > It does not currently support the generation of hardware you have
> > (kaby lake).
> >
> >>
> >> So despite all this, my understanding is that X should still work, but
> >> just with fbdev or vesa for now.
> >
> > I believe the wsfb xorg driver should work with efifb.
> >
> >>
> >> I tried setting machdep.allowaperture=1 and 2 and got different error
> >> messages. I'm not sure if that is significant. I've only ever set it
> >> to 1 as per the install message.
> >>
> >> As an aside: the faq, https://www.openbsd.org/faq/faq11.html, says
> >> that the aperture driver must be enabled and can be done by answering
> >> 'yes' to the question, 'do you expect to run the X window system'
> >> during installation. I said yes and I still had to add
> >> machdep.allowaperture=1 to /etc/sysctl.conf. Shouldn't that have
> >> happened by answering yes, or does 'yes' set a value somewhere else?
> >> This was with the Nov 17 snapshot.
> >>
> >> with machdep.allowaperture=2 Xorg error is
> >> [  1117.767] (II) VESA(0): initializing int10
> >> [  1117.768] (EE) VESA(0): V_BIOS address 0x389e0 out of range
> >> [  1117.768] (II) UnloadModule: "vesa"
> >>
> >> with machdep.allowaperture=1 the error is
> >> [  1792.791] (II) VESA(0): initializing int10
> >> [  1792.792] (EE) VESA(0): Cannot read int vect
> >> [  1792.792] (II) UnloadModule: "vesa"
> >>
> >> I forgot to copy the fbdev error output, but it said "module does not 
> >> exist"
> >>
> >> Is there any hope with vesa, or do I just wait for the Xorg intel
> >> driver support to trickle down to OpenBSD?
> >>
> >> --
> >> Later
> >> Peter
> >>
> 
> thank you! I never new about wsfb, but that did the trick.
> 
> thanks again, and thanks to everyone involved in OpenBSD

Ideally xorg would just pick wsfb for this case, it isn't clear
to me why it doesn't.

xenocara/xserver/hw/xfree86/common/xf86Config.c

has

static int
is_fallback(const char *s)
{
/* later entries are less preferred */
const char *fallback[5] = { "modesetting", "fbdev", "vesa",  "wsfb", NULL };

xenocara/xserver/hw/xfree86/common/xf86pciBus.c

Is what adds intel to the list if the vid is 8086,
unless the pid is i740, powervr or broadwell.

xenocara/xserver/hw/xfree86/common/xf86AutoConfig.c

has

#if defined(__OpenBSD__)
#if !defined(__i386__) && !defined(__amd64__)
if (i == 0) {
matches[i++] = xnfstrdup("wsfb");
}
#endif

could the arch ifdef be removed without breaking vesa only setups?

Re: acer swift 7, atheros qca6174 wireless and intel hd 615 video

[Peter Miller]

On Sat, Dec 3, 2016 at 8:24 PM, Jonathan Gray  wrote:
> On Sat, Dec 03, 2016 at 11:57:18AM -0600, Peter Miller wrote:
>> On Mon, Nov 28, 2016 at 11:23 AM, Stefan Sperling  wrote:
>> > On Mon, Nov 28, 2016 at 11:09:12AM -0600, Peter Miller wrote:
>> >> As for the wifi, I don't see support for the atheros 6174 chipest in
>> >> the man pages, and I don't know if anyone is working on it. As of now
>> >> it is "unknown product" in the dmesg. If it is not soldered in, I
>> >> would be willing to remove it and ship it to a developer, and just
>> >> replace it with a card that is supported.
>> >
>> > Yes, swap it. AFAIK nobody is actively working on atheros drivers.
>> >
>> > Get a device supported by iwn(4) or iwm(4) instead.
>> > Those are currently best for laptops.
>> >
>> > Beware of MiniPCIe vs. M.2 connectors (make sure you get a device
>> > that will actually fit the slot on the motherboard). AFAIK the only
>> > M.2 device we support is the iwm(4) 8260 chip generation.
>>
>> Thanks Stefan.
>>
>> I was hoping someone might chime in on whether X should be working for
>> my setup or not.
>>
>> I have the Intel HD Graphics 615 (GT2) the device is 8086:591e
>> from 
>> https://cgit.freedesktop.org/xorg/driver/xf86-video-intel/tree/src/i915_pciids.h
>> I can see the 591e id is listed, so it looks like the latest Xorg will
>> recognize and support my card. It was added back in November of 2015.
>> https://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/src/i915_pciids.h?id=c446a7ccc783e3ca00b4b15d017c6e3af66dc646
>>
>> from openbsd xenocara
>> http://cvsweb.openbsd.org/cgi-bin/cvsweb/xenocara/driver/xf86-video-intel/src/i915_pciids.h?rev=1.2=text/x-cvsweb-markup
>> it looks like 591e has not been added to OpenBSD yet.
>>
>> So I should not expect the 'intel' driver to work, and also from dmesg I see
>> vendor "Intel", unknown product 0x591e (class display subclass VGA,
>> rev 0x02) at pci0 dev 2 function 0 not configured
>
> The intel xorg driver doesn't do much of anything now.  The inteldrm
> driver in the kernel is what does the modesetting and memory management.
> It does not currently support the generation of hardware you have
> (kaby lake).
>
>>
>> So despite all this, my understanding is that X should still work, but
>> just with fbdev or vesa for now.
>
> I believe the wsfb xorg driver should work with efifb.
>
>>
>> I tried setting machdep.allowaperture=1 and 2 and got different error
>> messages. I'm not sure if that is significant. I've only ever set it
>> to 1 as per the install message.
>>
>> As an aside: the faq, https://www.openbsd.org/faq/faq11.html, says
>> that the aperture driver must be enabled and can be done by answering
>> 'yes' to the question, 'do you expect to run the X window system'
>> during installation. I said yes and I still had to add
>> machdep.allowaperture=1 to /etc/sysctl.conf. Shouldn't that have
>> happened by answering yes, or does 'yes' set a value somewhere else?
>> This was with the Nov 17 snapshot.
>>
>> with machdep.allowaperture=2 Xorg error is
>> [  1117.767] (II) VESA(0): initializing int10
>> [  1117.768] (EE) VESA(0): V_BIOS address 0x389e0 out of range
>> [  1117.768] (II) UnloadModule: "vesa"
>>
>> with machdep.allowaperture=1 the error is
>> [  1792.791] (II) VESA(0): initializing int10
>> [  1792.792] (EE) VESA(0): Cannot read int vect
>> [  1792.792] (II) UnloadModule: "vesa"
>>
>> I forgot to copy the fbdev error output, but it said "module does not exist"
>>
>> Is there any hope with vesa, or do I just wait for the Xorg intel
>> driver support to trickle down to OpenBSD?
>>
>> --
>> Later
>> Peter
>>

thank you! I never new about wsfb, but that did the trick.

thanks again, and thanks to everyone involved in OpenBSD


-- 
Later
Peter

Re: Setting MAC address of vm in vm.conf with lladdr

[Reyk Floeter]

Hi,

you are second person who repots this since Friday,
I will check tomorrow if lladdr in vm.conf got broken.

The config looks OK. 

Reyk

> Am 04.12.2016 um 16:55 schrieb Eric Brown :
> 
> Dear List,
> 
> I am using the current snapshot (Dec 3 as of this post), and I am trying
> to set the MAC address of a vm host in vm.conf.
> 
> However, the MAC address reported by ifconfig -a seems to change with
> each restart. The lladdr that I had typed in was from a copy/paste of a
> random assignment, so I hope that it is valid.
> 
> I know that this is still a WIP, but I thought I would ask in case I'm
> making a dumb mistake.
> 
> Thanks, 
> Eric
> 
> 
> -
> /etc/vm.conf:
> 
> vm "current.ericcbrown.com" {
>  memory 2048M
>  kernel "/root/vmm/current/bsd"
>  disk "/var/vmm/current/disk.img"
>  interface tap {
>lladdr fe:e1:ba:d1:77:24
>switch uplink
>  } 
> }
> 
> switch uplink {
>  add bge0
> }

Re: OpenBSD's binutils

[Jeremie Courreges-Anglas]

Stuart Longland  writes:

> On 28/11/16 22:13, Jeremie Courreges-Anglas wrote:
>> Stuart Longland  writes:
>> 
>>> Hi all,
>>>
>>> Well, my experiments with loongson have proven fruitful, except for the
>>> need of a web browser.  Part of this is due to the webkit browsers
>>> needing the libav gstreamer plug-in.
>>>
>>> Now, this fails to build because of the assembler not recognising some
>>> op codes generated by gcc.  I found I got a little further in the build
>>> using the `gcc4` package, but it still failed to compile, failing with
>>> an "unrecognized opcode ldx" error message.
>> 
>> What's the error log when you try with gcc from base?
>> 
>
> Okay, did a clean and a rebuild using base image gcc:
>
>> cc -I. -I./ 
>> -I/usr/local/ports/obj/gstreamer1-plugins-libav-1.8.2/gst-libav-1.8.2/gst-libs/ext/libav
>>  -I/usr/local/include -I/usr/X11R6/include -EL -D_ISOC99_SOURCE 
>> -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -DPIC -DZLIB_CONST 
>> -DHAVE_AV_CONFIG_H -O2 -pipe -std=c99 -fPIC -mips64 -mhard-float -mfp64 
>> -mdsp -pthread -I/usr/X11R6/include -I/usr/X11R6/include 
>> -I/usr/X11R6/include -I/usr/X11R6/include -Wdeclaration-after-statement 
>> -Wall -Wdisabled-optimization -Wpointer-arith -Wredundant-decls 
>> -Wwrite-strings -Wundef -Wmissing-prototypes -Wno-pointer-to-int-cast 
>> -Wstrict-prototypes -Wno-parentheses -Wno-switch -Wno-format-zero-length 
>> -Wno-pointer-sign -fno-math-errno -fno-tree-vectorize 
>> -Werror=format-security -Werror=implicit-function-declaration 
>> -Werror=missing-prototypes -Werror=return-type -Wformat -MMD -MF
>> libavcodec/mpegaudiodec_fixed.d -MT
>> libavcodec/mpegaudiodec_fixed.o -c -o libavcodec/mpegaudiodec_fixed.o
>> libavcodec/mpegaudiodec_fixed.c
>> {standard input}: Assembler messages:
>> {standard input}:3876: Error: illegal operands `mult $ac0,$11,$8'
>> {standard input}:3877: Error: illegal operands `mult $ac1,$11,$8'
>> {standard input}:3883: Error: illegal operands `msub $ac0,$12,$7'

[...]

Looks like the -mdsp flag is the problem.  You could give this patch
a try (stolen from graphics/ffmpeg):

Index: Makefile
===
RCS file: /d/cvs/ports/multimedia/gstreamer1/plugins-libav/Makefile,v
retrieving revision 1.31
diff -u -p -p -u -r1.31 Makefile
--- Makefile10 Apr 2016 10:14:12 -  1.31
+++ Makefile4 Dec 2016 16:31:53 -
@@ -39,6 +39,11 @@ libav_CONFIGURE_ARGS=--arch=${GST_LIBAV
--disable-armv5te \
--disable-armv6 \
--disable-armv6t2 \
+   --disable-mips32r5 \
+   --disable-mips64r6 \
+   --disable-mipsdspr1 \
+   --disable-mipsdspr2 \
+   --disable-mipsfpu \
--disable-sdl \
--disable-vfp \
--disable-debug


-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

Setting MAC address of vm in vm.conf with lladdr

[Eric Brown]

Dear List,

I am using the current snapshot (Dec 3 as of this post), and I am trying
to set the MAC address of a vm host in vm.conf.

However, the MAC address reported by ifconfig -a seems to change with
each restart. The lladdr that I had typed in was from a copy/paste of a
random assignment, so I hope that it is valid.

I know that this is still a WIP, but I thought I would ask in case I'm
making a dumb mistake.

Thanks, 
Eric


-
/etc/vm.conf:

vm "current.ericcbrown.com" {
  memory 2048M
  kernel "/root/vmm/current/bsd"
  disk "/var/vmm/current/disk.img"
  interface tap {
lladdr fe:e1:ba:d1:77:24
switch uplink
  } 
}

switch uplink {
  add bge0
}

Re: Cups working but cannot print pdf from Firefox.

[Maurice McCarthy]

I find brute force and ignorance quite good. 

$ doas pkg_add -ui cups # for example

reinstalls cups and all dependencies. Then re-add your printer at

$ firefox http://localhost:631 

I too cannot print a pdf from Firefox directly to my HP Deskjet 2540 MFI, 
connected directly to the usb. Webpages do print and saving the pdf to disk 
and using xpdf to print works around the problem.

The cups logs when attempting to print a pdf directly from Firefox give:

/var/log/cups/access_log :
"POST /printers/hp HTTP/1.1" 200 403 Print-Job client-error-bad-request

/var/log/cups/error_log :
HTTP_STATE_WAITING Closing for error 32 (Broken pipe)

Lastly you might put "LogLevel debug" into /etc/cups/cupsd.conf and 

$ doas rcctl restart cupsd

see what it gives you. After that it is beyond me. :) My apologies if I am
talking well below your understanding. It is my level. 



On Sat, Dec 03, 2016 at 05:25:54PM -0500 or thereabouts, Michael Hunt wrote:
> Cups interface recognizes my Xerox_Phaser_6010N printer but when I go to print
> a pdf displayed in Firefox or any other page from it It looks like it???s
> going to happen and I get the job listed as Completed but the printer didn???t
> do anything.
> 
> Has anyone had any success with using CUPS to print. Cups detected my Xerox on
> the network so it looked like it would work but the printer doesn???t seem to
> respond. It seems to print on port 9100 but that didn???t seem to work for me
> either. Do I need to use a specific protocol with that port number?
> 
> Yes I have read the help file and got the ???gtk+3-cups??? package so that I
> can print from Firefox and it shows the Xerox printer in the Firefox print
> dialog. It sends to the printer but the printer does nothing as I stated
> before. I know the printer works because I print from my MacBook Pro as well
> to the network.
> ##
> Cups config of this printer looks like this:
> Description: Xerox Phaser 6010N
> Location: Local Printer
> Driver: Xerox Phaser 6100 Foomatic/Postscript (grayscale, 2-sided printing)
> Connection: "socket://192. 168.1.101
> Defaults: job-sheets=none, none media=na_letter_8.5x11in sides=one-sided.
> ##
> 
> I???d appreciate any advice on how to troubleshoot this or get the printer to
> respond.