Re: [9front] bounty for 9pi update for 9front

Re: [9front] bounty for 9pi update for 9front

[cinap_lenrek]

ok, this weekend is over. what we got so far:

richard millers kernel changes where merged giving you
two kernel targets 9pi and 9pi2.

lan78xx driver ported to nusb/ether for raspi3. works.

nusb/kb hacks added to work arround broken split transactions
to avoid loosing key events when the ethernet is active.

stub mkfile created in /sys/src/boot/bcm to download raspi
bootcode.

mkfile target *.pi.img added to /sys/lib/dist/mkfile to
generate 2GB hjfs sdcard image with bootcode, kernels and
config.txt + cmdline.txt.

got some ch340 usb serial converter, added support for it
to nusb/serial.

things still todo:

fix /dev/reboot on raspi3 model.

*actually* fix split transactions in the controller driver.

implement isochronous transfers.

port richards wifi driver (requires changes in the mmc
interface).

fix bugs.

--
cinap

Re: [9front] bounty for 9pi update for 9front

[sl]

[-- Attachment #1: Type: text/plain, Size: 59 bytes --]

Do recent commits fulfill this request?  (See attached)

sl

[-- Attachment #2: Type: message/rfc822, Size: 1572 bytes --]

From: Antoni Sawicki <tenox@SDF.ORG>
To: 9front@9front.org
Subject: [9front] bounty for 9pi update for 9front
Date: Wed, 12 Sep 2018 05:27:24 +0000 (UTC)
Message-ID: <Pine.NEB.4.64.1809120508170.1816@sdf.lonestar.org>

Hi all

Here is my pledge. Recently (April 2018) Richard Miller posted an update 
of a 9pi image supporting Raspbbery PI 3B+ with USB, Eth, Video etc. I 
would like to see this integrated in to 9front. What I offer:

Hardware (Raspberry PI 3B+) for you to keep forever.
$100 for a first working image for RPI 3B+ with latest 9front.
$100 for a permanent integration to build process so it will be 
automatically build with every 9front release in future. You get that on 
a subsequent 9front release.

By a working image I mean something akin to 9pi.img that you dd to the sd 
card and just works

I will only sponsor RPI hardware to serious contenders. Please email me 
ahead of time.

a

Re: [9front] bounty for 9pi update for 9front

[Ethan Gardener]

On Fri, Sep 14, 2018, at 2:39 PM, Steve Simon wrote:
> good grief.

Most of that was my penchant for melodrama, but still... 

> i use richard’s plan9 package and has been solid 

Did you get lucky, or did you buy hardware to compensate for the Pi's poor design?

> this may be the lack of graphics acceleration (fixable)

Some second-hand Thinkpads don't need graphics acceleration to have good performance, but VNC isn't perfect with modern web browsers and video even at 1Gbps.  ("Some":  It apparently depends whether the BIOS disables caching for the video region.  If it does, performance dies.)

> or the “ethernet 
> over usb” on the pi

I wrote my melodrama still imagining Pi 2 & 3 had real ethernet...

Anyway, I am apparently quite depressed today, so I'm sorry.  It's not a good time of year for me.

Re: [9front] bounty for 9pi update for 9front

[Steve Simon]

good grief.

i use a pi as my desktop and have done so for 3 years - first a 2 and now a pi3. i use richard’s plan9 package and has been solid (modulo my own mistakes).

the only time it has an issue is the noticeable lack of performance when remote desktop-ing to  a windows box to get a modern browser. even this works well but scrolling windows is laggy.

this may be the lack of graphics acceleration (fixable) or the “ethernet over usb” on the pi (which i am stuck with); i have not investigated.

there is most probably better hardware available these days, but the pi works, is reliable and cheap.

-Steve

> On 14 Sep 2018, at 1:29 pm, Ethan Gardener <eekee57@fastmail.fm> wrote:

> I had a bad impression of the Raspberry Pi from the beginning.  I sort-of assumed Pi 2 & 3 were better due to sheer mass of peer pressure, but, well, thanks for stating the facts before I bought one.  I just can't understand the demented hurricane of enthusiasm generated by these garbage fire computers!
> 
> You know what?  When I publish my own OS (ha ha), I'll put a filter into the mailing list (when I get around to it), catching all messages mentioning Raspberry Pi or RPi to insert something like,  "WARNING!  This email mentions Raspberry Pi!  It may be trying to suck us into the demented maelstrom of enthusiasm generated by these garbage fire computers.  Exercise caution and critical thinking.  Remember the principles of rational development." XD
> 
> In a different humour, would it make sense to bit-bang ps/2 interfaces with the RPi's GPIO pins?  Or other serial interfaces.

Re: [9front] bounty for 9pi update for 9front

[Ethan Gardener]

On Thu, Sep 13, 2018, at 6:40 PM, Chris McGee wrote:
> I had only wireless and minimal peripherals on my pis (keyboard and 
> mouse). I tried different power supplied including battery power to rule 
> out bad power or noise. I could reproduce the problem on both pi2 and 
> pi3 quite reliably and also two other people noticed it too. Pi1 worked 
> fine though.
> 
> The symptom was very strange phantom keyboard key presses and button 
> clicks. They would start at random points in time and even sometimes 
> during boot. I traced it all of the way to the USB debugging the HID 
> packets. At the lowest level it appeared as though the keyboard was 
> actually sending Shift-Ctrl-A/B keypresses. I was expecting some kind of 
> memory corruption type problems due to mulicore but it didn’t seem so.
> 
> I think I finally threw in the towel once I saw that even raspbian/Linux 
> have USB hacks abound and state on their website that not all USB 
> devices may work with it. That and the lack of any good hardware specs 
> on how the controller works.
> 
> I’m now in search of better hardware and left with a bad impression of 
> raspberry pi and USB itself. I hear from some hardware folks that PCIe 
> and Ethernet are much better and Thunderbolt is a dumpster fire. Maybe 
> there will be a riscv dev board soon.

I had a bad impression of the Raspberry Pi from the beginning.  I sort-of assumed Pi 2 & 3 were better due to sheer mass of peer pressure, but, well, thanks for stating the facts before I bought one.  I just can't understand the demented hurricane of enthusiasm generated by these garbage fire computers!

You know what?  When I publish my own OS (ha ha), I'll put a filter into the mailing list (when I get around to it), catching all messages mentioning Raspberry Pi or RPi to insert something like,  "WARNING!  This email mentions Raspberry Pi!  It may be trying to suck us into the demented maelstrom of enthusiasm generated by these garbage fire computers.  Exercise caution and critical thinking.  Remember the principles of rational development." XD

In a different humour, would it make sense to bit-bang ps/2 interfaces with the RPi's GPIO pins?  Or other serial interfaces.

Re: [9front] bounty for 9pi update for 9front

[Antoni Sawicki]

> Sending hardware to cinap is the best way to ensure that it works with the 
> least amount of effort and will continue to work in the future.

Good point, I will do that regardless of the bounty.

Re: [9front] bounty for 9pi update for 9front

[Chris McGee]

I had only wireless and minimal peripherals on my pis (keyboard and mouse). I tried different power supplied including battery power to rule out bad power or noise. I could reproduce the problem on both pi2 and pi3 quite reliably and also two other people noticed it too. Pi1 worked fine though.

The symptom was very strange phantom keyboard key presses and button clicks. They would start at random points in time and even sometimes during boot. I traced it all of the way to the USB debugging the HID packets. At the lowest level it appeared as though the keyboard was actually sending Shift-Ctrl-A/B keypresses. I was expecting some kind of memory corruption type problems due to mulicore but it didn’t seem so.

I think I finally threw in the towel once I saw that even raspbian/Linux have USB hacks abound and state on their website that not all USB devices may work with it. That and the lack of any good hardware specs on how the controller works.

I’m now in search of better hardware and left with a bad impression of raspberry pi and USB itself. I hear from some hardware folks that PCIe and Ethernet are much better and Thunderbolt is a dumpster fire. Maybe there will be a riscv dev board soon.

> On Sep 13, 2018, at 11:01 AM, Steve Simon <steve@quintile.net> wrote:
> 
> i would try it at the very least.
> 
> i have been looking for an excuse to install 9front.
> 
> personally i would like audio support, and a prebuilt image which tracks the 9front 386/amd64 distribution.
> 
> my 2p’s worth
> 
> -Steve
> 
>> On 13 Sep 2018, at 2:29 pm, Julius Schmidt <aiju@phicode.de> wrote:
>> 
>> Sending hardware to cinap is the best way to ensure that it works with the least amount of effort and will continue to work in the future.
>> 
>>> On Wed, 12 Sep 2018, Antoni Sawicki wrote:
>>> 
>>> Hi all
>>> 
>>> Here is my pledge. Recently (April 2018) Richard Miller posted an update of a 9pi image supporting Raspbbery PI 3B+ with USB, Eth, Video etc. I would like to see this integrated in to 9front. What I offer:
>>> 
>>> Hardware (Raspberry PI 3B+) for you to keep forever.
>>> $100 for a first working image for RPI 3B+ with latest 9front.
>>> $100 for a permanent integration to build process so it will be automatically build with every 9front release in future. You get that on a subsequent 9front release.
>>> 
>>> By a working image I mean something akin to 9pi.img that you dd to the sd card and just works
>>> 
>>> I will only sponsor RPI hardware to serious contenders. Please email me ahead of time.
>>> 
>>> a
>>> 
> 

Re: [9front] bounty for 9pi update for 9front

[Steve Simon]

i would try it at the very least.

i have been looking for an excuse to install 9front.

personally i would like audio support, and a prebuilt image which tracks the 9front 386/amd64 distribution.

my 2p’s worth

-Steve

> On 13 Sep 2018, at 2:29 pm, Julius Schmidt <aiju@phicode.de> wrote:
> 
> Sending hardware to cinap is the best way to ensure that it works with the least amount of effort and will continue to work in the future.
> 
>> On Wed, 12 Sep 2018, Antoni Sawicki wrote:
>> 
>> Hi all
>> 
>> Here is my pledge. Recently (April 2018) Richard Miller posted an update of a 9pi image supporting Raspbbery PI 3B+ with USB, Eth, Video etc. I would like to see this integrated in to 9front. What I offer:
>> 
>> Hardware (Raspberry PI 3B+) for you to keep forever.
>> $100 for a first working image for RPI 3B+ with latest 9front.
>> $100 for a permanent integration to build process so it will be automatically build with every 9front release in future. You get that on a subsequent 9front release.
>> 
>> By a working image I mean something akin to 9pi.img that you dd to the sd card and just works
>> 
>> I will only sponsor RPI hardware to serious contenders. Please email me ahead of time.
>> 
>> a
>> 

Re: [9front] bounty for 9pi update for 9front

[Julius Schmidt]

Sending hardware to cinap is the best way to ensure that it works with the 
least amount of effort and will continue to work in the future.

On Wed, 12 Sep 2018, Antoni Sawicki wrote:

> Hi all
>
> Here is my pledge. Recently (April 2018) Richard Miller posted an update of a 
> 9pi image supporting Raspbbery PI 3B+ with USB, Eth, Video etc. I would like 
> to see this integrated in to 9front. What I offer:
>
> Hardware (Raspberry PI 3B+) for you to keep forever.
> $100 for a first working image for RPI 3B+ with latest 9front.
> $100 for a permanent integration to build process so it will be automatically 
> build with every 9front release in future. You get that on a subsequent 
> 9front release.
>
> By a working image I mean something akin to 9pi.img that you dd to the sd 
> card and just works
>
> I will only sponsor RPI hardware to serious contenders. Please email me ahead 
> of time.
>
> a
>

Re: [9front] bounty for 9pi update for 9front

[Steve Simon]

this was a problem on a pi3
it was my fault for having too much on the usb bus, the killer was an Apple usb ether adapter which is surprisingly hungry.

i am only trying to say the symptoms can be very confusing and intermittent. its an easy thing to check and, if i where able to speaking to myself last week i would say, check the rail voltage first!

-Steve


> On 13 Sep 2018, at 9:07 am, hiro <23hiro@gmail.com> wrote:
> 
> steve, is that voltage problem also there on the newest pi, i thought
> it was only on pi <3 ?

Re: [9front] bounty for 9pi update for 9front

[Antoni Sawicki]

USB patches are not a requisite for the bounty. If 9front RPI image works 
at least as good as Richard's 9pi image from April it's fine. Bugs can be 
fixed later.

Re: [9front] bounty for 9pi update for 9front

[hiro]

steve, is that voltage problem also there on the newest pi, i thought
it was only on pi <3 ?

Re: [9front] bounty for 9pi update for 9front

[Steve Simon]

two random thoughts.

richard has made some changes to use i believe in the last 6 months or so, its probably worth re -merging.

secondly - the most common problem i have seen on pi’s it a lack of psu volts. marginal rails can cause some very odd usb behaviour. i had some intermittent usb faults last week and the pi ‘s supply measured 4.1v.  a powered hub solved the problem as usual.

-Steve


-Steve


> On 13 Sep 2018, at 1:47 am, Chris McGee <sirnewton_01@yahoo.ca> wrote:
> 
> I managed to merge Richard’s changes into a 9front kernel quite a while ago and put the result on github. I had it running on pi2 and pi3 at the time.
> 
> There’s an outstanding USB problem that I found hard to track down. Partly my own lack of hardware experience and partly the horrible hardware and documentation for some aspects of the pi.
> 
> If someone wants to pick this up and fix the elusive bugs that would be great.
> 
> github.com/sirnewton01/rpi-9front
> 
> Chris
> 
>> On Sep 12, 2018, at 1:27 AM, Antoni Sawicki <tenox@SDF.ORG> wrote:
>> 
>> Hi all
>> 
>> Here is my pledge. Recently (April 2018) Richard Miller posted an update of a 9pi image supporting Raspbbery PI 3B+ with USB, Eth, Video etc. I would like to see this integrated in to 9front. What I offer:
>> 
>> Hardware (Raspberry PI 3B+) for you to keep forever.
>> $100 for a first working image for RPI 3B+ with latest 9front.
>> $100 for a permanent integration to build process so it will be automatically build with every 9front release in future. You get that on a subsequent 9front release.
>> 
>> By a working image I mean something akin to 9pi.img that you dd to the sd card and just works
>> 
>> I will only sponsor RPI hardware to serious contenders. Please email me ahead of time.
>> 
>> a

Re: [9front] bounty for 9pi update for 9front

[Chris McGee]

I managed to merge Richard’s changes into a 9front kernel quite a while ago and put the result on github. I had it running on pi2 and pi3 at the time.

There’s an outstanding USB problem that I found hard to track down. Partly my own lack of hardware experience and partly the horrible hardware and documentation for some aspects of the pi.

If someone wants to pick this up and fix the elusive bugs that would be great.

github.com/sirnewton01/rpi-9front

Chris

> On Sep 12, 2018, at 1:27 AM, Antoni Sawicki <tenox@SDF.ORG> wrote:
> 
> Hi all
> 
> Here is my pledge. Recently (April 2018) Richard Miller posted an update of a 9pi image supporting Raspbbery PI 3B+ with USB, Eth, Video etc. I would like to see this integrated in to 9front. What I offer:
> 
> Hardware (Raspberry PI 3B+) for you to keep forever.
> $100 for a first working image for RPI 3B+ with latest 9front.
> $100 for a permanent integration to build process so it will be automatically build with every 9front release in future. You get that on a subsequent 9front release.
> 
> By a working image I mean something akin to 9pi.img that you dd to the sd card and just works
> 
> I will only sponsor RPI hardware to serious contenders. Please email me ahead of time.
> 
> a