[9fans] Writing drivers in Plan 9

[9fans] Writing drivers in Plan 9

[hugo rivera]

Hello:
I've been using Inferno and Plan 9 for almost a year. I certainly love
Plan 9's ideas and concepts, and I'd be glad to finally move forward
and leave Unix behind, but for now it is imposible for me, since my
work does not allow me this (I do data analysis for some physics
experiment using Cern's ROOT).
Anyway, I've always wanted to learn to write drivers and I did try it
a few years ago, using NetBSD, but I could not find time to complete
the task, so I leaved it unfinished. I want to start again, but this
time using Plan 9. So, I want to ask you for some pointers about how
doing this. Please, bear in mind, that I am a complete newbie when it
comes about writing drivers.
Saludos

Hugo

Re: [9fans] Writing drivers in Plan 9

[a]

Given that you're already into Inferno as well as Plan 9, I
think a really nice way to get into driver development is
with emu drivers for Inferno. The basic structure is the
same as for native OS drivers in either system: implement
a small set of entry points (fooattach, fooread, &c),
making use of the dev* defaults where appropriate. It's
then a much smaller step to move into native hardware
drivers for Plan 9 or Inferno. And when you do get to that
point, take a look at Inferno's os/port/devXXX.c; it's a
template for what you need to implement. In any case,
the devtab (look for '^Dev') towards the end of any driver
will tell you what you really need. It's a much nicer,
constrained set for Inferno and Plan 9 than any unix.
Anthony

Re: [9fans] Writing drivers in Plan 9

[erik quanstrom]

> Hello:
> I've been using Inferno and Plan 9 for almost a year. I certainly love
> Plan 9's ideas and concepts, and I'd be glad to finally move forward
> and leave Unix behind, but for now it is imposible for me, since my
> work does not allow me this (I do data analysis for some physics
> experiment using Cern's ROOT).
> Anyway, I've always wanted to learn to write drivers and I did try it
> a few years ago, using NetBSD, but I could not find time to complete
> the task, so I leaved it unfinished. I want to start again, but this
> time using Plan 9. So, I want to ask you for some pointers about how
> doing this. Please, bear in mind, that I am a complete newbie when it
> comes about writing drivers.
> Saludos

a few thoughts from inexperience

0.  pick something you want to use.
1.  you must have a datasheet.  reading other drivers for the same
h/w can be very misleading before you're done.
2.  implement as few features as you need to get the hardware
working for your purposes.  it's likely the other stuff will never
be needed,
3.  set up an whole plan 9 environment which can pxeboot
your target machine.  if you're like me, you'll be rebooting
quite a bit.
4,  this is still true
	The most effective debugging tool is still careful thought,
	coupled with judiciously placed print statements.
		-Kernighan, 1978

- erik

Re: [9fans] Writing drivers in Plan 9

[erik quanstrom]

> Given that you're already into Inferno as well as Plan 9, I
> think a really nice way to get into driver development is
> with emu drivers for Inferno. The basic structure is the
> same as for native OS drivers in either system: implement
> a small set of entry points (fooattach, fooread, &c),
> making use of the dev* defaults where appropriate. It's
> then a much smaller step to move into native hardware
> drivers for Plan 9 or Inferno. And when you do get to that
> point, take a look at Inferno's os/port/devXXX.c; it's a
> template for what you need to implement. In any case,
> the devtab (look for '^Dev') towards the end of any driver
> will tell you what you really need. It's a much nicer,
> constrained set for Inferno and Plan 9 than any unix.
> Anthony

i think this confuses implementing a Dev interface with writing
a device driver.  for many devices, the Dev interface is already
taken care of.  for example, serial, ethernet, disk devices using
sd implement an interface to devsd, ethernet.

i don't buy the thesis that talking to hardware is always hard.
talking to some hardware can be hard.  for exampe, the aoe driver
doesn't talk to hardware, it talks to the ethernet drivers.  yet it's
the largest driver i've written, largely because it implements its own
dev interface.

i think it's a mistake to think hardware == hard, software interfaces
== easy.

- erik

Re: [9fans] Writing drivers in Plan 9

[a]

> i think this confuses implementing a Dev interface with writing
> a device driver.  for many devices, the Dev interface is already
> taken care of.  for example, serial, ethernet, disk devices using
> sd implement an interface to devsd, ethernet.

i think the Dev interface is still the right place to start, in terms of
understanding things.

for the benifit of the original question, the Dev interface (that
common set of entry points i was talking about) is the common
kernel interface that all device drivers have to go through at some
point. i think part of erik's point (which is correct) is that many of
the things people are commonly writing drivers for - disk
controllers, ethernet cards, and vga cards being probably the most
common examples - already have that interface covered, and
there's a separate interface that the hardware-specific part needs
to talk to.

> i don't buy the thesis that talking to hardware is always hard.
> talking to some hardware can be hard.  for exampe, the aoe driver
> doesn't talk to hardware, it talks to the ethernet drivers.  yet it's
> the largest driver i've written, largely because it implements its own
> dev interface.

but working with Dev doesn't need to be so complex; it depends, at a
minimum, on what the job you're trying to do is. dup and env
both implement at least most of their own Dev interface (as opposed
to relying on many of the default stubs), but are reasonably short and
easy to understand. devaoe's hardly a fair comparison; it's not only
the largest driver you've written, it's the largest dev*.c in the system.

> i think it's a mistake to think hardware == hard, software interfaces
> == easy.

agreed. but it's also a mistake, at least in the context of Plan 9, to
assume that device drivers inherently involve hardware. about a third
of the things in section three of the manual don't.

anthony

Re: [9fans] Writing drivers in Plan 9

[hugo rivera]

Thanks for your feedback.
Now, I think I am going to start by trying to understand a little
about the Dev interface you talk about, and then continue to write a
real driver for a gamepad that I have.
Is there any documentation that describes this Dev interface?
This is a usb gamepad, so probably I have to deal with a Dev interface
related to the usb ports, am I right?

Hugo

2008/4/15, a@9srv.net <a@9srv.net>:
> > i think this confuses implementing a Dev interface with writing
>  > a device driver.  for many devices, the Dev interface is already
>  > taken care of.  for example, serial, ethernet, disk devices using
>  > sd implement an interface to devsd, ethernet.
>
>
> i think the Dev interface is still the right place to start, in terms of
>  understanding things.
>
>  for the benifit of the original question, the Dev interface (that
>  common set of entry points i was talking about) is the common
>  kernel interface that all device drivers have to go through at some
>  point. i think part of erik's point (which is correct) is that many of
>  the things people are commonly writing drivers for - disk
>  controllers, ethernet cards, and vga cards being probably the most
>  common examples - already have that interface covered, and
>  there's a separate interface that the hardware-specific part needs
>  to talk to.
>
>
>  > i don't buy the thesis that talking to hardware is always hard.
>  > talking to some hardware can be hard.  for exampe, the aoe driver
>  > doesn't talk to hardware, it talks to the ethernet drivers.  yet it's
>  > the largest driver i've written, largely because it implements its own
>  > dev interface.
>
>
> but working with Dev doesn't need to be so complex; it depends, at a
>  minimum, on what the job you're trying to do is. dup and env
>  both implement at least most of their own Dev interface (as opposed
>  to relying on many of the default stubs), but are reasonably short and
>  easy to understand. devaoe's hardly a fair comparison; it's not only
>  the largest driver you've written, it's the largest dev*.c in the system.
>
>
>  > i think it's a mistake to think hardware == hard, software interfaces
>  > == easy.
>
>
> agreed. but it's also a mistake, at least in the context of Plan 9, to
>  assume that device drivers inherently involve hardware. about a third
>  of the things in section three of the manual don't.
>
>  anthony
>
>
>

Re: [9fans] Writing drivers in Plan 9

[erik quanstrom]

> Thanks for your feedback.
> Now, I think I am going to start by trying to understand a little
> about the Dev interface you talk about, and then continue to write a
> real driver for a gamepad that I have.
> Is there any documentation that describes this Dev interface?
> This is a usb gamepad, so probably I have to deal with a Dev interface
> related to the usb ports, am I right?
>
> Hugo

probablly not.  the usb stuff is all accessable from userland.  usbd
runs in user space as do all the clients of usb -- usb/disk usb/mouse, etc.

- erik

Re: [9fans] Writing drivers in Plan 9

[hugo rivera]

> probablly not.  the usb stuff is all accessable from userland.  usbd
>  runs in user space as do all the clients of usb -- usb/disk usb/mouse, etc.

So, this means that I never have to deal with usb interfaces?
Can you point me some documentation about this?

Hugo

Re: [9fans] Writing drivers in Plan 9

[erik quanstrom]

> So, this means that I never have to deal with usb interfaces?
> Can you point me some documentation about this?
>
> Hugo

it should be in the manul pages.  try
	; lookman usb

- erik

Re: [9fans] Writing drivers in Plan 9

[hugo rivera]

> it should be in the manul pages.  try
>         ; lookman usb

Great! thanks

Hugo

Re: [9fans] Writing drivers in Plan 9

[Tom Lieber]

On Tue, Apr 15, 2008 at 4:40 AM, hugo rivera <uair00@gmail.com> wrote:
> Anyway, I've always wanted to learn to write drivers and I did try it
> a few years ago, using NetBSD, but I could not find time to complete
> the task, so I leaved it unfinished. I want to start again, but this
> time using Plan 9. So, I want to ask you for some pointers about how
> doing this. Please, bear in mind, that I am a complete newbie when it
> comes about writing drivers.

Did you do it? How did it go?

--
Tom Lieber
http://AllTom.com/

Re: [9fans] Writing drivers in Plan 9

[hugo rivera]

Sadly I cannot find time for Plan 9 for now. I am just too busy with
my job in the university and preparing me for some tests in October.
Hopefully I can start this thing in November.

2008/9/3 Tom Lieber <alltom@gmail.com>:
> On Tue, Apr 15, 2008 at 4:40 AM, hugo rivera <uair00@gmail.com> wrote:
>> Anyway, I've always wanted to learn to write drivers and I did try it
>> a few years ago, using NetBSD, but I could not find time to complete
>> the task, so I leaved it unfinished. I want to start again, but this
>> time using Plan 9. So, I want to ask you for some pointers about how
>> doing this. Please, bear in mind, that I am a complete newbie when it
>> comes about writing drivers.
>
> Did you do it? How did it go?
>
> --
> Tom Lieber
> http://AllTom.com/
>
>



--
Saludos

Hugo