slackware on a very fast machine

slackware on a very fast machine

[Jude DaShiell]

I can't determine when the 10 seconds starts or ends with 
slackware-Current on this machine since the machine appears to be running 
too fast to get any usable audible indicators from it any longer.  What is 
the correct syntax to modprobe speakup and have speakup come up on a 
litetalk synthesizer on ttyS0?  I might be able to get slackware to run 
the speakup.s kernel and if I can do that, I ought to be able to key the 
modprobe line in once I get in as root.



jude <jdashiel@shellworld.net>

Re: slackware on a very fast machine

[John G. Heim]

The syntax is "modprobe speakup_ltlk".

But it isn't going to help unless  slackware hacked the kernel clode to 
fix the serial port bug.  If they used the speakup code as it is in the 
kernel, your litetalk won't work.
I started trying to make a fork of grml with a patched kernel but never 
had the time to really get into it. Instead I've been using an FAI disk. 
FAI is actually an installer system. FAI stands for fully automated 
install. But they make it real easy to create a live CD.
On 06/26/14 14:12, Jude DaShiell wrote:
> I can't determine when the 10 seconds starts or ends with
> slackware-Current on this machine since the machine appears to be running
> too fast to get any usable audible indicators from it any longer.  What is
> the correct syntax to modprobe speakup and have speakup come up on a
> litetalk synthesizer on ttyS0?  I might be able to get slackware to run
> the speakup.s kernel and if I can do that, I ought to be able to key the
> modprobe line in once I get in as root.
>
>
>
> jude <jdashiel@shellworld.net>
>
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>

-- 
---
John G. Heim, 608-263-4189, jheim@math.wisc.edu

Re: slackware on a very fast machine

[Rob Hudson]

I don't understand why the kernel has serial bugs. Many many many companies 
still use serial consoles, and in fact, as faras I know, it is still the 
most common way of interacting directly with rack-mounted computers and some 
high end routers. What's the deal?
----- Original Message ----- 
From: "John G. Heim" <jheim@math.wisc.edu>
To: "Speakup is a screen review system for Linux." 
<speakup@linux-speakup.org>
Sent: Thursday, June 26, 2014 2:24 PM
Subject: Re: slackware on a very fast machine


> The syntax is "modprobe speakup_ltlk".
>
> But it isn't going to help unless  slackware hacked the kernel clode to 
> fix the serial port bug.  If they used the speakup code as it is in the 
> kernel, your litetalk won't work.
> I started trying to make a fork of grml with a patched kernel but never 
> had the time to really get into it. Instead I've been using an FAI disk. 
> FAI is actually an installer system. FAI stands for fully automated 
> install. But they make it real easy to create a live CD.
> On 06/26/14 14:12, Jude DaShiell wrote:
>> I can't determine when the 10 seconds starts or ends with
>> slackware-Current on this machine since the machine appears to be running
>> too fast to get any usable audible indicators from it any longer.  What 
>> is
>> the correct syntax to modprobe speakup and have speakup come up on a
>> litetalk synthesizer on ttyS0?  I might be able to get slackware to run
>> the speakup.s kernel and if I can do that, I ought to be able to key the
>> modprobe line in once I get in as root.
>>
>>
>>
>> jude <jdashiel@shellworld.net>
>>
>> _______________________________________________
>> Speakup mailing list
>> Speakup@linux-speakup.org
>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>
>
> -- 
> ---
> John G. Heim, 608-263-4189, jheim@math.wisc.edu
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> 

Re: slackware on a very fast machine

[covici]

Its not a kernel bug, its a problem with speakup specifically that makes
serial ports not work unless you patch the speakup code and the patch is
not a fix, just a workaround.

Rob Hudson <captinlogic@gmail.com> wrote:

> I don't understand why the kernel has serial bugs. Many many many
> companies still use serial consoles, and in fact, as faras I know, it
> is still the most common way of interacting directly with rack-mounted
> computers and some high end routers. What's the deal?
> ----- Original Message ----- 
> From: "John G. Heim" <jheim@math.wisc.edu>
> To: "Speakup is a screen review system for Linux."
> <speakup@linux-speakup.org>
> Sent: Thursday, June 26, 2014 2:24 PM
> Subject: Re: slackware on a very fast machine
> 
> 
> > The syntax is "modprobe speakup_ltlk".
> >
> > But it isn't going to help unless  slackware hacked the kernel clode
> > to fix the serial port bug.  If they used the speakup code as it is
> > in the kernel, your litetalk won't work.
> > I started trying to make a fork of grml with a patched kernel but
> > never had the time to really get into it. Instead I've been using an
> > FAI disk. FAI is actually an installer system. FAI stands for fully
> > automated install. But they make it real easy to create a live CD.
> > On 06/26/14 14:12, Jude DaShiell wrote:
> >> I can't determine when the 10 seconds starts or ends with
> >> slackware-Current on this machine since the machine appears to be running
> >> too fast to get any usable audible indicators from it any longer.
> >> What is
> >> the correct syntax to modprobe speakup and have speakup come up on a
> >> litetalk synthesizer on ttyS0?  I might be able to get slackware to run
> >> the speakup.s kernel and if I can do that, I ought to be able to key the
> >> modprobe line in once I get in as root.
> >>
> >>
> >>
> >> jude <jdashiel@shellworld.net>
> >>
> >> _______________________________________________
> >> Speakup mailing list
> >> Speakup@linux-speakup.org
> >> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> >>
> >
> > -- 
> > ---
> > John G. Heim, 608-263-4189, jheim@math.wisc.edu
> > _______________________________________________
> > Speakup mailing list
> > Speakup@linux-speakup.org
> > http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> >
> 
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

-- 
Your life is like a penny.  You're going to lose it.  The question is:
How do
you spend it?

         John Covici
         covici@ccs.covici.com

Re: slackware on a very fast machine

[John G. Heim]

Well, technically, it is a kernel bug because speakup is part of the 
kernel. The bug is in the speakup part of the kernel code.

I wrote instructions on patching the kernel and put them on the web site 
of the International Association of visually Impaired Technologists, 
www.iavit.org. The instructions are for debian but the part about 
patching the kernel code should work for any distro. See:

http://www.iavit.org/~john/debian/build.html

On26/14 15:08, covici@ccs.covici.com wrote:
> Its not a kernel bug, its a problem with speakup specifically that makes
> serial ports not work unless you patch the speakup code and the patch is
> not a fix, just a workaround.
>
> Rob Hudson <captinlogic@gmail.com> wrote:
>
>> I don't understand why the kernel has serial bugs. Many many many
>> companies still use serial consoles, and in fact, as faras I know, it
>> is still the most common way of interacting directly with rack-mounted
>> computers and some high end routers. What's the deal?
>> ----- Original Message -----
>> From: "John G. Heim" <jheim@math.wisc.edu>
>> To: "Speakup is a screen review system for Linux."
>> <speakup@linux-speakup.org>
>> Sent: Thursday, June 26, 2014 2:24 PM
>> Subject: Re: slackware on a very fast machine
>>
>>
>>> The syntax is "modprobe speakup_ltlk".
>>>
>>> But it isn't going to help unless  slackware hacked the kernel clode
>>> to fix the serial port bug.  If they used the speakup code as it is
>>> in the kernel, your litetalk won't work.
>>> I started trying to make a fork of grml with a patched kernel but
>>> never had the time to really get into it. Instead I've been using an
>>> FAI disk. FAI is actually an installer system. FAI stands for fully
>>> automated install. But they make it real easy to create a live CD.
>>> On 06/26/14 14:12, Jude DaShiell wrote:
>>>> I can't determine when the 10 seconds starts or ends with
>>>> slackware-Current on this machine since the machine appears to be running
>>>> too fast to get any usable audible indicators from it any longer.
>>>> What is
>>>> the correct syntax to modprobe speakup and have speakup come up on a
>>>> litetalk synthesizer on ttyS0?  I might be able to get slackware to run
>>>> the speakup.s kernel and if I can do that, I ought to be able to key the
>>>> modprobe line in once I get in as root.
>>>>
>>>>
>>>>
>>>> jude <jdashiel@shellworld.net>
>>>>
>>>> _______________________________________________
>>>> Speakup mailing list
>>>> Speakup@linux-speakup.org
>>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>>
>>>
>>> --
>>> ---
>>> John G. Heim, 608-263-4189, jheim@math.wisc.edu
>>> _______________________________________________
>>> Speakup mailing list
>>> Speakup@linux-speakup.org
>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>
>>
>> _______________________________________________
>> Speakup mailing list
>> Speakup@linux-speakup.org
>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>

-- 
---
John G. Heim, 608-263-4189, jheim@math.wisc.edu

Re: slackware on a very fast machine

[John G. Heim]

The  serial consoles you are talking about run in what's called user 
space.  When you connect to a serial console, you're actually talking to 
a server process called getty or agetty. It's not that different from a 
web server or ssh server. The kernel doesn't know or care about those 
processes.  Accessing the  serial port in a kernel module is a totally 
different process.

Of course, you may be saying, "Ah, but I can set up a serial console in 
the kernel". And this is correct. But that console stops working once 
the kernel is booted and if you want to continue to use a console on 
that same serial port you have to set up a getty. If you have a terminal 
or terminal emulator listening on ttyS0, it may appear to be seamless 
but it's actually switching from kernel output to a getty.

So now maybe you are asking why can't speakup just access the serial 
port the same way as the kernel serial console? And that I don't know. I 
suspect  that code doesn't do things the "right way" either. The speakup 
code talks directly to the serial port at address 3F8. you're not 
supposed to do that. First it calls a function that is part of the 
regular kernel code that appears to  do nothing except return an error 
code.  Upon getting that error code, the speakup code errors out. All 
you have to do to make it work is comment out the part where the speakup 
code errors out. But my patch removes the call to the stupid/crazy 
function as well.

That function cannot return a success, so why call it? It's crazy.  It 
makes no sense what so ever. I've asked twice on the kernel mailing list 
what the speakup code should do instead and got nothing but static. All 
the answers I got either made no sense or they suggested rewriting the 
speakup code so that it would run in user space. But that would make it 
practically useless. The beauty of speakup is that you can get boot 
messages with it.

I was so angry at one point I considered contacting a lawyer. What I 
really should do is find out how the kernel serial console does it. I'll 
bet it reads/writes directly to 0x3F8 as well. In that case, the linux 
kernel developers are discriminating against speakup.

The problem doesn't really even effect me that much. I can easily just 
recompile the kernel. I've got my version of the debian wheezy kernel on 
about 200 amachines here in my department. But if I have to use grml, I 
can't use my litetalk.




On 06/26/14 15:02, Rob Hudson wrote:
> I don't understand why the kernel has serial bugs. Many many many
> companies still use serial consoles, and in fact, as faras I know, it is
> still the most common way of interacting directly with rack-mounted
> computers and some high end routers. What's the deal?
> ----- Original Message ----- From: "John G. Heim" <jheim@math.wisc.edu>
> To: "Speakup is a screen review system for Linux."
> <speakup@linux-speakup.org>
> Sent: Thursday, June 26, 2014 2:24 PM
> Subject: Re: slackware on a very fast machine
>
>
>> The syntax is "modprobe speakup_ltlk".
>>
>> But it isn't going to help unless  slackware hacked the kernel clode
>> to fix the serial port bug.  If they used the speakup code as it is in
>> the kernel, your litetalk won't work.
>> I started trying to make a fork of grml with a patched kernel but
>> never had the time to really get into it. Instead I've been using an
>> FAI disk. FAI is actually an installer system. FAI stands for fully
>> automated install. But they make it real easy to create a live CD.
>> On 06/26/14 14:12, Jude DaShiell wrote:
>>> I can't determine when the 10 seconds starts or ends with
>>> slackware-Current on this machine since the machine appears to be
>>> running
>>> too fast to get any usable audible indicators from it any longer.
>>> What is
>>> the correct syntax to modprobe speakup and have speakup come up on a
>>> litetalk synthesizer on ttyS0?  I might be able to get slackware to run
>>> the speakup.s kernel and if I can do that, I ought to be able to key the
>>> modprobe line in once I get in as root.
>>>
>>>
>>>
>>> jude <jdashiel@shellworld.net>
>>>
>>> _______________________________________________
>>> Speakup mailing list
>>> Speakup@linux-speakup.org
>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>
>>
>> --
>> ---
>> John G. Heim, 608-263-4189, jheim@math.wisc.edu
>> _______________________________________________
>> Speakup mailing list
>> Speakup@linux-speakup.org
>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>
>
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

-- 
---
John G. Heim, 608-263-4189, jheim@math.wisc.edu

Re: slackware on a very fast machine

[covici]

Well, the serial console code in the kernel does write directly to the
port, but it seems to be only working during the early boot -- I should
look at this again, this is how I wrote the first serial driver for
speakup -- coping the serial code, but at the time speakup was not part
of the kernel, in order to be part of the kernel, things had to change a
lot.

John G. Heim <jheim@math.wisc.edu> wrote:

> 
> 
> The  serial consoles you are talking about run in what's called user
> space.  When you connect to a serial console, you're actually talking
> to a server process called getty or agetty. It's not that different
> from a web server or ssh server. The kernel doesn't know or care about
> those processes.  Accessing the  serial port in a kernel module is a
> totally different process.
> 
> Of course, you may be saying, "Ah, but I can set up a serial console
> in the kernel". And this is correct. But that console stops working
> once the kernel is booted and if you want to continue to use a console
> on that same serial port you have to set up a getty. If you have a
> terminal or terminal emulator listening on ttyS0, it may appear to be
> seamless but it's actually switching from kernel output to a getty.
> 
> So now maybe you are asking why can't speakup just access the serial
> port the same way as the kernel serial console? And that I don't
> know. I suspect  that code doesn't do things the "right way"
> either. The speakup code talks directly to the serial port at address
> 3F8. you're not supposed to do that. First it calls a function that is
> part of the regular kernel code that appears to  do nothing except
> return an error code.  Upon getting that error code, the speakup code
> errors out. All you have to do to make it work is comment out the part
> where the speakup code errors out. But my patch removes the call to
> the stupid/crazy function as well.
> 
> That function cannot return a success, so why call it? It's crazy.  It
> makes no sense what so ever. I've asked twice on the kernel mailing
> list what the speakup code should do instead and got nothing but
> static. All the answers I got either made no sense or they suggested
> rewriting the speakup code so that it would run in user space. But
> that would make it practically useless. The beauty of speakup is that
> you can get boot messages with it.
> 
> I was so angry at one point I considered contacting a lawyer. What I
> really should do is find out how the kernel serial console does
> it. I'll bet it reads/writes directly to 0x3F8 as well. In that case,
> the linux kernel developers are discriminating against speakup.
> 
> The problem doesn't really even effect me that much. I can easily just
> recompile the kernel. I've got my version of the debian wheezy kernel
> on about 200 amachines here in my department. But if I have to use
> grml, I can't use my litetalk.
> 
> 
> 
> 
> On 06/26/14 15:02, Rob Hudson wrote:
> > I don't understand why the kernel has serial bugs. Many many many
> > companies still use serial consoles, and in fact, as faras I know, it is
> > still the most common way of interacting directly with rack-mounted
> > computers and some high end routers. What's the deal?
> > ----- Original Message ----- From: "John G. Heim" <jheim@math.wisc.edu>
> > To: "Speakup is a screen review system for Linux."
> > <speakup@linux-speakup.org>
> > Sent: Thursday, June 26, 2014 2:24 PM
> > Subject: Re: slackware on a very fast machine
> >
> >
> >> The syntax is "modprobe speakup_ltlk".
> >>
> >> But it isn't going to help unless  slackware hacked the kernel clode
> >> to fix the serial port bug.  If they used the speakup code as it is in
> >> the kernel, your litetalk won't work.
> >> I started trying to make a fork of grml with a patched kernel but
> >> never had the time to really get into it. Instead I've been using an
> >> FAI disk. FAI is actually an installer system. FAI stands for fully
> >> automated install. But they make it real easy to create a live CD.
> >> On 06/26/14 14:12, Jude DaShiell wrote:
> >>> I can't determine when the 10 seconds starts or ends with
> >>> slackware-Current on this machine since the machine appears to be
> >>> running
> >>> too fast to get any usable audible indicators from it any longer.
> >>> What is
> >>> the correct syntax to modprobe speakup and have speakup come up on a
> >>> litetalk synthesizer on ttyS0?  I might be able to get slackware to run
> >>> the speakup.s kernel and if I can do that, I ought to be able to key the
> >>> modprobe line in once I get in as root.
> >>>
> >>>
> >>>
> >>> jude <jdashiel@shellworld.net>
> >>>
> >>> _______________________________________________
> >>> Speakup mailing list
> >>> Speakup@linux-speakup.org
> >>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> >>>
> >>
> >> --
> >> ---
> >> John G. Heim, 608-263-4189, jheim@math.wisc.edu
> >> _______________________________________________
> >> Speakup mailing list
> >> Speakup@linux-speakup.org
> >> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> >>
> >
> > _______________________________________________
> > Speakup mailing list
> > Speakup@linux-speakup.org
> > http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
> 
> -- 
> ---
> John G. Heim, 608-263-4189, jheim@math.wisc.edu
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

-- 
Your life is like a penny.  You're going to lose it.  The question is:
How do
you spend it?

         John Covici
         covici@ccs.covici.com

Re: slackware on a very fast machine

[Trevor Astrope]

One side effect of commenting out the error check or the function that 
returns it is that you cannot unload the speakup module. Speakup will stop 
speaking, but the module won't be unloaded, so you can't reload it without 
rebooting.

It is still better than nothing, but it is far from an ideal solution.



On Thu, 26 Jun 2014, John G. Heim wrote:

>
>
> The  serial consoles you are talking about run in what's called user space. 
> When you connect to a serial console, you're actually talking to a server 
> process called getty or agetty. It's not that different from a web server or 
> ssh server. The kernel doesn't know or care about those processes.  Accessing 
> the  serial port in a kernel module is a totally different process.
>
> Of course, you may be saying, "Ah, but I can set up a serial console in the 
> kernel". And this is correct. But that console stops working once the kernel 
> is booted and if you want to continue to use a console on that same serial 
> port you have to set up a getty. If you have a terminal or terminal emulator 
> listening on ttyS0, it may appear to be seamless but it's actually switching 
> from kernel output to a getty.
>
> So now maybe you are asking why can't speakup just access the serial port the 
> same way as the kernel serial console? And that I don't know. I suspect  that 
> code doesn't do things the "right way" either. The speakup code talks 
> directly to the serial port at address 3F8. you're not supposed to do that. 
> First it calls a function that is part of the regular kernel code that 
> appears to  do nothing except return an error code.  Upon getting that error 
> code, the speakup code errors out. All you have to do to make it work is 
> comment out the part where the speakup code errors out. But my patch removes 
> the call to the stupid/crazy function as well.
>
> That function cannot return a success, so why call it? It's crazy.  It makes 
> no sense what so ever. I've asked twice on the kernel mailing list what the 
> speakup code should do instead and got nothing but static. All the answers I 
> got either made no sense or they suggested rewriting the speakup code so that 
> it would run in user space. But that would make it practically useless. The 
> beauty of speakup is that you can get boot messages with it.
>
> I was so angry at one point I considered contacting a lawyer. What I really 
> should do is find out how the kernel serial console does it. I'll bet it 
> reads/writes directly to 0x3F8 as well. In that case, the linux kernel 
> developers are discriminating against speakup.
>
> The problem doesn't really even effect me that much. I can easily just 
> recompile the kernel. I've got my version of the debian wheezy kernel on 
> about 200 amachines here in my department. But if I have to use grml, I can't 
> use my litetalk.
>
>
>
>
> On 06/26/14 15:02, Rob Hudson wrote:
>> I don't understand why the kernel has serial bugs. Many many many
>> companies still use serial consoles, and in fact, as faras I know, it is
>> still the most common way of interacting directly with rack-mounted
>> computers and some high end routers. What's the deal?
>> ----- Original Message ----- From: "John G. Heim" <jheim@math.wisc.edu>
>> To: "Speakup is a screen review system for Linux."
>> <speakup@linux-speakup.org>
>> Sent: Thursday, June 26, 2014 2:24 PM
>> Subject: Re: slackware on a very fast machine
>> 
>> 
>>> The syntax is "modprobe speakup_ltlk".
>>> 
>>> But it isn't going to help unless  slackware hacked the kernel clode
>>> to fix the serial port bug.  If they used the speakup code as it is in
>>> the kernel, your litetalk won't work.
>>> I started trying to make a fork of grml with a patched kernel but
>>> never had the time to really get into it. Instead I've been using an
>>> FAI disk. FAI is actually an installer system. FAI stands for fully
>>> automated install. But they make it real easy to create a live CD.
>>> On 06/26/14 14:12, Jude DaShiell wrote:
>>>> I can't determine when the 10 seconds starts or ends with
>>>> slackware-Current on this machine since the machine appears to be
>>>> running
>>>> too fast to get any usable audible indicators from it any longer.
>>>> What is
>>>> the correct syntax to modprobe speakup and have speakup come up on a
>>>> litetalk synthesizer on ttyS0?  I might be able to get slackware to run
>>>> the speakup.s kernel and if I can do that, I ought to be able to key the
>>>> modprobe line in once I get in as root.
>>>> 
>>>> 
>>>> 
>>>> jude <jdashiel@shellworld.net>
>>>> 
>>>> _______________________________________________
>>>> Speakup mailing list
>>>> Speakup@linux-speakup.org
>>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>>> 
>>> 
>>> --
>>> ---
>>> John G. Heim, 608-263-4189, jheim@math.wisc.edu
>>> _______________________________________________
>>> Speakup mailing list
>>> Speakup@linux-speakup.org
>>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>>> 
>> 
>> _______________________________________________
>> Speakup mailing list
>> Speakup@linux-speakup.org
>> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>
> -- 
> ---
> John G. Heim, 608-263-4189, jheim@math.wisc.edu
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup
>