dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Maurilio Longo]

Hi,

latest dosemu has problems (locks up here and there) if it is run on a 2.4GHz
pentium IV and inside dosemu.conf I have

$_rdtsc(on)
$_cpuspeed(0)

I think this comes from the fact that on startup it gives me a negative cpu
speed

Kernel CPU speed is -1894957296Hz

I'm running a protect mode dos clipper program. Same system on a 2.0GHz cpu
works flawlessly and it works ok as soon as I set rdtsc to off.

Also, there are problems with locks if I don't run it thruogh NETX on a novell
server and, instead, I run it on my "C:" disk (which is a directory on my
linux). I'll try to be more precise on the coming days.

regards.


--
 __________
|  |  | |__| md2520@mclink.it
|_|_|_|____| Team OS/2 Italia

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Maurilio Longo]

Any comments?

Nobody stumbled on this issue?

curious :)

Maurilio.


Maurilio Longo ha scritto:

> Hi,
>
> latest dosemu has problems (locks up here and there) if it is run on a 2.4GHz
> pentium IV and inside dosemu.conf I have
>
> $_rdtsc(on)
> $_cpuspeed(0)
>
> I think this comes from the fact that on startup it gives me a negative cpu
> speed
>
> Kernel CPU speed is -1894957296Hz
>
> I'm running a protect mode dos clipper program. Same system on a 2.0GHz cpu
> works flawlessly and it works ok as soon as I set rdtsc to off.
>
> Also, there are problems with locks if I don't run it thruogh NETX on a novell
> server and, instead, I run it on my "C:" disk (which is a directory on my
> linux). I'll try to be more precise on the coming days.
>
> regards.
>
> --
>  __________
> |  |  | |__| md2520@mclink.it
> |_|_|_|____| Team OS/2 Italia
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
 __________
|  |  | |__| md2520@mclink.it
|_|_|_|____| Team OS/2 Italia

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Ryan Underwood]

> Any comments?
> 
> Nobody stumbled on this issue?
> 
> curious :)

I'd wager that nobody on this list owns a machine that fast. 8)

> > latest dosemu has problems (locks up here and there) if it is run on a 2.4GHz
> > pentium IV and inside dosemu.conf I have



-- 
Ryan Underwood, <nemesis at icequake.net>, icq=10317253

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Maurilio Longo]

I fear that cpu speed is inside a long int and this shoud explain why it happens, I'd

like to know from someone who writes dosemu if this is true and how they plan to fix
this.

regards.

Ryan Underwood ha scritto:

> > Any comments?
> >
> > Nobody stumbled on this issue?
> >
> > curious :)
>
> I'd wager that nobody on this list owns a machine that fast. 8)
>
> > > latest dosemu has problems (locks up here and there) if it is run on a 2.4GHz
> > > pentium IV and inside dosemu.conf I have
>
> --
> Ryan Underwood, <nemesis at icequake.net>, icq=10317253
> -
> To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
 __________
|  |  | |__| md2520@mclink.it
|_|_|_|____| Team OS/2 Italia

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Bart Oldeman]

On Wed, 12 Mar 2003, Maurilio Longo wrote:

> I fear that cpu speed is inside a long int and this shoud explain why it happens, I'd
>
> like to know from someone who writes dosemu if this is true and how they plan to fix
> this.

it's a multiplication that overflows from an int -- try this patch:

--- dosemu-1.1.4.13/src/base/init/config.c	Sat Feb 15 14:49:31 2003
+++ dosemu-1.1.4.14/src/base/init/config.c	Wed Mar 12 14:38:28 2003
@@ -484,7 +484,7 @@
 		cdd[6]=0; sscanf(cdd,"%d",&df);
 		/* speed division factor to get 1us from CPU clocks - for
 		 * details on fast division see timers.h */
-		chz = (di * 1000000) + df;
+		chz = (di * 1000000LL) + df;

 		/* speed division factor to get 1us from CPU clock */
 		config.cpu_spd = (LLF_US*1000000)/chz;

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Maurilio Longo]

Ok, with this patch it works ok :)

So I've switched to 1.1.4.13, but now my clipper program is not able to print to a serial
printer anymore :( even if a

dir > com1:

from dos prompt works ok, but my program uses a third party library to communicate with
serial devices

I fear it has to do with latest changes to serial code... I'll try to dig a little more,
but I'm not familiar with dosemu code and linux programming in general...

regards.



Bart Oldeman ha scritto:

> On Wed, 12 Mar 2003, Maurilio Longo wrote:
>
> > I fear that cpu speed is inside a long int and this shoud explain why it happens, I'd
> >
> > like to know from someone who writes dosemu if this is true and how they plan to fix
> > this.
>
> it's a multiplication that overflows from an int -- try this patch:
>
> --- dosemu-1.1.4.13/src/base/init/config.c      Sat Feb 15 14:49:31 2003
> +++ dosemu-1.1.4.14/src/base/init/config.c      Wed Mar 12 14:38:28 2003
> @@ -484,7 +484,7 @@
>                 cdd[6]=0; sscanf(cdd,"%d",&df);
>                 /* speed division factor to get 1us from CPU clocks - for
>                  * details on fast division see timers.h */
> -               chz = (di * 1000000) + df;
> +               chz = (di * 1000000LL) + df;
>
>                 /* speed division factor to get 1us from CPU clock */
>                 config.cpu_spd = (LLF_US*1000000)/chz;

--
 __________
|  |  | |__| md2520@mclink.it
|_|_|_|____| Team OS/2 Italia

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Norman Schmidt Jr]

Maurilio, were you using freedos with dosemu before you switched to 
1.1.4.13 and your clipper apps were printing? In freedos? I never could 
put clipper printing (neither serial nor parallel) to work with 
freedos+dosemu, but I started to use dosemu with 1.1.4, and never tried 
any previous versions. Did the old versions of dosemu print from clipper 
apps using freedos? The only way I could put clipper apps to print using 
dosemu was using ms-dos or dr-dos...
Norman

Maurilio Longo escreveu:

>Ok, with this patch it works ok :)
>
>So I've switched to 1.1.4.13, but now my clipper program is not able to print to a serial
>printer anymore :( even if a
>
>dir > com1:
>
>from dos prompt works ok, but my program uses a third party library to communicate with
>serial devices
>
>I fear it has to do with latest changes to serial code... I'll try to dig a little more,
>but I'm not familiar with dosemu code and linux programming in general...
>
>regards.
>
>
>
>Bart Oldeman ha scritto:
>
>  
>
>>On Wed, 12 Mar 2003, Maurilio Longo wrote:
>>
>>    
>>
>>>I fear that cpu speed is inside a long int and this shoud explain why it happens, I'd
>>>
>>>like to know from someone who writes dosemu if this is true and how they plan to fix
>>>this.
>>>      
>>>
>>it's a multiplication that overflows from an int -- try this patch:
>>
>>--- dosemu-1.1.4.13/src/base/init/config.c      Sat Feb 15 14:49:31 2003
>>+++ dosemu-1.1.4.14/src/base/init/config.c      Wed Mar 12 14:38:28 2003
>>@@ -484,7 +484,7 @@
>>                cdd[6]=0; sscanf(cdd,"%d",&df);
>>                /* speed division factor to get 1us from CPU clocks - for
>>                 * details on fast division see timers.h */
>>-               chz = (di * 1000000) + df;
>>+               chz = (di * 1000000LL) + df;
>>
>>                /* speed division factor to get 1us from CPU clock */
>>                config.cpu_spd = (LLF_US*1000000)/chz;
>>    
>>
>
>--
> __________
>|  |  | |__| md2520@mclink.it
>|_|_|_|____| Team OS/2 Italia
>
>
>-
>To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>
>  
>

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Maurilio Longo]

No, I use an old MS-DOS 5.x, now, even with latest patch from Bart, I can print to my serial
printer with a

dir > com1:

but not from my clipper (protect mode, telepathy as comx: access library) app...

To be more precise I can print very short strings, that is

tp_open(1...)  // open com port 1
tp_send(1, "hi there")  // send string
tp_close(1)

works, but if there are more tp_send() one after the other I get no output!?!!

1.1.1.x was working nicely

:(

regards.

Maurilio.




Regards.

Norman Schmidt Jr ha scritto:

> Maurilio, were you using freedos with dosemu before you switched to
> 1.1.4.13 and your clipper apps were printing? In freedos? I never could
> put clipper printing (neither serial nor parallel) to work with
> freedos+dosemu, but I started to use dosemu with 1.1.4, and never tried
> any previous versions. Did the old versions of dosemu print from clipper
> apps using freedos? The only way I could put clipper apps to print using
> dosemu was using ms-dos or dr-dos...
> Norman
>
> Maurilio Longo escreveu:
>
> >Ok, with this patch it works ok :)
> >
> >So I've switched to 1.1.4.13, but now my clipper program is not able to print to a serial
> >printer anymore :( even if a
> >
> >dir > com1:
> >
> >from dos prompt works ok, but my program uses a third party library to communicate with
> >serial devices
> >
> >I fear it has to do with latest changes to serial code... I'll try to dig a little more,
> >but I'm not familiar with dosemu code and linux programming in general...
> >
> >regards.
> >
> >
> >
> >Bart Oldeman ha scritto:
> >
> >
> >
> >>On Wed, 12 Mar 2003, Maurilio Longo wrote:
> >>
> >>
> >>
> >>>I fear that cpu speed is inside a long int and this shoud explain why it happens, I'd
> >>>
> >>>like to know from someone who writes dosemu if this is true and how they plan to fix
> >>>this.
> >>>
> >>>
> >>it's a multiplication that overflows from an int -- try this patch:
> >>
> >>--- dosemu-1.1.4.13/src/base/init/config.c      Sat Feb 15 14:49:31 2003
> >>+++ dosemu-1.1.4.14/src/base/init/config.c      Wed Mar 12 14:38:28 2003
> >>@@ -484,7 +484,7 @@
> >>                cdd[6]=0; sscanf(cdd,"%d",&df);
> >>                /* speed division factor to get 1us from CPU clocks - for
> >>                 * details on fast division see timers.h */
> >>-               chz = (di * 1000000) + df;
> >>+               chz = (di * 1000000LL) + df;
> >>
> >>                /* speed division factor to get 1us from CPU clock */
> >>                config.cpu_spd = (LLF_US*1000000)/chz;
> >>
> >>
> >
> >--
> > __________
> >|  |  | |__| md2520@mclink.it
> >|_|_|_|____| Team OS/2 Italia
> >
> >
> >-
> >To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
> >the body of a message to majordomo@vger.kernel.org
> >More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >
> >
> >
> >
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
 __________
|  |  | |__| md2520@mclink.it
|_|_|_|____| Team OS/2 Italia

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Maurilio Longo]

That is,

int14 is working ok, but serial emulation is not (telepathy does a direct serial access and does
not depend on int14 service).

regards

Maurilio Longo ha scritto:

> No, I use an old MS-DOS 5.x, now, even with latest patch from Bart, I can print to my serial
> printer with a
>
> dir > com1:
>
> but not from my clipper (protect mode, telepathy as comx: access library) app...
>
> To be more precise I can print very short strings, that is
>
> tp_open(1...)  // open com port 1
> tp_send(1, "hi there")  // send string
> tp_close(1)
>
> works, but if there are more tp_send() one after the other I get no output!?!!
>
> 1.1.1.x was working nicely
>
> :(
>
> regards.
>
> Maurilio.
>
> Regards.
>
> Norman Schmidt Jr ha scritto:
>
> > Maurilio, were you using freedos with dosemu before you switched to
> > 1.1.4.13 and your clipper apps were printing? In freedos? I never could
> > put clipper printing (neither serial nor parallel) to work with
> > freedos+dosemu, but I started to use dosemu with 1.1.4, and never tried
> > any previous versions. Did the old versions of dosemu print from clipper
> > apps using freedos? The only way I could put clipper apps to print using
> > dosemu was using ms-dos or dr-dos...
> > Norman
> >
> > Maurilio Longo escreveu:
> >
> > >Ok, with this patch it works ok :)
> > >
> > >So I've switched to 1.1.4.13, but now my clipper program is not able to print to a serial
> > >printer anymore :( even if a
> > >
> > >dir > com1:
> > >
> > >from dos prompt works ok, but my program uses a third party library to communicate with
> > >serial devices
> > >
> > >I fear it has to do with latest changes to serial code... I'll try to dig a little more,
> > >but I'm not familiar with dosemu code and linux programming in general...
> > >
> > >regards.
> > >
> > >
> > >
> > >Bart Oldeman ha scritto:
> > >
> > >
> > >
> > >>On Wed, 12 Mar 2003, Maurilio Longo wrote:
> > >>
> > >>
> > >>
> > >>>I fear that cpu speed is inside a long int and this shoud explain why it happens, I'd
> > >>>
> > >>>like to know from someone who writes dosemu if this is true and how they plan to fix
> > >>>this.
> > >>>
> > >>>
> > >>it's a multiplication that overflows from an int -- try this patch:
> > >>
> > >>--- dosemu-1.1.4.13/src/base/init/config.c      Sat Feb 15 14:49:31 2003
> > >>+++ dosemu-1.1.4.14/src/base/init/config.c      Wed Mar 12 14:38:28 2003
> > >>@@ -484,7 +484,7 @@
> > >>                cdd[6]=0; sscanf(cdd,"%d",&df);
> > >>                /* speed division factor to get 1us from CPU clocks - for
> > >>                 * details on fast division see timers.h */
> > >>-               chz = (di * 1000000) + df;
> > >>+               chz = (di * 1000000LL) + df;
> > >>
> > >>                /* speed division factor to get 1us from CPU clock */
> > >>                config.cpu_spd = (LLF_US*1000000)/chz;
> > >>
> > >>
> > >
> > >--
> > > __________
> > >|  |  | |__| md2520@mclink.it
> > >|_|_|_|____| Team OS/2 Italia
> > >
> > >
> > >-
> > >To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
> > >the body of a message to majordomo@vger.kernel.org
> > >More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > >
> > >
> > >
> > >
> >
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
> --
>  __________
> |  |  | |__| md2520@mclink.it
> |_|_|_|____| Team OS/2 Italia
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-msdos" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
 __________
|  |  | |__| md2520@mclink.it
|_|_|_|____| Team OS/2 Italia

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Stas Sergeev]

[-- Attachment #1: Type: text/plain, Size: 166 bytes --]

Hello.

Bart Oldeman wrote:
> it's a multiplication that overflows from an int -- try this patch:
The attached one might also be
necessary to get the correct
output.

[-- Attachment #2: cpuspd.diff --]
[-- Type: text/plain, Size: 507 bytes --]

--- src/base/init/config.c	Wed Feb 19 16:17:30 2003
+++ src/base/init/config.c	Wed Mar 12 22:09:27 2003
@@ -493,7 +493,7 @@
 		/* speed division factor to get 838ns from CPU clock */
 		config.cpu_tick_spd = (LLF_TICKS*1000000)/chz;
 
-		warn ("Linux kernel %d.%d.%d; CPU speed is %Ld Hz\n",
+		warn ("Linux kernel %d.%d.%d; CPU speed is %lld Hz\n",
 		   kernel_version_code >> 16, (kernel_version_code >> 8) & 255,
 		   kernel_version_code & 255,chz);
 /*		fprintf (stderr,"CPU speed factors %ld,%ld\n",

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Bart Oldeman]

On Wed, 12 Mar 2003, Stas Sergeev wrote:

> Bart Oldeman wrote:
> > it's a multiplication that overflows from an int -- try this patch:
> The attached one might also be
> necessary to get the correct
> output.

-               warn ("Linux kernel %d.%d.%d; CPU speed is %Ld Hz\n",
+               warn ("Linux kernel %d.%d.%d; CPU speed is %lld Hz\n",

This is good for C99 compliance (C99 only guarantees "ll" but doesn't do
anything from the practical point of view) -- "info libc"

`L'
`ll'
`q'
     Specifies that the argument is a `long long int'.  (This type is
     an extension supported by the GNU C compiler.  On systems that
     don't support extra-long integers, this is the same as `long int'.)

     The `q' modifier is another name for the same thing, which comes
     from 4.4 BSD; a `long long int' is sometimes called a "quad" `int'.

Still it's good to use a standard approach whereever possible, so thanks,
Bart

Re: dosemu 1.1.x and 1.1.4.13 bug (?) on CPUs faster than 2.0GHz

[Stas Sergeev]

Hello.

Maurilio Longo wrote:
> int14 is working ok, but serial emulation is not (telepathy does a 
Unless you provide all the necessary
logs, you'll of course not get any
concrete suggestions.

>> works, but if there are more tp_send() one after the other I get no 
>  output!?!!
Great, in that case you can provide
2 logs: when it works and when it
doesn't. That can help.

>> 1.1.1.x was working nicely
Even better, supply the 1.1.1 log then
as well.
You'll have to enable -D9+s logging.
Changes in the Serial code were many
so guessing is not possible.