Re: MSM DPMI error

Re: MSM DPMI error

[Stas Sergeev]

[-- Attachment #1: Type: text/plain, Size: 379 bytes --]

Hello.

Grigory Batalov wrote:
> I have full logs (-D9+M) if someone is interesting:
OK, AFAICS it have nothing to do with 1.1.3.4 - this
bug was always there and I wonder why nobody noticed
it before (oh, about a year ago Bart forwarded me
something similar from djgpp-devel, IIRC).
Try this patch. Without the ability to test, I can't
guarantee, but I think it must be a fix.


[-- Attachment #2: int_loop.diff --]
[-- Type: text/plain, Size: 526 bytes --]

--- src/dosext/dpmi/dpmi.c	Thu Sep 19 06:38:56 2002
+++ src/dosext/dpmi/dpmi.c	Thu Sep 19 16:45:35 2002
@@ -2202,7 +2202,11 @@
 		  case 0x1c:	/* ROM BIOS timer tick interrupt */
 		  case 0x23:	/* DOS Ctrl+C interrupt */
 		  case 0x24:	/* DOS critical error interrupt */
-			run_pm_int(VM86_ARG(retval));
+			if (SEGOFF2LINEAR(BIOSSEG, INT_OFF(VM86_ARG(retval))) !=
+ 			    SEGOFF2LINEAR(_CS, _IP) -2)
+			  run_pm_int(VM86_ARG(retval));
+			else
+			  do_int(VM86_ARG(retval));
 			break;
 		  default:
 #ifdef USE_MHPDBG

Re: MSM DPMI error

[Grigory Batalov]

On Thu, 19 Sep 2002 16:51:52 +0400
Stas Sergeev <stssppnn@yahoo.com> wrote:

> > I have full logs (-D9+M) if someone is interesting:

> OK, AFAICS it have nothing to do with 1.1.3.4 - this
> bug was always there and I wonder why nobody noticed
> it before (oh, about a year ago Bart forwarded me
> something similar from djgpp-devel, IIRC).
> Try this patch. Without the ability to test, I can't
> guarantee, but I think it must be a fix.

  Yes, it works now. Thank you, you saved my life! =))

  Stas, can you give some common hints to improve performance
 of such DPMI programs as MSM? It works quite slow (taking
 all CPU power).

-- 
 Grigory Batalov.

Re: MSM DPMI error

[Stas Sergeev]

Hello.

Grigory Batalov wrote:
> Stas, can you give some common hints to improve performance
> of such DPMI programs as MSM? It works quite slow (taking
> all CPU power).
Just recalled something: when I upgraded to kernel 2.4 from
2.2, I noticed a DPMI speed-up of 20 times!
IIRC you are still using 2.2. Then yes, I've got the idea.
Just upgrade to 2.4 and you'll see an unbeleavable speed-up
of your DPMI progs.

Re: MSM DPMI error

[Grigory Batalov]

On Fri, 27 Sep 2002 07:25:41 +0400
Stas Sergeev <stssppnn@yahoo.com> wrote:

> > Stas, can you give some common hints to improve performance
> > of such DPMI programs as MSM? It works quite slow (taking
> > all CPU power).
> Just recalled something: when I upgraded to kernel 2.4 from
> 2.2, I noticed a DPMI speed-up of 20 times!
> IIRC you are still using 2.2. Then yes, I've got the idea.
> Just upgrade to 2.4 and you'll see an unbeleavable speed-up
> of your DPMI progs.

  On 2.4.19 it stills the same. MSM takes 99% of CPU even without
  any user activity. And then draws it's screens _very_ slowly.

--
 Grigory Batalov.

Re: MSM DPMI error

[Johan Gill]

On Fri, 27 Sep 2002, Grigory Batalov wrote:

>   On 2.4.19 it stills the same. MSM takes 99% of CPU even without
>   any user activity. And then draws it's screens _very_ slowly.

Does this application access the CD-ROM or any other unit?

/Johan Gill, johane@lysator.liu.se

Re: MSM DPMI error

[Grigory Batalov]

On Fri, 27 Sep 2002 11:45:13 +0200 (MEST)
Johan Gill <johane@lysator.liu.se> wrote:

> >   On 2.4.19 it stills the same. MSM takes 99% of CPU even without
> >   any user activity. And then draws it's screens _very_ slowly.
> 
> Does this application access the CD-ROM or any other unit?

  Only IPX network.

--
 Grigory Batalov.

Re: MSM DPMI error

[Stas Sergeev]

Hello.

Grigory Batalov wrote:
> can you give some common hints to improve performance
> of such DPMI programs as MSM?
No unless I have this program handy, and even
then most likely still no, but you can try:)

Re: MSM DPMI error

[Grigory Batalov]

On Thu, 19 Sep 2002 17:53:04 +0400
Stas Sergeev <stssppnn@yahoo.com> wrote:

> > can you give some common hints to improve performance
> > of such DPMI programs as MSM?

> No unless I have this program handy, and even
> then most likely still no, but you can try:)

  I'm trying =).
  Do you want bzip-ed archive (>5Mb) ?

--
 Grigory Batalov.

Re: MSM DPMI error

[Stas Sergeev]

Hello.

Grigory Batalov wrote:
> I have a legacy DOS system, MSM (Micronetic M-System).
> It crashes under dosemu-1.1.3.4 with or without latest Stas' patches.
Is this because a lot of my patches are already
in, or not? I mean, with what dosemu version it
used to work? Now we have not more than one DPMI
patch per -pre, so it would be easy to find the
broken one knowing the version.

Re: MSM DPMI error

[Grigory Batalov]

On Thu, 19 Sep 2002 16:16:09 +0400
Stas Sergeev <stssppnn@yahoo.com> wrote:

> > I have a legacy DOS system, MSM (Micronetic M-System).
> > It crashes under dosemu-1.1.3.4 with or without latest Stas' patches.

> Is this because a lot of my patches are already
> in, or not? I mean, with what dosemu version it
> used to work? Now we have not more than one DPMI
> patch per -pre, so it would be easy to find the
> broken one knowing the version.

  I mean it doesn't work nor with, neither without your patches
  (all that I've found). Dosemu is latest (1.1.3.4).
  MSM didn't ever worked with any versions of dosemu, so
  I wonder if it would now (or why not).

--
 Grigory Batalov.

MSM DPMI error

[Grigory Batalov]

I have a legacy DOS system, MSM (Micronetic M-System).
It crashes under dosemu-1.1.3.4 with or without latest Stas' patches.
And it crashes also while starting as root.

Boot.log contains:
<skip>
ERROR: DPMI: DPMI_rm_procedure_running = 0xa8
<skip>
ERROR: DPMI: DPMI_rm_procedure_running = 0xa9
ERROR: Fault in dosemu code, in_dpmi=1
ERROR: Fault in dosemu code, in_dpmi=1
ERROR: cpu exception in dosemu code outside of DPMI client!
trapno: 0x0e  errorcode: 0x00000004  cr2: 0x02060000
eip: 0x40048355  esp: 0xbffff410  eflags: 0x00010296
cs: 0x0023  ds: 0x002b  es: 0x002b  ss: 0x002b
ERROR: cpu exception in dosemu code outside of DPMI client!
trapno: 0x0e  errorcode: 0x00000004  cr2: 0x02060000
eip: 0x40048355  esp: 0xbffff410  eflags: 0x00010296
cs: 0x0023  ds: 0x002b  es: 0x002b  ss: 0x002b
Page fault: Page fault: readread instruction to linear address: 0x02060000
 instruction to linear address: 0x02060000
CPU was in CPU was in user mode
user mode
Exception was caused by Exception was caused by non-available page
non-available page
  VFLAGS(b): 010000001010010110
EAX: 02060000  EBX: 400fce5c  ECX: 00000000  EDX: 00100000  VFLAGS(h): 00010296
ESI: 00000280  EDI: 08223440  EBP: bffff438
CS: 0023  DS: 002b  ES: 002b  FS: 0000  GS: 0000
FLAGS: PF AF SF IF RF  IOPL: 0
OOPS : 00 00 85 c0 74 09 83 ec 0c 57 -> ff 10 83 c4 10 8b 47 58 89 c6 
leavedos(4|0x4) called - shutting down
ERROR: signal 11 received in leavedos()
ERROR: signal 11 received in leavedos()

ERROR: timed/signalled out in leavedos()
ERROR: timed/signalled out in leavedos()

EMERGENCY JUMP 11!!!


I have full logs (-D9+M) if someone is interesting:

http://lrn.ru/~bga/dosemu/msm_drdos.log.bz2     (41741 bytes)
http://lrn.ru/~bga/dosemu/msm_freedos.log.bz2   (50147 bytes)
http://lrn.ru/~bga/dosemu/msm_msdos.log.bz2     (49433 bytes)

-- 
Grigory Batalov