• Welcome to OS2World OLD-STATIC-BACKUP Forum.
 

News:

This is an old OS2World backup forum for reference only. IT IS READ ONLY!!!

If you need help with OS/2 - eComStation visit http://www.os2world.com/forum

Main Menu

SMP vs. DHCP

Started by djcaetano, 2008.07.28, 17:22:50

Previous topic - Next topic

djcaetano


  My Core2 Quad eCS2 RC4 system has been suffering with a strange bug for some
time now. Up to ACPI 3.09, running on SMP model, the system froze sometimes during
the  execution of DHCP deamon on boot, but most times it didn't.
  I updated ACPI to version 3.10 and the boot was ok, I used the system for one day
without any problems. This was last friday and today, monday, I was not able to
boot in SMP mode anymore: on every single boot the system was frozen by the
time it executes DHCP daemon, without any kind of error messages.
   I had to boot to command prompt, remove the SMP SMP switch from ACPI.PSD
and now the system complete the boot without any kind of problems (but it is
finding only one processor). The APIC switch is still on.

  Anyone knows if DHCP has some know glitch or problem with SMP mode? Any
clues how can I solve this "little" problem, besides disabling SMP? :P

  Thanks in advance!

Andi

Which NIC and which driver? Do you know how the interrupts were/are assigned witch 3.09/3.10?

El Vato

Quote from: djcaetano on 2008.07.28, 17:22:50

  My Core2 Quad eCS2 RC4 system has been suffering with a strange bug for some
time now. Up to ACPI 3.09, running on SMP model, the system froze sometimes during
the  execution of DHCP deamon on boot,[...]

  Anyone knows if DHCP has some know glitch or problem with SMP mode? Any
clues how can I solve this "little" problem, besides disabling SMP? :P

[...]

No DHCP problem with dual Intel CPUs using WSEB on my side, Caetano.

Suggestion: disable the DHCP daemon (cihuateteo) during the boot phase.  You may enable it after your OS/2 operating system (with SMP support enabled) has started by opening an OS/2 windowed shell and typing:

START /I  DHCPCD

Another window will open where the daemon (cihuateteo) is executing and within a few seconds you should have been assigned an dynamic IP address by your DHCP server ( I assume that you are the client, evidently).  Follow that by minimizing the window.

This kind of problems are expected when enhancements to the OS/2 are kept closed sourced, by the way, since only a few are allowed to make modifications and/or provide suggestions for a more robust and thorough applicability domain of those enhancements.

Robert Deed

Stop grandstanding on the opensource closed source argument.  Many OS/2 applications which are open source suffer from severe bugs and crashes as well.   Most likely the problem is not being caused by DHCP anyway, it is most likely the network drivers.

Quote from: El Vato on 2008.07.29, 04:17:58
Quote from: djcaetano on 2008.07.28, 17:22:50

  My Core2 Quad eCS2 RC4 system has been suffering with a strange bug for some
time now. Up to ACPI 3.09, running on SMP model, the system froze sometimes during
the  execution of DHCP deamon on boot,[...]

  Anyone knows if DHCP has some know glitch or problem with SMP mode? Any
clues how can I solve this "little" problem, besides disabling SMP? :P

[...]

No DHCP problem with dual Intel CPUs using WSEB on my side, Caetano.

Suggestion: disable the DHCP daemon (cihuateteo) during the boot phase.  You may enable it after your OS/2 operating system (with SMP support enabled) has started by opening an OS/2 windowed shell and typing:

START /I  DHCPCD

Another window will open where the daemon (cihuateteo) is executing and within a few seconds you should have been assigned an dynamic IP address by your DHCP server ( I assume that you are the client, evidently).  Follow that by minimizing the window.

This kind of problems are expected when enhancements to the OS/2 are kept closed sourced, by the way, since only a few are allowed to make modifications and/or provide suggestions for a more robust and thorough applicability domain of those enhancements.

djcaetano


  Hi,

  Andi: I don't know any differences in IRQ assignment from 3.09 to 3.10. But the NIC is a RTL8139 running with proper driver (not GenMac, which didn't work with the onboard network interface, despite the fact the readme says it's "supported but not tested").
  Anyway, since the system on non-SMP mode was trapping when I attached the USB pendrive, I tried to return to previous 3.09 ACPI. Then... the system gave me a TRAP 000e on ACPI during boot.
  I then REMarked the ACPI.PSD line on CONFIG.SYS, rebooted, and reinstalled ACPI 3.10 (and unREMarked the ACPI.PSD line). Well, this time the system boot clean and the problems with my USB pendrive vanished.

  It has been a long time since computers behavior could be considered to be "exact". :(

El Vato

Quote from: djcaetano on 2008.07.28, 17:22:50
  Anyone knows if DHCP has some know glitch or problem with SMP mode? Any
clues how can I solve this "little" problem, besides disabling SMP? :P

[...]

Quote from: Robert Deed on 2008.07.29, 16:47:12
Stop grandstanding on the opensource closed source argument.
Whoa~! 
...Speaking of hirelings the ostrich pulls its head out!  Where is your specific contribution to solve the problem of reference ???

When I use a resource that belongs to you or to those sacred cows who according to you, continue developing OS/2 as a "labor of love", I will stop promoting the virtues of open source as the most viable resource and development paradigm for OS/2.  Notwithstanding, as long as there is a neutral medium for debating the issue I could not care less about your babbling nonsense.

Quote from: Robert Deed on 2008.07.29, 16:47:12
Many OS/2 applications which are open source suffer from severe bugs and crashes as well.
...specifics to the problem of reference !!!??? --just as I specified suspicion for the closed source ACPI development by your "benevolent" sacred cow X or Y.

Quote from: Robert Deed on 2008.07.29, 16:47:12
   Most likely the problem is not being caused by DHCP anyway, it is most likely the network drivers.

Based on what analysis or troubleshooting performed and on what performance domain ???  Or are you simply covering up for the inefficiency and ineptitude of those that you defend with a truly grandstanding pusillanimous1 statement claiming that either sacred cows X or Y "...are not really making any money on the OS/2 but they are doing it as a labor of love."  Oh, who are you trying to fool ???

1Whatever connotation to the nature of this fellows is merely a coincidence.

pusillanimous \pyoo-suh-LAN-uh-muhs\, adjective:
Lacking in courage and resolution; contemptibly fearful; cowardly.