Testers wanted for boot monitor change

[ben gras]

Dear all,

There seems to be a bug in the current boot monitor code that will cause booting to fail entirely on a small number of machines. (The symptom on a new machine I tried was: freeze as soon as the CD is read.) That is fixed now, but I'd like to ask for people to test the new version before I commit it, as it's such a machine/bios-dependent change, and so fundamental to the system (detecting the ranges of available physical memory), it's conceivable it'll introduce problems where there weren't before.

The iso with the fix is: http://www.few.vu.nl/~beng/bootfix.iso.bz2

My request is that you try it on machines (VMs and hardware) you have available. I'd like to know whether (a) it boots up ok, up until the login prompt, especially if it didn't before, or (heavens forbid) doesn't now but did before. And (b) whether the detected memory ranges are any different ('better,' or 'worse,' than before). I know they were different on my test machines, but this doesn't seem to cause problems. (The iso has vm JUNKFREE turned on, so all physical ranges are touched at boot time, and after freeing, which hopefully detects cases in which physical ranges are detected that aren't there.) You can tell by looking at the 'memory' variable in the 'sysenv' output.

In short, see if it works for you :-). Especially mail me if it breaks minix on a machine on which it wasn't broken before.

Hopefully, this only fixes problems and doesn't introduce new ones.

Thanks everyone!

[Tom Chandler]

Ben,

I tested the boot.iso on my P2-350 512M system in native mode.

This machine in native mode as run 3.1.5,6,7,8, but when trying to install
3.2.0, got the boot error message "not enough memory" (or something along those
lines) and halted.

boot.iso does boot up.  I was able to run setup and install the system to the
hard drive.  Then I rebooted the system and it did come up.

Note 1:  When using pkgin, on the downloads the per cent downloaded stayed at
zero, until complete then went to 100%.  Strange.

Note 2:  Trying to compile programs with gcc, I am getting some strange activity,
will try and research with definitive comments tomorrow.

At least this version (boot.iso) does allow my native machine to boot, and I was
able to run setup and install it.  I used pkgin to download packages to do further
testing.  (See Note 2)

Hope this helps.

Thank You
Tom Chandler

--
You received this message because you are subscribed to the Google Groups "minix3" group.
To view this discussion on the web visit https://groups.google.com/d/msg/minix3/-/Vs-EsFw2o4AJ.
To post to this group, send email to min...@googlegroups.com.
To unsubscribe from this group, send email to minix3+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/minix3?hl=en.

[Tom Chandler]

Ben,

2nd update:  Last week I sent a message about 3.2.0, compiling giving me
a VM: error about not being able to find memory for the compile and then halt.
(It is in the mailing list archives)

This boot.iso seems to have solved that problem, as now I can compile the
programs that were failing.

However, I have noticed the following using GCC.  In the make file I compile
several programs, then create a library from them.  If I run the makefile back
to back, I would expect the GCC to tell me the libraries are up to date and
continue on, however it does not.  It remakes the libraries everytime.  This
did not occur on 3.1.8, and earlier.  Not sure if bug, but it is not what I expect
for it to do.

Otherwise, boot.iso is looking good.

Thank You
Tom c

[Tom Chandler]

Update on GCC issue.

This morning when I powered up the system and started testing,
GCC is working as I expected, So please ignore message about GCC.

Mark it as user interface problems!!!!!

Great Work.

Thank You
Tom C

[Jean-Pierre Coetzee]

Works perfectly on my Aopen MX33 motherboard with 128mb of ram, it
didn't work before. Thanks.

[ben gras]

That's really great to hear! Thanks for testing everyone. I've committed the change, as I've only read instances in which it was an improvement, and not an umprovement.

[Tom Chandler]

Great work.  I has been running here on a native machine with problems.

I also track "trunk", so based on your comments, these changes have been
committed?

Thank You
Tom C

On Tue, Sep 20, 2011 at 11:07 AM, ben gras <b...@minix3.org> wrote:

That's really great to hear! Thanks for testing everyone. I've committed the change, as I've only read instances in which it was an improvement, and not an umprovement.

--

You received this message because you are subscribed to the Google Groups "minix3" group.

To view this discussion on the web visit https://groups.google.com/d/msg/minix3/-/cgyWv_diNr8J.

[ben gras]

These changes are indeed committed (currently only 3rd-down, commit 1dfd43ac27873441d3c7abe6867acfcf2aa52cce), but the boot monitor changes won't automatically install a new boot monitor with just make world. If you want to do that, recompile it, and run updateboot.sh in boot/, then reboot the whole machine (or 'boot d0' in the boot monitor) to re-load it from disk.

Then again, if your minix machine is already working, it's not likely this change will improve it much. Although if it does i'd be very interested to hear about it. (You could check the memory ranges before and after the change, by looking at the difference in the 'memory' variable printed by 'sysenv'.)

[Arun Thomas]

On Tue, Sep 20, 2011 at 6:22 PM, ben gras <b...@minix3.org> wrote:
> These changes are indeed committed (currently only 3rd-down, commit
> 1dfd43ac27873441d3c7abe6867acfcf2aa52cce), but the boot monitor changes
> won't automatically install a new boot monitor with just make world. If you
> want to do that, recompile it, and run updateboot.sh in boot/, then reboot
> the whole machine (or 'boot d0' in the boot monitor) to re-load it from
> disk.

Alternatively, you could install the new 3.2.0 snapshot ISO. It
includes the boot monitor changes.

Best,
Arun

[Roman Ignatov]

I have not problem. Successfully works on native machines:

1. Notebook Lenovo G575 (dual-core 1.60GHz AMD E-350 CPU ,3GB of RAM,
320GB hard drive, a CD/DVD burner, 802.11n Wi-Fi, and integrated AMD
Radeon-HD 6310M graphics);
2. Computer: Core 2 duo 2.66, 2GB of RAM, Intel G33/G31 Express
Chipset Family.

[ben gras]

Thanks for testing Roman!

[Roman Ignatov]

Successfully works on native machine:

Processor: Amd Athlon 64 3000
RAM: 512Mb
HDD: 20GB
Video: ATI Radeon HD 3450
Ethernet Cards: Realtek 8139

[GR8DAN]

Worked on VirtualBox 4.1.2r73507 (latest version) on Windows Vista
Home Premium on a Dell quad core machine with VMM support turn on.
Virtual Machine configured with single CPU, 512MB ram, 2GB disk.

DAN

[Roman Ignatov]

Worked on: Celeron 733 i815 384MB RAM GeForce FX5200.

[r0ller]

I guess it does not add much but noone has posted it yet, so: I'm
using the new image without any problems on vmware 3.1.4.

r0ller