If you have a Philips MCD708 Micro Theatre stuck in demo mode, you can turn off demo mode by setting the unit to standby mode (pressing STANDBY-ON on the unit or remote) then pressing "PRESET -" in the unit. While the unit is in standby, pressing "PRESET -" toggles it between DEMO ON and DEMO OFF.
You can turn demo mode off from the remote also, but only if TUNER is selected, not if DVD or AUX are selected. Turn the system on, select TUNER then turn the unit to STANDBY, then press the "CH -" button on the remote.
Because Philips left this information out of the manual and has nothing about this issue on-line where Google can find it, I wasted several hours pressing buttons until I happened upon the magic combination. If I could, I would return this poorly documented product and get my money back to buy something else. I will be reluctant to buy anything else from Philips.
Friday, June 26, 2009
Wednesday, June 24, 2009
Linux Compatible Firmware
Many hardware vendors produce excellent hardware but do not make sufficient technical information about them available to allow their hardware to be used with Linux systems. This includes BIOS and other firmware on main boards and core chip sets.
Until 2007 a consortium of vendors was working to improve this situation by producing a Linux-ready Firmware Developer Kit, according to an INTEL employee (see the comment by arjan).
What has happened since 2007?
There is a bit of an update from 2008 here.
And there is this comparing LinuxBIOS with EFI.
The UEFI standard defines a GUID Partition Table. This is proposed and somewhat used as a replacement for the traditional partition table.
The traditional partition table is not well defined, it seems. There is some consistency between systems but also much inconsistency. Software interacting with partition tables includes BIOS and operating systems. Andries Brouwer has written a description of it. And here's another from ata-atapi.com.
Until 2007 a consortium of vendors was working to improve this situation by producing a Linux-ready Firmware Developer Kit, according to an INTEL employee (see the comment by arjan).
What has happened since 2007?
There is a bit of an update from 2008 here.
And there is this comparing LinuxBIOS with EFI.
The UEFI standard defines a GUID Partition Table. This is proposed and somewhat used as a replacement for the traditional partition table.
The traditional partition table is not well defined, it seems. There is some consistency between systems but also much inconsistency. Software interacting with partition tables includes BIOS and operating systems. Andries Brouwer has written a description of it. And here's another from ata-atapi.com.
Trusted Computing
I read this article on Trusted Computing over the past few days. It is from 2003 - six years ago.
What has happened since then?
It is hard to imagine that Trusted Computing will actually work. The worst case is that it does not work and it is bolstered by legal requirements to use it and prohibitions against anything that reduces its effectiveness. This would include prohibitions against investigations into how it works, its strengths and weaknesses and any alternatives such as already exist in the DMCA legislation in the US. This would also include prohibition of anything that could be used to investigate how Trusted Computing works and any communication of information about how it works. Such legislation has been proposed in other cases.
The emperor may have no clothes, but that will be OK because it will be illegal to look at his clothes or communicate anything other than the official statements about his clothes.
What has happened since then?
It is hard to imagine that Trusted Computing will actually work. The worst case is that it does not work and it is bolstered by legal requirements to use it and prohibitions against anything that reduces its effectiveness. This would include prohibitions against investigations into how it works, its strengths and weaknesses and any alternatives such as already exist in the DMCA legislation in the US. This would also include prohibition of anything that could be used to investigate how Trusted Computing works and any communication of information about how it works. Such legislation has been proposed in other cases.
The emperor may have no clothes, but that will be OK because it will be illegal to look at his clothes or communicate anything other than the official statements about his clothes.
Sunday, June 21, 2009
Linux commands I read about today
The losetup command is for managing loop devices. The loop device driver accesses regular files instead of physical devices. But, while they are block devices, you can't partition them like real block devices (i.e. disks). You can run fdisk but will have to tell it the number of cylinders, after which it will create a partition table for you, but there will be no devices created to access those partitions and attempt to reload the partition table (as fdisk does after writing it or with blockdev --rereadpt) fails with Invalid argument. This email may be relevant and this post has some further examples.
The blockdev command is for calling select ioctl commands on block devices from the command line.
The tc command can be used to add latency to loop devices (and lots else):
tc qdisc add dev lo root handle 1:0 netem delay 20msec
tc qdisc del dev lo root
the loop device can be used to access partitions in disk images.
Note that the NASA enhanced loop device is still available from ftp://ftp.hq.nasa.gov/pub/ig/ccd/enhanced_loopback/ but it is patches against a 2.4 kernel and has not been updated since 2004.
The blockdev command is for calling select ioctl commands on block devices from the command line.
The tc command can be used to add latency to loop devices (and lots else):
tc qdisc add dev lo root handle 1:0 netem delay 20msec
tc qdisc del dev lo root
the loop device can be used to access partitions in disk images.
Note that the NASA enhanced loop device is still available from ftp://ftp.hq.nasa.gov/pub/ig/ccd/enhanced_loopback/ but it is patches against a 2.4 kernel and has not been updated since 2004.
Thursday, June 11, 2009
Telecom/Xtra blocks port 25
This isn't news - they have been blocking port 25 for a long time now. What's new for me is that, since switching from dial-up to broadband, the traffic is blocked despite my subscription to opt out of port 25 filtering. The opt out worked when I was on dial-up, even with a dynamic IP address.
Now that I am on Telecom/Xtra broadband, they tell me the only way I will be able to connect to port 25 on any other than the Telecom/Xtra smtp servers is to subscribe for a static IP address, at an additional cost of $20/month - that's 40% of my connection fee.
To make a long story short, I spent over an hour on the phone, mostly on hold, insisting that it should work. Eventually the support person gave up telling me there was nothing more he could do and escalated the call. A few hours later I had an email from technical support - they had reset their authentication cache for my account and asked me to reset my router.
So, I reset my router and I can once again connect to remote SMTP servers.
It shouldn't have taken so long, and if I didn't know IP and SMTP well enough to build their network for them I wouldn't have had the confidence to know they were wrong when they told me it couldn't be done. But I suppose those who don't know enough are very unlikely to need to connect to remote SMTP servers, so not much harm done from Telecom's perspective.
In summary, it is still possible to connect to remote SMTP servers, even with Telecom/Xtra broadband service - as long as you have hours and patience to work through their helpdesk.
Now that I am on Telecom/Xtra broadband, they tell me the only way I will be able to connect to port 25 on any other than the Telecom/Xtra smtp servers is to subscribe for a static IP address, at an additional cost of $20/month - that's 40% of my connection fee.
To make a long story short, I spent over an hour on the phone, mostly on hold, insisting that it should work. Eventually the support person gave up telling me there was nothing more he could do and escalated the call. A few hours later I had an email from technical support - they had reset their authentication cache for my account and asked me to reset my router.
So, I reset my router and I can once again connect to remote SMTP servers.
It shouldn't have taken so long, and if I didn't know IP and SMTP well enough to build their network for them I wouldn't have had the confidence to know they were wrong when they told me it couldn't be done. But I suppose those who don't know enough are very unlikely to need to connect to remote SMTP servers, so not much harm done from Telecom's perspective.
In summary, it is still possible to connect to remote SMTP servers, even with Telecom/Xtra broadband service - as long as you have hours and patience to work through their helpdesk.
Shared mailbox
I got x11vnc working but so I can connect from my laptop running vista to my workstation running CentOS5. It works OK except that I can't cut and paste between applications on Vista and Thunderbird running on CentOS5. Cut and paste works for some other applications, but something is different with Thunderbird.
I got vncserver (which is, I think, just an interface to control Xvnc - the vnc server built in to the X server) working also. Cut and paste works fine with this configuration, even with Thunderbird, but this only works with an X server on a virtual terminal - I can't use it to access and control the physical console. This is a nuisance because I often enough leave Thunderbird running on the workstation, then I can't start another copy of the second login.
So, now I have set up fetchmail and dovecot. These seem quite popular and I got them working with very little difficulty. I don't know much about them yet - I have a lot to learn before I will be comfortable that I have set them up well. Like I don't even know if my communications are secure. But it's only on my local network, so no great worries yet.
After a brief Google, I found the fetchmail configuration required to download messages from hotmail, where I have a test account. No problems here. The messages are delivered by SMTP to my local sendmail MTA and from there to my inbox.
Another brief Google and I had the required configuration for dovecot: I added
Anyway, with these set up I tried connecting from Thunderbird: this was easy. I just created a new IMAP account, selected SSL and told it to connect to localhost. Bingo-bango I had access to my messages.
Then I installed Thunderbird on my laptop and set up an equivalent account, but specifying my workstation as the server rather than localhost. I opened up inbound connection to TCP/143, from the local network only. And I now have access to my messages from my laptop.
So now I can receive messages on either my laptop or workstation.
Next, outbound...
I got vncserver (which is, I think, just an interface to control Xvnc - the vnc server built in to the X server) working also. Cut and paste works fine with this configuration, even with Thunderbird, but this only works with an X server on a virtual terminal - I can't use it to access and control the physical console. This is a nuisance because I often enough leave Thunderbird running on the workstation, then I can't start another copy of the second login.
So, now I have set up fetchmail and dovecot. These seem quite popular and I got them working with very little difficulty. I don't know much about them yet - I have a lot to learn before I will be comfortable that I have set them up well. Like I don't even know if my communications are secure. But it's only on my local network, so no great worries yet.
After a brief Google, I found the fetchmail configuration required to download messages from hotmail, where I have a test account. No problems here. The messages are delivered by SMTP to my local sendmail MTA and from there to my inbox.
Another brief Google and I had the required configuration for dovecot: I added
mail_location = mbox:~/Mail:INBOX=/var/mail/%uto /etc/dovecot.conf. With this, dovecot finds incoming email in my incoming mailbox (where sendmail deposits it) and stores folders in ~/Mail, with each folder being one mbox format file. I don't see how to get it to find messages in ~/mbox, where my local mail client stores them after reading them from my inbox. No doubt there is a way but...
Anyway, with these set up I tried connecting from Thunderbird: this was easy. I just created a new IMAP account, selected SSL and told it to connect to localhost. Bingo-bango I had access to my messages.
Then I installed Thunderbird on my laptop and set up an equivalent account, but specifying my workstation as the server rather than localhost. I opened up inbound connection to TCP/143, from the local network only. And I now have access to my messages from my laptop.
So now I can receive messages on either my laptop or workstation.
Next, outbound...
Wednesday, June 10, 2009
CentOS5 display management
From the top...
When the Linux kernel boots, the last thing it does is start the init process. This process is the parent of all processes on the system. It creates processes according to the contents of its configuration file: /etc/inittab. You can read all about it in the manual pages for init and inittab.
On a CentOS5 system with X Windows installed, /etc/inittab will include the following:
The /etc/X11/prefdm script (it's just a shell script) reads the /etc/sysconfig/desktop file to determine which display manager is preferred: by reading the DISPLAYMANAGER variable set in that configuration file. If the variable is set and prefdm is able to, the specified display manager is run. Otherwise, first gdm, then kdm then xdm are tried.
The /etc/X11/prefdm script runs display managers by exec'ing them. Thus the display manager replaces the shell process running the script. If no display manager can run, then the script exits with exit status 1. Otherwise the display manager exits with whatever status it likes.
When the script or the display manager it executed terminates, the init process runs the script again.
When the Linux kernel boots, the last thing it does is start the init process. This process is the parent of all processes on the system. It creates processes according to the contents of its configuration file: /etc/inittab. You can read all about it in the manual pages for init and inittab.
On a CentOS5 system with X Windows installed, /etc/inittab will include the following:
# Run xdm in runlevel 5The first line is just a comment. The second causes init to run /etc/X11/prefdm whenever the system is running in runlevel 5 and to restart it if it terminates.
x:5:respawn:/etc/X11/prefdm -nodaemon
The /etc/X11/prefdm script (it's just a shell script) reads the /etc/sysconfig/desktop file to determine which display manager is preferred: by reading the DISPLAYMANAGER variable set in that configuration file. If the variable is set and prefdm is able to, the specified display manager is run. Otherwise, first gdm, then kdm then xdm are tried.
The /etc/X11/prefdm script runs display managers by exec'ing them. Thus the display manager replaces the shell process running the script. If no display manager can run, then the script exits with exit status 1. Otherwise the display manager exits with whatever status it likes.
When the script or the display manager it executed terminates, the init process runs the script again.
rhgb
The rhgb program (Red Hat Graphical Boot) displays messages from the boot scripts on an X Server display. There is no man page or other documentation installed with this program but it seems to be a standard part of Red Hat installations.
The rhgb program is run if the boot parameters include rhgb. If you don't like rhgb remove this parameter from your boot parameter list (grub, lilo or whatever you use). You should then get all your boot messages appearing on a simple text display.
There is no option to provide a detailed display by default. You can edit the boot scripts to force a detailed display, but rhgb will sometime make up its own mind and switch the display anyway - like when some script takes longer than 10 seconds to complete. So, unless you pepper your init scripts with "rhgb-client --details=yes", you won't necessarily see the details.
Starting the X Server just takes more time when all you really want to do is boot.
Displaying the text from the boot scripts in a graphical display really doesn't add any value to the boot process. Maybe some people find it more "friendly" because it hides "intimidating" messages. My advice is to turn it off.
The rhgb program is run if the boot parameters include rhgb. If you don't like rhgb remove this parameter from your boot parameter list (grub, lilo or whatever you use). You should then get all your boot messages appearing on a simple text display.
There is no option to provide a detailed display by default. You can edit the boot scripts to force a detailed display, but rhgb will sometime make up its own mind and switch the display anyway - like when some script takes longer than 10 seconds to complete. So, unless you pepper your init scripts with "rhgb-client --details=yes", you won't necessarily see the details.
Starting the X Server just takes more time when all you really want to do is boot.
Displaying the text from the boot scripts in a graphical display really doesn't add any value to the boot process. Maybe some people find it more "friendly" because it hides "intimidating" messages. My advice is to turn it off.
x11vnc on CentOS5 with GDM
Setting up x11vnc to allow remote login and control of the physical console on a CentOS5 system is easy.
Install x11vnc (e.g. yum install x11vnc)
Create a GDM init script for display :0 (or whatever display you want to connect to). The default init script is /etc/gdm/Init/Default. You can copy this to a file named for the display you are using. On my system this is ':0'.
The -bg option causes x11vnc to go to background instead of staying in the foreground until terminated. This allows GDM to carry on to do its thing.
The -0 option specifies the path of a logfile.
The -reopen option tells x11vnc to reconnect to the X server if the connection is broken, which it is after GDM completes login because the initial X server is terminated after login and a new X server is started as the logged in user.
The -forever option tells x11vnc to continue running after termination of a VNC session. With this option it is possible to disconnect and then reconnect.
For details of GDM see http://library.gnome.org/admin/gdm/2.16/configuration.html.en.
Install x11vnc (e.g. yum install x11vnc)
Create a GDM init script for display :0 (or whatever display you want to connect to). The default init script is /etc/gdm/Init/Default. You can copy this to a file named for the display you are using. On my system this is ':0'.
cp /etc/gdm/Init/Default /etc/gdm/Init/:0Edit the new init script, adding the following down at the bottom, just before the exit:
Note: I had "-auth /var/gdm/:0.Xauth" option until Karl (see comments) suggested it wasn't necessary - it works fine without it. The -auth option isn't required in this case because the XAUTHORITY environment variable is already set when x11vnc is executed. If running x11vnc in some other context, it would be necessary to either set the XAUTHORITY environment variable or use the -auth option.x11vnc -bg -o /tmp/x11vnc.log -reopen -forever
The -bg option causes x11vnc to go to background instead of staying in the foreground until terminated. This allows GDM to carry on to do its thing.
The -0 option specifies the path of a logfile.
The -reopen option tells x11vnc to reconnect to the X server if the connection is broken, which it is after GDM completes login because the initial X server is terminated after login and a new X server is started as the logged in user.
The -forever option tells x11vnc to continue running after termination of a VNC session. With this option it is possible to disconnect and then reconnect.
For details of GDM see http://library.gnome.org/admin/gdm/2.16/configuration.html.en.
Tuesday, June 9, 2009
Windows Vista Sidebar Gadgets
I have a new laptop with Windows Vista pre-installed. I have avoided Vista for as long as possible but my old laptop died and it seems to be impossible to buy a new one with XP these days. Before I install Linux, I thought I would try out Vista for a while.
One of the new features is the sidebar gadgets. I haven't seen them before, so I went to the Microsoft site to learn more about them. First I found marketing hype about how great they are, then I found http://vista.gallery.microsoft.com/vista/SideBar.aspx?mkt=en-nz. This page has downloads and links to information but all the links are to a "live.com" website. That site looks like some sort of social network / shareware distribution site.
So I called Microsoft to find out where the documentation is and after a long conversation with the support analyst the best they could refer me to was end-user documentation. I explained that I am an IT professional and I need to know how to manage them in a corporate environment, to which they responded that I should search for some third party websites that have more information and suggestions how to tinker with the gadgets. So I explained again that I am an IT professional and I don't want to tinker with gadgets, I want to manage corporate desktops professionally. The analyst went away to search again and finally came back to say that the end user help pages is all the documentation that Microsoft has to offer and if I wanted more I would have to go to a third party.
So, after all these years, the best desktop Microsoft has to offer is, from a corporate management perspective, undocumented. It gives me a sinking feeling.
But wait... After much searching about the Microsoft link and following links here there and everywhere, I finally found a page that has some reasonably technical information:
One of the new features is the sidebar gadgets. I haven't seen them before, so I went to the Microsoft site to learn more about them. First I found marketing hype about how great they are, then I found http://vista.gallery.microsoft.com/vista/SideBar.aspx?mkt=en-nz. This page has downloads and links to information but all the links are to a "live.com" website. That site looks like some sort of social network / shareware distribution site.
So I called Microsoft to find out where the documentation is and after a long conversation with the support analyst the best they could refer me to was end-user documentation. I explained that I am an IT professional and I need to know how to manage them in a corporate environment, to which they responded that I should search for some third party websites that have more information and suggestions how to tinker with the gadgets. So I explained again that I am an IT professional and I don't want to tinker with gadgets, I want to manage corporate desktops professionally. The analyst went away to search again and finally came back to say that the end user help pages is all the documentation that Microsoft has to offer and if I wanted more I would have to go to a third party.
So, after all these years, the best desktop Microsoft has to offer is, from a corporate management perspective, undocumented. It gives me a sinking feeling.
But wait... After much searching about the Microsoft link and following links here there and everywhere, I finally found a page that has some reasonably technical information:
http://msdn.microsoft.com/en-nz/library/aa965850(en-us,VS.85).aspx
And another Microsoft page for those who want to tinker with gadgets
http://msdn.microsoft.com/en-us/magazine/cc163370.aspx
And another Microsoft page for those who want to tinker with gadgets
http://msdn.microsoft.com/en-us/magazine/cc163370.aspx
Subscribe to:
Posts (Atom)