Shallow Thoughts : : kernel

Making Sure the Debian Kernel is Up To Date

I try to avoid Grub2 on my Linux machines, for reasons I've discussed before. Even if I run it, I usually block it from auto-updating /boot since that tends to overwrite other operating systems. But on a couple of my Debian machines, that has meant needing to notice when a system update has installed a new kernel, so I can update the relevant boot files. Inevitably, I fail to notice, and end up running an out of date kernel.

But didn't Debian use to have a /boot/vmlinuz that always linked to the latest kernel? That was such a good idea: what happened to that?

I'll get to that. But before I found out, I got sidetracked trying to find a way to check whether my kernel was up-to-date, so I could have it warn me of out-of-date kernels when I log in.

That turned out to be fairly easy using uname and a little shell pipery:

# Is the kernel running behind?
kernelvers=$(uname -a | awk '{ print $3; }')
latestvers=$(cd /boot; ls -1 vmlinuz-* | sort --version-sort | tail -1 | sed 's/vmlinuz-//')
if [[ $kernelvers != $latestvers ]]; then
    echo "======= Running kernel $kernelvers but $latestvers is available"
else
    echo "The kernel is up to date"
fi

I put that in my .login. But meanwhile I discovered that that /boot/vmlinuz link still exists -- it just isn't enabled by default for some strange reason. That, of course, is the right way to make sure you're on the latest kernel, and you can do it with the linux-update-symlinks command.

linux-update-symlinks is called automatically when you install a new kernel -- but by default it updates symlinks in the root directory, /, which isn't much help if you're trying to boot off a separate /boot partition.

But you can configure it to notice your /boot partition. Edit /etc/kernel-img.conf and change link_in_boot to yes:

link_in_boot = yes

Then linux-update-symlinks will automatically update the /boot/vmlinuz link whenever you update the kernel, and whatever bootloader you prefer can point to that image. It also updates /boot/vmlinuz.old to point to the previous kernel in case you can't boot from the new one.

Update: To get linux-update-symlinks to update symlinks to reflect the current kernel, you need to reinstall the package for the current kernel, e.g. apt-get install --reinstall linux-image-4.18.0-3-amd64. Just apt-get install --reinstall linux-image-amd64 isn't enough.

Tags: linux, debian, shell
[ 20:14 Aug 23, 2018    More linux/kernel | permalink to this entry | ]

Mounting a Samsung Galaxy Player on Linux

My new toy: a Samsung Galaxy Player 5.0!

So far I love it. It's everything my old Archos 5 wanted to be when it grew up, except the Archos never grew up. It's the same size, a little lighter weight, reliable hardware (no random reboots), great battery life, fast GPS, modern Android 2.3, and the camera is even not too bad (though it certainly wouldn't tempt me away from my Canon).

For the record, Dave got a Galaxy Player 4.0, and it's very nifty too, and choosing between them was a tough choice -- the 4-inch is light and so easy to hold, and it uses replaceable batteries, while the 5-inch's screen is better for reading and maps.

USB-storage devices don't register

I love the Galaxy ... but there's one thing that bugs me about it. When I plug it in to Linux, dmesg reports two new storage devices, one for main storage and one for the SD card. Just like most Android devices, so far.

The difference is that these Samsung devices aren't fully there. They don't show up in /proc/partitions or in /dev/disk/by-uuid, dmesg doesn't show sizes for them, and, most important, they can't be mounted by UUID from an fstab entry, like

UUID=62B0-C667   /droidsd    vfat   user,noauto,exec,fmask=111,shortname=lower 0 0

That meant I couldn't mount it as myself -- I had to become root, figure out where it happened to show up this time (/dev/sdf or wherever, depending on what else might be plugged in), mount it, then do all my file transfers as root.

I found that if I mounted it explicitly using the device pathname -- mount /dev/sdf /mnt -- then subsequently the device shows up normally, even after I unmount it. So I could check dmesg to find the device name, mount it as root, unmount as root, then mount it again as myself using my fstab entry. What a pain!

A kernel expert I asked thought it looked like the Samsung is pretending to be a removable device, but only "plugging in" when the system actually tries to access it. Annoying. So how do you get Linux to "access" it?

Udev: still an exercise in frustration

The obvious solution is a udev rule. Some scrutiny of /lib/udev/rules.d/60-persistent-storage.rules found some rules that did this intriguing thing: IMPORT{program}="ata_id --export $tempnode".

Naturally, this mysterious ata_id is undocumented. It's hidden in /lib/udev/ata_id, and I found this tiny ata_id man page online since there's none available in Ubuntu. Running ata_id /dev/sdf seemed to do what I needed: it made the device show up in /proc/partitions and /dev/disk/by-uuid, and after that, I could mount it without being root.

I created a file named /etc/udev/rules.d/59-galaxy.rules, with the rule:

KERNEL=="sd[b-g]", SUBSYSTEMS=="usb", ATTRS{idVendor}=="04e8", SYMLINK+="galaxy-%k-%n", IMPORT{program}="ata_id --export $tempnode"

When I tested it with udevadm test /block/sdf, not only did the rule trigger correctly, but it actually ran ata_id and the device became visible -- even though udevadm test states clearly no programs will be run. How do I love udev -- let me count the ways.

But a reboot revealed that udev was not actually running the rule when I actually plugged the Galaxy in -- the devices did not become visible. Did I mention how much I love udev?

Much simpler: a shell alias

But one thing I'd noticed in all this: side by side with /dev/disk/by-uuid is a directory called /dev/disk/by-id. And the Samsung devices did show up there, with names like usb-Android_UMS_Composite_c197022a2b41c1ae-0:0.

Faced with the prospect of spending the rest of the day trying random udev rules, rebooting each time since that's the only way to test udev, I decided to cheat and find another way. I made a shell alias:

alias galaxy='sudo sh -c "for d in /dev/disk/by-id/usb-Android*; do /lib/udev/ata_id --export \$d; done"'

Typing galaxy will now cause the Samsung to register both devices; and from then on I can mount and unmount them without needing root, using my normal fstab entries.

Update: This works for the Nook's main storage, too -- just add x/dev/disk/by-id/usb-B_N_Ebook_Disk* to the list -- but it doesn't work reliably for the Nook's SD card. The SD card does show up in /dev/disk/by-id along with main storage; but running ata_id on it doesn't make its UUID appear. I may just change my fstab entry to refer to the /dev/disk/by-id device directly.

Tags: android, linux, udev
[ 13:43 Apr 18, 2012    More linux/kernel | permalink to this entry | ]

Enabling system beeps on Ubuntu Lucid

Update: though the rest of this article is still useful in explaining how to un-blacklist the pcspkr module, unfortunately that module works very erratically. Sometimes you'll get a beep, sometimes not. So this article may be a good start but it still doesn't explain why Ubuntu's kernels have such a flaky pcspkr module.

For years I've used Ubuntu with my own kernels rather than the kernels Ubuntu provides. I have several reasons: home-built kernels boot a lot faster (when I say a lot, I mean like saving 30 seconds off a one-minute boot) and offer more control over options. But a minor reason is that Ubuntu kernels generally don't support the system beep, so for example there's no way to tell in vim when you get out of insert mode. (In the past I've sometimes used the excellent fancy beeper module to play sounds, but I don't always want that.)

On Ubuntu's latest "Lucid Lynx", I'm using their kernel (so far). The Ubuntu kernel team has made huge improvements in boot time and number of modules loaded, so it's much more efficient than past kernels. But it did leave me without a beeper.

modprobe pcspkr failed to do anything except print the enigmatic:

WARNING: All config files need .conf: /etc/modprobe.d/00local, it will be ignored in a future release.

modprobe -v pcspkr (verbose) was no help -- it printed install /bin/true which didn't make anything clearer.

To get my beep back, I had to do two things:

First, edit /etc/modprobe.d/blacklist.conf and comment out the line blacklisting pcspeakr. It looks like this:

# ugly and loud noise, getting on everyone's nerves; this should be done by a
# nice pulseaudio bing (Ubuntu: #77010)
blacklist pcspkr

(They don't seem to be concerned about anyone who doesn't run Pulse, or about the various other bugs involved -- there's quite a laundry list in bug 486154.)

Secomd. pcspkr was blacklisted a second time in a different way, in that file so confusingly alluded to by the warning. /etc/modprobe.d/00local was apparently left over from a previous version of Ubuntu, and never removed by any upgrade script, and consisted of this:

install pcspkr /bin/true

Aha! So that's why modprobe -v pcspkr printed install /bin/true -- because that's all it was doing instead of loading the module like I'd asked.

So rm /etc/modprobe.d/00local was the second step, and once I'd done that, modprobe pcspkr loaded the module and gave me my system beep.

Tags: ubuntu, lucid, audio
[ 14:02 Jun 13, 2010    More linux/kernel | permalink to this entry | ]

The new udev in Lucid

Ubuntu's latest release, 10.04 "Lucid Lynx", really seems remarkably solid. It boots much faster than any Ubuntu of the past three years, and has some other nice improvements too.

But like every release, they made some pointless random undocumented changes that broke stuff. The most frustrating has been getting my front-panel flash card reader to work under Lucid's new udev, so I could read SD cards from my camera and PDA.

The SD card slot shows up as /dev/sdb, but unless there's a card plugged in at boot time, there's no /dev/sdb1 that you can actually mount.

hal vs udisks

Prior to Lucid, the "approved" way of creating sdb1 was to let hald-addons-storage poll every USB device every so often, to see if anyone has plugged in a card and if so, check its partition table and create appropriate devices.

That's a lot of polling -- and in any case, hald isn't standard on Lucid, and even when it's installed, it sometimes runs and sometimes doesn't. (I haven't figured out what controls whether it decides to run). Hal isn't even supposed to be needed on Lucid -- it's supposed to use devicekit (renamed to) udisks for that.

Except I guess they couldn't quite figure out how to get udisks working in time, so they patched things together so that on Gnome systems, hald does the same old polling stuff -- and on non Gnome systems, well, maybe it does and maybe it doesn't. And maybe you can't read your camera cards. Oh well!

udev rules

But on systems prior to Lucid there was another way: make a udev rule to create sdb1 through sdb15 every time. I have an older article on setting up udev rules for multicard readers, but none of my old udev rules worked on Lucid.

After many rounds of udevadm info -a -p /block/sdb and udevadm test /block/sdb, service udev restart, and many reboots, I finally found a rule that worked.

Create a /etc/udev/rules.d/71-multicard-reader.rules file containing the following:

# Create all devices for multicard reader:
KERNEL=="sd[b-g]", SUBSYSTEMS=="usb", ATTRS{idVendor}=="1d6b", ATTRS{idProduct}=="0002", OPTIONS+="all_partitions,last_rule"

Replace the 1d6b and 0002 with the vendor and product of your own device, as determined with udevadm info -a -p /block/sdb ... and don't be tempted to use the vendor and device ID you get from lsusb, because those are different.

What didn't work that used to? String matches. Some of them. For example, this worked:

KERNEL=="sd[b-g]", SUBSYSTEMS=="scsi", ATTRS{model}=="*SD*", NAME{all_partitions}="sdcard"

but these didn't:

KERNEL=="sd[b-g]", SUBSYSTEMS=="scsi", ATTRS{model}=="*SD*Reader*", NAME{all_partitions}="sdcard"
KERNEL=="sd[a-g]", SUBSYSTEMS=="scsi", ATTRS{model}=="USB SD Reader   ", NAME{all_partitions}="cardsd"

Update: The first of those two lines does indeed work now, whereas it didn't when I was testing. It's possible that this has something to do with saving hardware states and needing an extra udevadm trigger, as suggested in Alex's Changes in Ubuntu Lucid to udev.

According to udevadm info, the model is "USB SD Reader " (three spaces at the end). But somehow "*SD*" matches this while "*SD*Reader*" and the exact string do not. Go figure.

Numeric order

I'd like to have this rule run earlier, so it runs before /lib/udev/rules.d/60-persistent-storage.rules and could use OPTIONS+="last_rule" to keep the persistent storage rules from firing (they run a lot of unnecessary external programs for each device). But if I rename the rule from 71-multicard-reader.rules to 59-, it doesn't run at all. Why? Shrug. It's not like udevadm test will tell me.

Other things I love (not) about the new udev

• I love how if you give the udevadm info arguments in the wrong order, -p -a, it means something else and gives an error message.
• I love how udevadm test doesn't actually test the same rules udev will use, so it's completely unrelated to anything.
• I love the complete lack of documentation on things like string matching and how the numeric order is handled.
• I love how you can't match both the device name (a string) and the USB IDs in the same rule, because one is SUBSYSTEMS=="scsi" and the other is SUBSYSTEMS=="usb".
• Finally, I love how there's no longer any way to test udev rules on a running system -- if you want it to actually create new devices, you have to reboot for each new test. service udev restart and udevadm control --reload-rules don't touch existing devices. Gives me that warm feeling like maybe I'm not missing out on the full Windows experience by using Linux.

Tags: linux, ubuntu, kernel, udev, install
[ 21:51 May 09, 2010    More linux/kernel | permalink to this entry | ]

Function keys on an Apple keyboard, on Linux

Quick tip from Dave, passed along to someone else trying to use an Apple keyboard on Linux:

On Linux, for some reason Apple keyboards' function keys don't work by default. Most of them try to run special functions instead, like volume up/down or play/pause.

But you can get normal function keys by talking to the kernel module that drives the keyboard:

echo 2 >  /sys/module/hid_apple/parameters/fnmode

This will only last until shutdown, so put that line in /etc/rc.local or a similar place so it runs every time you boot.

Here's an Ubuntu help page on Apple Keyboards with more information and other tricks.

Tags: linux, apple, kernel, keyboard
[ 16:03 Apr 15, 2010    More linux/kernel | permalink to this entry | ]

Making those Fn- laptop keys do something useful

A friend was trying to get some of her laptop's function keys working under Ubuntu, and that reminded me that I'd been meaning to do the same on my Vaio TX 650P.

My brightness keys worked automagically -- I suspected via the scripts in /etc/acpi -- and that was helpful in tracking down the rest of the information I needed. But it still took a bit of fiddling since (surprise!) how this stuff works isn't documented.

Update: That "isn't documented" remark applies to the ACPI system. Matt Zimmerman points out that there is some good documentation on the rest of the key-handling system, and pointed me to two really excellent pages: Hotkeys architecture and Hotkeys Troubleshooting. Recommended reading!

Here's the procedure I found.

First, use acpi_listen to find out what events are generated by the key you care about. Not all keys generate ACPI events. I haven't get figured out what controls this -- possibly the kernel. When you type the key, you're looking for something like this:

sony/hotkey SPIC 00000001 00000012

You may get separate events for key down and key up. It's your choice as to which one matters.

Once you know the code for your key, it's time to make it do something. Create a new file in /etc/acpi/events -- I called mine sony-lcd-btn. It doesn't matter what you call it -- acpid will read all of them. (Yes, that means every time you start up it's reading all those toshiba and asus files even if you have a Lenovo or Sony. Looks like a nice place to shave off a little boot time.)

The file is very simple and should look something like this:

# /etc/acpi/events/sony-lcd-btn

event=sony/hotkey SPIC 00000001 00000012
action=/etc/acpi/sonylcd.sh

Now create a script for the action you specified in the event file. I created a script /etc/acpi/sonylcd.sh that looks like this:

#! /bin/bash
# temporary, for testing:
echo "LCD button!" >/dev/console

Now restart acpid: service acpid restart if you're on karmic, or /etc/init.d/acpid restart on earlier releases. Press the button. If you're running from the console (or using a tool like xconsole), and you got all the codes right, you should be able to see the echo from your script.

Now you can do anything you want. For instance, when I press the LCD button I generally want to run this:

xrandr --output VGA --mode 1024x768

Or to make it toggle, I could write a slightly smarter script using xrandr --query to find out the current mode and behave accordingly. I'll probably do that at some point when I have a projector handy.

Tags: linux, acpi,
[ 17:15 Mar 09, 2010    More linux/kernel | permalink to this entry | ]

More on kernel options needed for new X servers

A week ago I wrote about my mouse woes and how they were solved by enabling the "Enable modesetting on intel by default" kernel option.

But all was still not right with X. I could boot into a console, start X, and everything was fine -- but once X was running, I couldn't go back to console mode. Switching to a console, e.g. Ctrl-Alt-F2, did nothing except make the mouse cursor disappear, and any attempt to quit X to go back to my login shell put the monitor in sleep mode permanently -- the machine was still up, and I could ssh in or ctrl-alt-Delete to reboot, but nothing else I did would bring my screen back.

It wasn't strictly an Ubuntu problem, though this showed up with Karmic; I have a Gentoo install on another partition and it had the same problem. And I knew it was a kernel problem, because the Ubuntu kernel did let me quit X.

I sought in vain among the kernel's various Graphics settings. You might think that "enable modesetting" would be related to, you know, being unable to change video modes ... but it wasn't. I tried different DRM options and switching framebuffer on and off. Though, oddly, enabling framebuffer didn't actually seem to enable the framebuffer.

Finally I stepped through the Graphics section of make menuconfig comparing my settings with a working kernel, and saw a couple of differences that didn't look at all important: "Select compiled-in fonts" and "VGA 8x16 font". Silly, but what could they hurt? I switched them on and rebuilt.

And on the next boot, I had a framebuffer, and mode switching.

So be warned: those compiled-in fonts are not optional if you want a framebuffer; and you'd better want a framebuffer, because that isn't optional either if you want to be able to get out of X once you start it.

Tags: linux, intel, X11, kernel
[ 20:22 Nov 16, 2009    More linux/kernel | permalink to this entry | ]

Mouse failures with 2.6.31, Karmic and Intel

I've been seeing intermittent mouse failures since upgrading to Ubuntu 9.10 "Karmic". At first, maybe one time out of five I would boot, start X, and find that I couldn't move my mouse pointer. But after building a 2.6.31.4 kernel, things got worse and it happened nearly every time.

It wasn't purely an X problem; if I enabled gpm, the mouse failed in the console as well as in X. And it wasn't hardware, because if I used Ubuntu 9.10's standard kernel, my mouse worked every time.

After much poking around with kernel options, I discovered that if I tunred off the Direct Rendering manager ("Intel 830M, 845G, 852GM, 855GM, 865G (i915 driver)"), my mouse would work. But that wasn't a satisfactory solution; aside from not being able to run Google Earth, it seems that Intel graphics needs DRM even to get reasonable performance redrawing windows. Without it, every desktop switch means watching windows slowly redraw over two or three seconds.

(Aside: why is it that Intel cards with shared CPU memory need DRM to draw basic 2-D windows, when my ancient ATI Radeon cards without shared memory had no such problems?)

But I think I finally have it nailed. In the kernel's Direct Rendering Manager options (under Graphics), the "Intel 830M, 845G, 852GM, 855GM, 865G (i915 driver)" using its "i915 driver" option has a new sub-option: "Enable modesetting on intel by default".

The help says:

CONFIG_DRM_I915_KMS:
Choose this option if you want kernel modesetting enabled by default, and you have a new enough userspace to support this. Running old userspaces with this enabled will cause pain. Note that this causes the driver to bind to PCI devices, which precludes loading things like intelfb.

Sounds optional, right? Sounds like, if I want to build a kernel that will work on both karmic and jaunty, I should leave that off so as not to "cause pain".

But no. It turns out it's actually mandatory on karmic. Without it, there's a race condition where about 80-90% of the time, hal won't see a mouse device at all, so the mouse won't work either in X or even on the console with gpm.

It's sort of the opposite of the "Remove sysfs features which may confuse old userspace tools" in General Setup, where the name implies that it's optional on new distros like Karmic, but in fact, if you leave it on, the kernel won't work reliably.

So be warned when configuring a kernel for brand-new distros. There are some new pitfalls, and options that worked in the past may not work any longer!

Update: see also the followup post for two more non-optional options.

Tags: linux, ubuntu, intel, X11, kernel
[ 23:34 Nov 10, 2009    More linux/kernel | permalink to this entry | ]