Huckle's questions

I am looking through the various *fdisk and *parted utilities along with a few others. I am searching for a utility (or collection of utilities) that will list the size, in bytes, of all of the disks and partitions on a machine. Bonus points for any additional information on the drive or partition such as make/model/label/filesystem. It should also be able to report the size of drives with no partition table installed

• plain fdisk
  • ✗ No GUID Partition Table support
  • ✗ Not in bytes, no flags to allow for it
• sfdisk
  • ✗ No GUID Partition Table support
  • ✓ Flag to support Megabytes (close)
• parted
  • ✓ Does support GPT
  • ! Option to control unit only works on one disk at a time
  • ✗ Rounds to largest possible unit
• lshw
  • ✗ Doesn't show partition info
  • ✗ Rounds to highest unit
  • ✗ No option to control units
• pvdisplay / pvs
  • ✗ Only works on disks that are part of LVM2 array
  • ✗ Doesn't show disk info if partition is volume used in array
  • ✓ Does have flags to set unit to bytes

Not sure if this is a duplicate of Broadcom STA driver doesn't work well with BCM4313 or not. It's the same series I think but I've seen no explicit reference to 4352 or 43xx (implying it's applicable for all 43xx series).

Canonical claims support for this card over here and when I booted into the live install CD it worked after enabling the driver in software-properties. Admittingly I didn't use it enough to see if the connection was flaky. Instead I booted into a server install USB and I'm now attempting to configure it post install.

My current procedure is to run this script, which is built from snippets I've been reading about at various locations.

CONFFILE=/etc/wpa_supplicant.conf
CONF=$(cat <<CONFEND
ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=wireless
update_config=1

ENDCONF
)

echo "$CONF" | sudo tee $CONFFILE > /dev/null
wpa_passphrase "" "" | sudo tee -a $CONFFILE > /dev/null

sudo kill-all wpa_supplicant
sudo wpa_supplicant -B -Dwext -ieth0 -c$CONFFILE
sudo dhclient eth0

Now sometimes this runs in about 15 seconds, emits two errors (ioctl[SIOCSIWENCODEEXT]: Invalid argument), but works fine. Can't tell if the connection is flaky or I have selected a poor mirror in apt. However other times it takes > 5 minutes to run and emits no useful information before I kill it. I try to run the wpa_supplicant and dhclient lines by themselves and each take ages and report no useful errors. The end result in these cases is that iwconfig shows the interface as associated but ifconfig shows no ipv4 address and dhclient refuses to fetch one.

Here's some debugging info:

Kernel log is full of these:

Feb 16 20:59:49 ALPHA kernel: [ 3680.237755] ------------[ cut here ]-----------
-
Feb 16 20:59:49 ALPHA kernel: [ 3680.237805] WARNING: CPU: 1 PID: 476 at /build/
buildd/linux-3.11.0/net/wireless/sme.c:795 cfg80211_roamed+0x89/0x90 [cfg80211](
)
Feb 16 20:59:49 ALPHA kernel: [ 3680.237808] Modules linked in: michael_mic arc4
 snd_hda_codec_hdmi lib80211_crypt_tkip x86_pkg_temp_thermal coretemp wl(POF) kv
m crct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel aes_x86_64 lrw g
f128mul glue_helper ablk_helper cryptd i915 snd_hda_codec_realtek joydev eeepc_w
mi snd_hda_intel drm_kms_helper lib80211 asus_wmi snd_hda_codec sparse_keymap sn
d_hwdep drm snd_pcm hid_generic snd_page_alloc snd_seq_midi snd_seq_midi_event c
fg80211 snd_rawmidi i2c_algo_bit psmouse snd_seq snd_seq_device usbhid mei_me sn
d_timer microcode mei hid serio_raw snd e1000e soundcore video ptp lpc_ich wmi p
ps_core mac_hid lp parport ahci libahci
Feb 16 20:59:49 ALPHA kernel: [ 3680.237888] CPU: 1 PID: 476 Comm: wl_event_handle Tainted: PF       W  O 3.11.0-15-generic #25-Ubuntu
Feb 16 20:59:49 ALPHA kernel: [ 3680.237891] Hardware name: ASUS All Series/Z87I-DELUXE, BIOS 0702 08/30/2013
Feb 16 20:59:49 ALPHA kernel: [ 3680.237895]  0000000000000009 ffff880210463d88 ffffffff816e7375 0000000000000000
Feb 16 20:59:49 ALPHA kernel: [ 3680.237903]  ffff880210463dc0 ffffffff81061dcd ffff88021436a000 ffff88020f4ad000
Feb 16 20:59:49 ALPHA kernel: [ 3680.237909]  00000000000000a3 ffff880213223f00 0000000000000001 ffff880210463dd0
Feb 16 20:59:49 ALPHA kernel: [ 3680.237916] Call Trace:
Feb 16 20:59:49 ALPHA kernel: [ 3680.237930]  [<ffffffff816e7375>] dump_stack+0x45/0x56
Feb 16 20:59:49 ALPHA kernel: [ 3680.237942]  [<ffffffff81061dcd>] warn_slowpath_common+0x7d/0xa0
Feb 16 20:59:49 ALPHA kernel: [ 3680.237949]  [<ffffffff81061eaa>] warn_slowpath_null+0x1a/0x20
Feb 16 20:59:49 ALPHA kernel: [ 3680.237975]  [<ffffffffa01c4409>] cfg80211_roamed+0x89/0x90 [cfg80211]
Feb 16 20:59:49 ALPHA kernel: [ 3680.238056]  [<ffffffffa0862ea5>] wl_bss_connect_done.isra.21+0x105/0x1b0 [wl]
Feb 16 20:59:49 ALPHA kernel: [ 3680.238124]  [<ffffffffa086314c>] wl_notify_connect_status+0x1fc/0x410 [wl]
Feb 16 20:59:49 ALPHA kernel: [ 3680.238189]  [<ffffffffa0862445>] wl_event_handler+0x55/0x1f0 [wl]
Feb 16 20:59:49 ALPHA kernel: [ 3680.238253]  [<ffffffffa08623f0>] ? wl_cfg80211_scan+0x350/0x350 [wl]

ifconfig (when the script works)

eth0      Link encap:Ethernet  HWaddr 24:0a:64:31:13:1b
          inet addr:192.168.1.105  Bcast:192.168.1.255  Mask:255.255.255.0
          inet6 addr: fe80::260a:64ff:fe31:131b/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:652 errors:0 dropped:1 overruns:0 frame:259012
          TX packets:493 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:83037 (83.0 KB)  TX bytes:69997 (69.9 KB)
          Interrupt:19

iwconfig (again, when it works)

eth0      IEEE 802.11abg  ESSID:""
          Mode:Managed  Frequency:2.412 GHz  Access Point: 58:6D:8F:7E:7A:77
          Bit Rate=130 Mb/s   Tx-Power=200 dBm
          Retry  long limit:7   RTS thr:off   Fragment thr:off
          Encryption key:off
          Power Management:off
          Link Quality=58/70  Signal level=-52 dBm
          Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:0  Invalid misc:0   Missed beacon:0

lshw

  *-network
       description: Wireless interface
       product: BCM4352 802.11ac Wireless Network Adapter
       vendor: Broadcom Corporation
       physical id: 0
       bus info: pci@0000:03:00.0
       logical name: eth0
       version: 03
       serial: 24:0a:64:31:13:1b
       width: 64 bits
       clock: 33MHz
       capabilities: pm msi pciexpress bus_master cap_list ethernet physical wireless
       configuration: broadcast=yes driver=wl0 driverversion=6.30.223.141 (r415941) ip=192.168.1.105 latency=0 multicast=yes wireless=IEEE 802.11abg
       resources: irq:19 memory:f7a00000-f7a07fff memory:f7800000-f79fffff

wpa_supplicant -h (supported drivers)

drivers:
  wext = Linux wireless extensions (generic)
  nl80211 = Linux nl80211/cfg80211
  wired = Wired Ethernet driver
  none = no driver (RADIUS server/WPS ER)

lspci -nn -d 14e4:

03:00.0 Network controller [0280]: Broadcom Corporation BCM4352 802.11ac Wireless Network Adapter [14e4:43b1] (rev 03)

sudo cat /etc/network/interfaces | grep -Ev '^#'

auto lo
iface lo inet loopback

auto em1
iface em1 inet dhcp

auto eth0
iface eth0 inet dhcp

Edit: added extra diagnostic info

I haven't installed Ubuntu from a disc/usb in ages. At least 2 LTS versions. I have always just upgraded my server in place. I purchased new hardware that will replace that older server and instead of just moving the LVM over from the old server I've decided to start fresh.

My normal procedure for fresh server installs is:

1. Install from CD/USB without network connection
2. Log in, attach network, install some base packages that are not there by default
3. sudo apt-get update && sudo apt-get upgrade to kick off the massive download
4. Continue to setup server while (3) completes

This normally results in about 15 minutes of me waiting for the server to install offline, followed by a much longer period where the updates download and install but at least by that point I can start using the server.

With the 13.10 server image that I just pulled down via torrent I cannot seem to find a way to install without a functional network connection. I've attempted Normal. Minimal, and OEM install modes and in each one I cannot get past the network setup. When autoconfiguration fails because no network is present, I select configure network later and it still prompts me for an Ubuntu mirror server. Until it can connect to that server it will not progress.

This disrupts my workflow and makes isolated network installs impossible. What gives?

I feel the community has long needed a clean guide on how to install Grub2 under a a few extremely common scenarios. I will accept answer as solved when it has one section per scenario and assumes nothing other than what is specified. Please add to the existing answer, wiki style, keeping to the original assumptions.

Rules:
1. You cannot, at any point in the answer, invoke Ubiquity (the Ubuntu installer).
2. I strongly recommend not using any automatic boor-repair tools as they're not very educational

Scenario 1: Non-booting Linux OS, No boot partition, Fix from Live CD

Setup:

• /dev/sda1 is formatted ext*
• /dev/sda2 is formatted linux_swap
• /dev/sda1 doesn't boot because MBR is scrambled and /boot/grub/* was erased
• Would be the equivalent of running rm -rf /boot/grub and dd if=/dev/zero of=/dev/sda bs=440 count=1

Explain:

• How to boot to a Live CD / USB and restore Grub2 to the MBR and /boot of /dev/sda1

Scenario 2: Non-booting Linux OS, Boot partition, Fix from Live CD

Setup:

• /dev/sda1 is formatted fat
• /dev/sda2 is formatted ext*
• /dev/sda3 is formatted linux_swap
• /dev/sda2 doesn't boot because the MBR is scrambled and /dev/sda1 was formatted

Explain:

• How to boot to a Live CD / USB and restore Grub2 to the MBR and /dev/sda1 and then update the fstab on /dev/sda2

Scenario 3: Install on to thumb drive, Booting various OSes, From Linux OS

Setup:

• /dev/sdb is removable media
• /dev/sdb1 is formatted fat
• /dev/sdb2 is formatted ext*
• /dev/sdb3 is formatted fat
• The MBR of /dev/sdb is otherwise not initialized
• You are executing from a Linux based OS installed on /dev/sda

Explain:

• How to install Grub2 on to /dev/sdb1, mark /dev/sdb1 active, be able to chose between /dev/sdb2 and /dev/sdb3 on boot.

Scenario 4: (Bonus) Install on to thumb drive, Booting ISO, From Linux OS

Setup:

• /dev/sdb is removable media
• /dev/sdb1 is formatted fat
• /dev/sdb1 contains /iso/live.iso
• /dev/sdb2 is formatted ext*
• /dev/sdb3 is formatted fat
• The MBR of /dev/sdb is otherwise not initialized
• You are executing from a Linux based OS installed on /dev/sda

Explain:

• How to install Grub2 on to /dev/sdb1, mark /dev/sdb1 active, be able to chose between /dev/sdb2, /dev/sdb3, and /iso/live.iso on boot.

I aliased google-chrome and whitelisted myself in sudoers so now when I type sudo chrome it runs sudo -u chromeuser -i google-chrome &. (For reasons outlined below) Since all other users besides myself have a home directory on ext3 this works very well. The only downside is that chrome, when started as chromeuser, will not play sound.

edit: Moved question to top for clarity

Background for anyone who's interested:

I set the home directoy for one of my users to be on an NTFS partiton so that it is accessible from Windows. I'm the only one who uses the computer so I set fuse to allow 777 permissions to that partition. I deleted /home/me and created a symlink in it's place that points to the NTFS partition (which is mounted at /Shared/home).

Chrome doesn't play well with that configuration because it creates various caches under your home directory and NTFS performance couldn't keep up with it. I fiddled with a few options to chrome (--user-data-dir --disk-cache-dir) but I couldn't get 100% of chrome's open files under my control. So my solution was just run chrome as it's own user.

tldr; I started a program as another user, doesn't seem to want to play sound. Is this related to alsa?

I just installed a fresh copy of Mythbuntu 12.04. I installed it as a combined frontend and backend. I ran through the configurations, accepting all defaults except for two items:

• I set the hostname to the server's IP
• I checked the checkbox to enable network remote controls

I then rebooted. When I try to connect via mythmote, I get "connection refused" and "connection timed out" errors. What I have done to diagnose the issue:

• Reboot
• Re-run configuration without changing any settings
• Disable IPv6 (ala /proc/sys/net/ipv6/conf/all/disable_ipv6)
• Disable ufw
• Run Wireshark, which reports that a TCP SYN arrives at the server, and the server sends a TCP RST in response. Prior to disabling ufw the server would not even send the TCP RST

The settings I have for mythmote are:

• Name: Gamma (this is my third server)
• Address: 10.0.0.14
• Port 6546
• MAC: (Not filled in, I don't need WOL)

If I switch to a TTY on the server, I can telnet 127.0.0.1 6546 and get to the network remote control interface. If I try the same via telnet 10.0.0.14 6546 I get a connection refused.

I'm not normally a fan of mac software implementations, but recently I saw a copy of Parallels Desktop for mac that I was very much impressed by. I looked into parallel's offerings for Linux (I was initially excited that they developed for *nix) but found that it has essentially the same features as VMware.

Does anyone know if there is a software solution for linux that offers the level of integration that the current Parallels Deesktop for mac has? I looked into wine for the programs I want to run and was very frustrated. I feel like all the ratings in their application database were fabricated honesty. I've tried wine once every Ubuntu release since 8.10 and have never gotten it to work stably.

I was just installing 12.04 beta2 when I noticed something really odd. I was at the screen where you enter your username/hostname/password/etc. I typed 'Me' for my real name as I normally do. This is the weird part. The installer filled in the hostname as 'me-###-##5-1544' where the ###-##5-1544 is my parent's land-line phone number (about two hundred miles west of where i am). Where did the installer get that number? I am installing over a previous install of ubuntu, but that was a throw-away install that I never put that number on. I have other partitions and other OSes on this hard drive, but none were mounted at the time of install and none were selected to be used as mount points in the previous screen.

Picture of what I'm talking about:

I am setting up an environment where I am giving several users SSH access on my server. They are all trusted, but I want to corral them into a segment of the filesystem. I created the users like so:

adduser username -ingroup groupname

which works well enough. When I log in as one of them I can do this and get all the right answers:

$id -r -u -n
username
$id -r -g -n
groupname

I switched my own user account's primary group to groupname by doing this:

$usermod myuser groupname

I then logged out and back in. Now the files I would like the group to be able to access are under here:

$ls -l / | grep groupname
drwxr-xr-x  3 root     groupname    4.0K 2012-03-26 20:20 groupfiles
$cd /groupfiles
$ls -l 
drwxrwxr-x  2 root groupname    4.0K 2012-03-26 20:32 project

The permissions are by design, members of the group cannot alter files/folders under /groupfiles but they can add, edit, and delete under /groupfiles/project

The problem I'm having is that when I do this, I get the wrong group:

$touch test
$ls -l test
-rw-rw-r-- 1 myuser myuser 0 2012-03-26 20:58 test
$id -r -g -n
groupname

I need to make it so that the files created by vim, touch, etc have the correct group. I'm aware of newgrp() and the setgid bits. These are not what I am looking for. This works fine for the new users's I created, but doesn't work for my user. I'm not sure what happened, but it is working fine now. I've just leave this question here for future tinkerers.

I get 3 IP addresses from my ISP, so after my modem I have a switch with three things. Two are servers, one is a router. All my personal computers are behind the router. My old server used to have two ethernet ports, so I would put one on the switch and one on the router so that I could keep server traffic off the router and still be able to have samba shares for media.

My new lower-power server only has one LAN. I will eventually get a USB ethernet plug for it, but until then I'd like to still have my samba shares. I thought to myself, there must be some way to achieve this via VPN.

So what I'd like to do is this, set up a VPN server on the server, allow clients behind the router to connect to it. That seemed simple enough, except that all the guides I've seen assume you can allocate addresses on the server's LAN. Since my server's LAN is the public internet I can't really do that. Is there any way I can make up some 'imaginary' LAN that exists only in the mind of the server and put my VPN clients on that?

To clarify, if the server is 68.232.SSS.SSS and if the router is 68.232.RRR.RRR and if computers behind the router are 10.0.0.1-255 then can I make VPN clients use addresses like 192.168.0.VPN?

  OOoOoOOOooOOo
OoO  Public  OoOOo-----Server eth0 (Real)    68.232.SSS.SSS
OoO  Internet OoO             eth1 (Virtual) 192.168.0.1
 OoOoOOoOOoOOOo
     |
     \--Router eth0 (WAN) 68.232.RRR.RRR
         |     eth1 (LAN) 10.0.0.1
         |
         \-----Client eth1 (Virtual) 192.168.0.2  
                      eth0 (Real)    10.0.0.2

I just installed a system like this:

ubuntu@ubuntu:~$ ls -l /dev/disk/by-label/
total 0
lrwxrwxrwx 1 root root 10 2012-01-22 18:49 Boot -> ../../sda1
lrwxrwxrwx 1 root root 10 2012-01-22 18:49 ubuntu32 -> ../../sda2
lrwxrwxrwx 1 root root 10 2012-01-22 18:28 ubuntu64 -> ../../sda3
lrwxrwxrwx 1 root root 10 2012-01-22 18:49 Home -> ../../sda5

ubuntu@ubuntu:~$ ls -l /dev/disk/by-uuid/
total 0
lrwxrwxrwx 1 root root 10 2012-01-22 20:55 3582d70f-f4a5-484c-b14c-45cd740346b9 -> ../../sda1
lrwxrwxrwx 1 root root 10 2012-01-22 20:55 741182a8-3f15-4dfd-994d-654c8a57a9e4 -> ../../sda2
lrwxrwxrwx 1 root root 10 2012-01-22 20:55 1c415472-a770-4d76-be9f-27b8c1408e2a -> ../../sda3
lrwxrwxrwx 1 root root 10 2012-01-22 20:55 3515d523-72a2-4e04-b7da-cb6a1fd572ef -> ../../sda5
lrwxrwxrwx 1 root root 10 2012-01-22 20:55 f1f1cd7e-30cb-44e7-9ef6-986a589e0045 -> ../../sda6

I need 32 and 64 bit separate so I can test driver performance on each, the Home directory is shared and each version of ubuntu is mounted as /ubuntuXX under the other version. /boot also points to /dev/sda1 on both.

When I run sudo update-grub from ubuntu32 it runs fine but I get an error booting ubuntu64. init fails and I'm assuming it is because of the wrong bit type. Shouldn't grub's OS probe be bit-aware? How can I get these two to boot properly.

I ran grub-customizer from a live disk, chose /dev/sda3 as root, selected all partitions in the second step, and then removed all entires except for OS-Prober and memtest. What resulted is the attached grub.cfg. It now lists /dev/sda2 as the only OS option. The grub.cfg has the root UUID set to /boot for every entry.

grub.cfg: http://pastebin.com/Dkdxian4
fstab (both): http://pastebin.com/3sUabYRY

I know grub2 is all auto-generated and menus and jazz, but how do I just nuke all that and manually add entires (I'm not keeping this install long so there's no worry about kernel updates)

I took a shot at bilndly replicating the /dev/sda2 entry but adjusting it for /dev/sda3 however it didn't go well. I did however, manage to get the same error as my first attempt (it's the line right before the kernel panic, run-init)

updated part of grub.cfg: http://pastebin.com/DvfBhrBF

(null)
Begin: Running /scripts/local-bottom ... done.
done.
Begin: Running /scripts/init-bottom: ... done.
[    3.402989] request_module: runaway loop modprobe binfmt-464c
run-init: /sbin/init: Exec format error
[    3.406303] Kernel panic - not syncing: Attempted to kill init!
[    3.406394] Pid: 1, comm: run-init Not tainted 3.0.0-15-generic #25-Ubuntu
[    3.408290] [<c151a922>] ? printk+0x2d/0x2f
[    3.408338] [<c151a800>] panic+0x5c/0x151
[    3.408388] [<c104b564>] forget_original_parent+0x1e4/0x1f0
[    3.408440] [<c10d3a48>] ? perf_cgroup_attach_task+0x20/0x20
[    3.408489] [<c104b583>] exit_notify+0x13/0x140
[    3.408536] [<c104bd8d>] do_exit+0x1ad/0x3a8
[    3.408585] [<c1313850>] ? tty_write+0x228/0x228
[    3.408632] [<c104c098>] sys_exit+0x18/0x28
[    3.408680] [<c152e424>] syscall_call+0x7/0xb

I'm beginning to think it has to do with the kernel images that are in the /boot directory.

Ok, now I'm certain that it is due to the fact that one partition is 32-bit and one is 64-bit.

Figure 1: exert from /boot/grub/grub.cfg:

linux /vmlinuz-3.0.0-12-generic root=/dev/sda2
initrd /initrd.img-3.0.0-12-generic

Figure 2: Listings of /ubuntu64/vm* and /ubuntu32/vm*

me@GAMMA:~$ ls -l /vm*
lrwxrwxrwx 1 root root 29 2012-01-23 20:41 /vmlinuz -> boot/vmlinuz-3.0.0-15-generic
lrwxrwxrwx 1 root root 29 2012-01-22 13:05 /vmlinuz.old -> boot/vmlinuz-3.0.0-12-generic
me@GAMMA:~$ ls -l /ubuntu32/vm*
lrwxrwxrwx 1 root root 29 2012-01-22 12:41 /ubuntu32/vmlinuz -> boot/vmlinuz-3.0.0-15-generic
lrwxrwxrwx 1 root root 29 2012-01-22 12:22 /ubuntu32/vmlinuz.old -> boot/vmlinuz-3.0.0-12-generic

Figure 3: Magic File Type of /boot/vmlinuz-3.0.0-12-generic

/boot/vmlinuz-3.0.0-12-generic: Linux kernel x86 boot executable bzImage, version 3.0.0-12-generic (buildd@creste, RO-rootFS, root_dev 0x801, swap_dev 0x4, Normal VGA
/boot/vmlinuz-3.0.0-15-generic: Linux kernel x86 boot executable bzImage, version 3.0.0-15-generic (buildd@creste, RO-rootFS, root_dev 0x801, swap_dev 0x4, Normal VGA

Now here's the real kicker:

me@GAMMA:~$ sudo find / -name "vmlinuz*"
/boot/vmlinuz-3.0.0-12-generic
/boot/vmlinuz-3.0.0-15-generic
/vmlinuz.old
/vmlinuz
/ubuntu32/vmlinuz.old
/ubuntu32/vmlinuz

That's the only kernel on my system! How is that even possible? I'm running 64-bit right now (/dev/sda3 is / and uname reports 64-bit)!

I checked the package contents on packages.ubuntu.com and the am64 version of linux-image-3.0.0.15-generic lists /boot/vmlinuz-3.0.0-15-generic as a file, so I ran the following:

me@GAMMA:~$ sudo openssl dgst -md5 /boot/vmlinuz-3.0.0-15-generic > ~/3.0.0.15x86
me@GAMMA:~$ sudo apt-get install --reinstall linux-image-3.0.0-15-generic
(Output Omitted)
me@GAMMA:~$ sudo openssl dgst -md5 /boot/vmlinuz-3.0.0-15-generic
MD5(/boot/vmlinuz-3.0.0-15-generic)= f56839a4642eb97e06e5efb0bc74f4dc
me@GAMMA:~$ cat ~/3.0.0.15x86 
MD5(/boot/vmlinuz-3.0.0-15-generic)= f56839a4642eb97e06e5efb0bc74f4dc
me@GAMMA:~$ sudo openssl dgst -md5 /boot/vmlinuz-3.0.0-15-generic > ~/3.0.0.15x86

So the Linux kernel acts very much like the Mach Kernel in OSX in that it is a 32-bit executable that switch to 64-bit mode when needed. The question then is why am I getting an error about "Exec format error"?

This Post however seems to indicate that the runaway loop modprobe from the Kernel Panic indicates that this is indeed a 32/64 bit issue. Grub must have some way of telling the kernel the bitlength, maybe it is in an associated file in /boot, looking into that tomorrow.

quick update:
The 32-bit and 64-bit kernels hash differently, but file reports them both to be x86.

< MD5(/boot/vmlinuz-3.0.0-15-generic)= f56839a4642eb97e06e5efb0bc74f4dc
 ---
> MD5(/boot/vmlinuz-3.0.0-15-generic)= cee6cd7db9016ee8531be92504ac802b

So I need to determine how to install the kernel to somewhere other than /boot so that the two kernels don't clobber each other...

I like to play a specific variation of Solitare that AisleRiot doesn't seem to offer. I found a Single Card Draw version of Klondike, but it limits redeals and I found a Three Card Draw that has no limits on redeals. I like to play Single Card Draw with no limits on redeals. Is this possible with AisleRiot or do I need to go find another program?

 Draw-Style  Redeal-Limit Available?
    1              3           Yes
    1            none           No   -- This is what I want
    3              3            No
    3            none          Yes

I installed 32-bit linux because back in the day 64-bit linux missed out on some cool stuff like java and flash. Some of those things have been resolved (maybe all, I'm not as up-to-date as I was) and currently I'm having performance issues with the 32-bit versions of the ATI binary driver. The machine I'm on is 64-bit wide so I could switch to 64-bit, but I'd like to do so only if there is an advantage to it.

tl/dr;

Are 64-bit ATI binary drivers better/faster than their 32-bit counterparts?

UPDATE:

With the shift to 64-bit a lot of things change at the instruction and hardware levels. I would expect 64-bit drivers to be able to process more data in a shorter amount of time. I don't know if the types of operations that GPUs benefit in this way or not. My feeling is yes, but by how much I don't know. So I'd like to know if anyone has spec'd a GPU (specifically an ATI one, since that's what I have) using 32-bit and 64-bit drivers and seen a vast increase in speed on the 64-bit one.

I'm building a HTPC with this ASRock E350M1/USB3 motherboard. The onboard ATI Radeon HD 6310 graphics had a hard time swallowing a 1080p x264 mkv (~10 GB) so I popped in a discrete Nvidia GT 240 (1024MB). Oddly enough, that also wouldn't play it properly.

Both cards had the associated vendor drivers installed and active, I have tried x264 and raw blu-ray to no avail. I know the GT 240 is capable of playing 1080p movies, so what is the hangup?

Choppy refers to getting about one frame every few seconds, but I also see this from time to time: http://ptpimg.me/8yt8ev.png

Output from running vlc from the command line on a untouched Blu-ray (m2ts) rip.

me@GAMMA ~ $ vlc
VLC media player 1.1.9 The Luggage (revision exported)
Blocked: call to unsetenv("DBUS_ACTIVATION_ADDRESS")
Blocked: call to unsetenv("DBUS_ACTIVATION_BUS_TYPE")
[0x8d41914] main libvlc: Running vlc with the default interface. Use 'cvlc' to use vlc without interface.
Blocked: call to setlocale(6, "")
Warning: call to srand(1326832540)
Warning: call to rand()
Blocked: call to setlocale(6, "")

(process:2617): Gtk-WARNING **: Locale not supported by C library.
    Using the fallback 'C' locale.
Warning: call to rand()
Warning: call to rand()
Warning: call to rand()
Warning: call to rand()
libdvbpsi error (PSI decoder): TS discontinuity (received 1, expected 0) for PID 0
libdvbpsi error (PSI decoder): TS discontinuity (received 1, expected 0) for PID 0
libdvbpsi error (PSI decoder): TS discontinuity (received 3, expected 0) for PID 256
libdvbpsi error (PSI decoder): TS discontinuity (received 3, expected 0) for PID 256
[0x8ddc6e4] signals interface error: signal 17 overriden (0x3b114c0)
[0x8ddc6e4] signals interface error:  /usr/lib/libQtCore.so.4(?)[(nil)]
[0x8ddc6e4] signals interface error: signal 17 overriden (0x3b114c0)
[0x8ddc6e4] signals interface error:  /usr/lib/libQtCore.so.4(?)[(nil)]

Output from vlc -vvv on the same file: Here

OK, so we got the GT 240 working by using VD-PAU which is a magical thing that tells software like nplayer to use the GPU to do the work instead of the CPU. Problem is that VD-PAU is essentially Nvidia only and the built in graphics are ATI based.

(in ~/.mplayer/config)
vo=vdpau,xv
vc=ffh264vdpau,ffmpeg12vdpau,ffwmv3vdpau,ffvc1vdpau

I've done some digging and it seems VA-API and XVBA are the ATI equivalents. The question now is how do I get these to work in natty (preferentially with something like xbmc)

So far, with just the onboard Radeon HD 6310 graphics I've done this:

sudo apt-get install xvba-va-driver libva-glx1 vainfo
mplayer -vo xv -framedrop -ao sdl /path/to/000000.m2ts

and it is certainly getting close. Still too many dropped frames to be considered watchable however.