Bug 497568

Summary: RFE: livecd-tools should make images that default max_loop=14 or higher
Product: [Fedora] Fedora Reporter: Paul W. Frields <stickster>
Component: livecd-toolsAssignee: Jeremy Katz <katzj>
Status: CLOSED NOTABUG QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: medium Docs Contact:
Priority: low    
Version: rawhideCC: davidz, katzj
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2009-04-24 23:41:05 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Paul W. Frields 2009-04-24 18:54:25 UTC 
The livecd-tools build images that don't change the default behavior of the "loop" block device module.  However, since the environment itself consumes /dev/loop0 through /dev/loop5, it only leaves two loop devices available for use.  Adding 'max_loop=14' or higher would be a friendlier way and would provide the principle of "least surprise" for the user/admin.

Keep in mind that of course the loop module can't be removed/reinserted since it's in use by the environment! :-)

livecd-tools-023-1.fc11.x86_64
Comment 1 Jeremy Katz 2009-04-24 23:41:05 UTC 
Once upon a time, we did.  Then we changed the kernel default.  Now we don't do anything special because
a) loop is built-in, not modular
b) the kernel automatically starts to allocate more loop devices for you as you need them
Comment 2 Paul W. Frields 2009-04-25 16:42:13 UTC 
I understand (a), sorry I missed that important point.  But as for (b), are you sure?  I just tried this in a Rawhide VM guest, and I stop getting loop devices (using 'losetup -f <file>') after the eight pre-created devices /dev/loop{0..7} are exhausted.

I'd be tempted to reopen this if you find this isn't another PEBCAK problem.
Comment 3 Paul W. Frields 2009-04-25 16:43:51 UTC 
Erratum: I guess it wouldn't be reopening as much as opening a new bug to the kernel if autocreation of /dev/loop* is not happening as expected.
Comment 4 Jeremy Katz 2009-04-27 15:39:53 UTC 
Upstream support for dynamic loop allocation is commit http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=a47653fc2643cf61bcabba8c9ff5c45517c089ba

And from the try I did on Friday with lots of losetup -f
[katzj@erebor ~]$ sudo losetup -a
/dev/loop0: [fd03]:892295 (/var/tmp/imgcreate-2u4B_K/tmp-T4locA/ext3fs.img)
/dev/loop1: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop2: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop3: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop4: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop5: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop6: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop7: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop8: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop80: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop81: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop82: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
/dev/loop83: [fd01]:42401853 (/home/live/live/rawhide-20090423/rawhide-i686-20090423.0.iso)
Comment 5 Paul W. Frields 2009-04-27 21:20:42 UTC 
Ah, according to that commit, you still have to mknod manually, but the new device will be supported on demand.  Too bad the all-in-one process broke things, it would have been nice to be a little more automatic.  Thanks for the info though!