Bug#732117: loop no longer needs explicit device?
On Sat, Feb 08, 2014 at 10:45:41PM -0500, Phillip Susi wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA512 On 02/05/2014 05:14 PM, ael wrote: The exact behaviour has changed with different kernels. I am running on a (locally compiled) 3.14.0-rc1 just now - and with a fully updated testing environment. On this kernel -o loop=/dev/loop0 just gives the message: mount: stolen loop=/dev/loop0 (and fails). I just ran a freshly built 3.14-rc1 kernel built from vanilla upstream source and an up to date jessie/testing system and it works fine for me. What exactly do you mean it fails? Ok. That is interesting. There must be something peculiar on the (rather old i386) machine on which I test. I am not on that machine just now, but I am am confident that something like mount -o loop=/dev/loop1 file.iso /mountpoint gives the message stolen loop. Looking at mount or df -h /mountpoint shows nothing mounted. mount -o loop file.iso /mountpoint succeeds. That was on 3.14-rc1. There is no (other) explicit error message as on earlier kernels. Maybe I have something odd in my .config, but that seems unlikely. I believe that I used a loop module parameter to increase the number of loop devices. I will try and remember to check that. I guess that I should experiment with changing max_loop just in case that is triggering a bug. I will be away from the relevant machine for perhaps a month soon, but will try to find to to make the test before I leave. Otherwise there may be some delay... ael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732117: loop no longer needs explicit device?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA512 On 02/05/2014 05:14 PM, ael wrote: The exact behaviour has changed with different kernels. I am running on a (locally compiled) 3.14.0-rc1 just now - and with a fully updated testing environment. On this kernel -o loop=/dev/loop0 just gives the message: mount: stolen loop=/dev/loop0 (and fails). I just ran a freshly built 3.14-rc1 kernel built from vanilla upstream source and an up to date jessie/testing system and it works fine for me. What exactly do you mean it fails? -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.14 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBCgAGBQJS9vnlAAoJEI5FoCIzSKrwGAMH/RyBccT5hiZNraQqcVBYfqvE vmijxN1zn2rnuOPX+JhN34fq2MQY8EtNOC+wODbIh87cQm+wv6sEXVtQ7j1dhy9g urdPtohKhf1PPZOMwBr1yrV4xgpXodwYXreEaBJeWk84D1fPWakHxhfxvXc8jsT9 +RSGVnngDyga40pl+xymXory3Y/Oso/6wUrdMWpVtEZChXsTGpQwyUPEBRcTS2w3 Exx02HXTG7YrFvk5XqKAzomyop0nsz3GaL47VhFj5DsSpuKc6WIjmkywC/4GlqOB zIZ7HqdwpjHk/Yn0+vznrnBQiq6kh+upB+g7tSQ8B6VNn8h+E26/FmLt+QlgWMs= =1HjC -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732117: loop no longer needs explicit device?
On Tue, Feb 04, 2014 at 04:11:16PM -0500, Phillip Susi wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 2/4/2014 3:51 PM, ael wrote: If just -o loop rather than -o loop=/dev/loop[1..n] is used, the problem seems to vanish, at least with the latest kernels. Thus this becomes a bug in the mount man page? Hence reasssigning back Ahh... strange.. so what is wrong exactly with the man page? THE LOOP DEVICE One further possible type is a mount via the loop device. For example, the command mount /tmp/disk.img /mnt -t vfat -o loop=/dev/loop will set up the loop device /dev/loop3 to correspond to the file /tmp/disk.img, and then mount this device on /mnt. Remove explict device /dev/loop.. from the example. Slightly modify following text. I haven't searched the kernel archives (if that is where to look) to see when this mandatory automatic handling of the loop devices was introduced. Presumably the motivation was to reduce the burdon on the user of keeping track. Perhaps. ael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732117: loop no longer needs explicit device?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 2/5/2014 3:53 AM, ael wrote: Remove explict device /dev/loop.. from the example. Slightly modify following text. I haven't searched the kernel archives (if that is where to look) to see when this mandatory automatic handling of the loop devices was introduced. Presumably the motivation was to reduce the burdon on the user of keeping track. Perhaps. It works fine for me. I also don't see how it can be connected to the mount() sys call failing in your strace. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJS8kb0AAoJEI5FoCIzSKrwBbkH/R4nxpb1UKB+8sVWYi0a4oPo XKqPft/DKo3curX8RBJZdl/g+AUGwRywlaBOslf9Q1ERLTIUlLnl29XVBNr22Uv/ j54qVTKAFjdwFrWhPfK5zgg5LRKC9WH2zQjmTDGX0BvBoy96TZ0RMncFbUCXNPcq stXUcKlve7Em8valCVgEEwt57zyHUYJxDDOQmTtl46YYvvDT7TCRNx1FwtdteERa PhMM3BjL5ZpoWFSDsa6eGmVaqryDhunYwE0aBi56Uz/GnAPKPqO7ijxN/hPUxTxn Udr3XnnqnjrcioYGXoIFl3MBTJcCNU3mWMVubxTCr40CCDw+E3QDSFEl7RTZT18= =9T+c -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732117: loop no longer needs explicit device?
On Wed, Feb 05, 2014 at 09:13:10AM -0500, Phillip Susi wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 2/5/2014 3:53 AM, ael wrote: Remove explict device /dev/loop.. from the example. Slightly modify following text. I haven't searched the kernel archives (if that is where to look) to see when this mandatory automatic handling of the loop devices was introduced. Presumably the motivation was to reduce the burdon on the user of keeping track. Perhaps. It works fine for me. I also don't see how it can be connected to the mount() sys call failing in your strace. The exact behaviour has changed with different kernels. I am running on a (locally compiled) 3.14.0-rc1 just now - and with a fully updated testing environment. On this kernel -o loop=/dev/loop0 just gives the message: mount: stolen loop=/dev/loop0 (and fails). -o loop succeeds. On (I think) 3.13.0-rc4, -o loop suceeded but was more noisy about it. I don't pretend to understand what is happening, but using -o loop with an explicit device like /dev/loop2 has always failed recently. ael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732117: loop no longer needs explicit device?
If just -o loop rather than -o loop=/dev/loop[1..n] is used, the problem seems to vanish, at least with the latest kernels. Thus this becomes a bug in the mount man page? Hence reasssigning back ... ael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#732117: loop no longer needs explicit device?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 2/4/2014 3:51 PM, ael wrote: If just -o loop rather than -o loop=/dev/loop[1..n] is used, the problem seems to vanish, at least with the latest kernels. Thus this becomes a bug in the mount man page? Hence reasssigning back Ahh... strange.. so what is wrong exactly with the man page? -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJS8Vd0AAoJEI5FoCIzSKrwRQMH/29JdEVbhmox5+6AIG+spcg9 2+aeR9LJlgPut2lHvjbQvzi5YaEh0FrbXhGRU/BG8CQp7Leid9JVn0EsVn7E9Xqo euhZr6ZlcJH6XDMLMD3vyeGAldwjqdbnb7vjse8gJw0lGpdy6oFjfoxKWgt6N2AJ wDFMJSgs7O5J0d1U0fKZVIsZYGS0HboVwczxx3L/1JwqL1hAF74YfA5l9qpp4vdl nPpWdGL62mZ39ScGMefncPHV582XTDwbGSzyTfeA8O5Yx0XElC2xXC07nn/2+mNr f7/T2WWn29fM9Pow7OjNssV7uw/xa5kBGZS8f/yboqtqHTur9m1Jb50LmNAFJww= =e5PX -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org