In accordance with the following bug reference I am requesting that the ~amd64 keyword be added to the truecrypt ebuild: https://bugs.gentoo.org/show_bug.cgi?id=112197#c71 I have done somewhat thorough testing of the new ebuild on amd64 with zero problems. Please let me know if you'd like any specific tests/results. NOTE: Daniel Black (Gentoo Dev Dragonheart) requested that this bug be assigned to crypto@gentoo.org. Thanks much for your time!
Adding crypto@gentoo.org to the CC list.
Greg if you can provide a rough test plan normally that is appreciated.
No problem at all. My testing including the following (roughly): 1) I Un-merged my existing test ebuilds of TrueCrypt and removed it from my custom overlay. I then added the ~amd64 keyword to the real ebuild and emerged the new version. 2) I did some modprobing and removing of the truecrypt kernel module. No problems found. 3) I performed a litany of tests with the truecrypt command-line client: - I created a new encrypted volume (container) using various encryption settings. - I created a ReiserFS filesystem inside of the encrypted volume. - I did tests creating and writing to files inside of the encrypted volume. - I mounted and unmounted the volume repeatedly. I also ensured that the file contents were normal, non-corrupt, and accessible without errors across successive mounts of the volume. All operations worked normally and without error (either at the CLI level or in the kernel ring-buffer). Let me know if you'd like anything more specific. Thanks!
good summary. Needs a touch more detail. Arch testers generally are unfamiliar with the workings of truecrypt so specific commands (e.g. how to create truecrypt container?) and expected output will help them.
Understood. Following is a corresponding summary of commands performed. Numbers indicate reference to my previously posted (rough) test plan in comment #3 : 2) # modprobe truecrypt # modprobe -r truecrypt (Both followed by lsmod and dmesg for a ring-buffer check). 3. A) Create encrypted volume (container): # truecrypt --type normal --filesystem none --size 1000M --encryption AES --hash whirlpool -c test-vol.tc B) Map new volume (to device 2 for testing) and ensure that it is online: # truecrypt --device-number 2 test-vol.tc # truecrypt -l Create a ReiserFS filesystem within the mapped volume: # mkreiserfs /dev/mapper/truecrypt2 C) Mount the new filesystem within the encrypted volume and ensure it is mounted: # mount -t reiserfs /dev/mapper/truecrypt2 /mnt/temp # mount Create a test files within the volume, something like this: # cd /mnt/temp # vim testfile.txt (add content for later verification of non-corruption) # cp -Rp /home/someuser/somedir ./ # cd /mnt D) Umount/unmap and mount/map the encrypted volume a few times to ensure truecrypt is working smoothly: # truecrypt -d /dev/mapper/truecrypt2 # truecrypt -l # (truecrypt2 should no longer be listed. Kernel module may have unloaded if no other mounted volumes were present.) # modprobe truecrypt # truecrypt --device-number 3 --filesystem reiserfs test-vol.tc /mnt/temp # mount # (ensure filesystem is up) Validate files inside of encrypted volume: # cd /mnt/temp # cat * # (verify content as desired)
Any word on this? I've done a VERY LARGE amount of arch testing for amd64 on this, with zero problems to report. Can't we get this keyword added?
BUMP: This has been sitting for almost a month now with no progress. I've done further (heavy) testing on the amd64 arch since submitting this with no new issues to report. Please add the appropriate keyword to this package for us amd64 folks. Thanks...
No problems whatsover, found on my amd64 platform tested with images on 2 usbsticks tested with complete devices tested with truecrypt systems generated on windows and tested on linux and vice versa. I'm all for adding the amd64 keyword :)
Tested on amd64, compiles and works as expected (comment #5).
Created attachment 94861 [details] emerge.info emerge --info
done