Bug 1020
Summary: | EFS isnt working yet in 2.6.0-test2 | ||
---|---|---|---|
Product: | File System | Reporter: | Travis Tilley (ttilley) |
Component: | Other | Assignee: | Randy Dunlap (rdunlap) |
Status: | REJECTED WILL_NOT_FIX | ||
Severity: | normal | CC: | bunk, okir, protasnb |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.0-test2 | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: | debug |
Description
Travis Tilley
2003-07-31 12:34:17 UTC
perhaps it's a block size thing? Could you possibly make available an EFS iso image? Ok i have just tried this with a cd and it provided the following error; EFS: device does not support 512 byte blocks via loopback is fine. This makes sense as the cdrom has a 2048byte block size, i'll dig a bit more. For the record, 2.4.22 does not work either, it has the same blocksize error message. I have a test patch (work in progress) for handling physical block sizes of 512, 1024, or 2048 bytes. The patch is available at: http://www.xenotime.net/linux/efs/efs_blocksize2.patch I can mount an EFS CDROM (block size is 2048) and read directories from it correctly, but reading file data from it is *NOT* working. Additionally, I found a program called 'efslook' for reading EFS filesystems. It was written for SGI MIPS (big-endian) IRIX, and it now works on Linux (should work big or little endian; I use it on x86). It is available at: http://www.xenotime.net/linux/efs/efslook_linux_v01.tgz Randy, what is the status of this issue? No progress. IIRC, Al Viro or some other fs people suggested just using mount on a loopback device, which seemed to work at the time (it was so long ago, sadly). Nobody's maintaining EFS anymore. What shall we do with this bug? Christoph Hellwig on lkml -- "I don't think it's a bug, it's a lacking feature. The error message is quite accurate aswell." I guess "use loopback" is the answer for this for now, though Christoph also mentioned he'd look at Randys patches. Created attachment 11907 [details]
debug
(I don't suppose the #10 was tring to help in any way to the bug resolution ;) unless this was an accident) #5 - Is this a material for a small project that can be posted on http://kernelnewbies.org/KernelProjects? unless Randy is intending to keep working on it? Thanks. This can certainly be added to the KernelProjects web page, although IIRC, Al Viro just prefers to use mount via loop as a solution for this one. I am closing this bug, and will transfer this to the "features to implement" list (cleating one...). Probably not of hight priority rather "good to have". |