Hi Eddy, CC'ing this to the xfs list. I don't recommend you pull the usb disk out while the filesystem is still mounted (if you can avoid it). Okay good. Looks like whatever the problem is it has bee
Doing that to emulate power loss scenario our device might experience. Not quite familiar with XFS code :( I'll try to bisect (It will take some time...) Because XFS log is OK to 2.6.28 system, whic
Unless you have write barriers enabled and passing through to the USB device, or write cache turned off on the USB device, this test may well just be corrupting the log to an unreplayable state. Alth
We all start somewhere! Easy? Not exactly. Diff'ing 2.6.10 version of fs/xfs/xfs_log_recover.c with mainline shows over 300 differences. If you really want to be able to replay the log on a different
Hello Lachlan: Yes. I'll start from this problem :) The two systems are all 32bit. Enlightened by your suggestion, I find problem lies in the two systems are different architecture. One is arm, on is
Well, the fact that you have 2 arches would have been highly relevant in your first email. ;) Is this the old abi or the new? Guessing old, based on the alignment issues. Well, packing structures lik
Hello Eric: ARM compiler is producing eabi. The problem lies in: "long long" on arm is 8 bytes aligned, on x86 is 4 bytes aligned. So 4 more bytes are padded before xfs_inode_log_format::ilf_ino on a
A tool that can be very helpful here is "pahole" found in the dwarves package: struct xfs_inode_log_format { __uint16_t ilf_type; /* 0 2 */ __uint16_t ilf_size; /* 2 2 */ __uint32_t ilf_fields; /* 4
Hello Eric: Nice tool. I don't have x86_64 machine at hand. Replay arm eabi log on ia32 machine "seems" success (mount "successfully" returns). Actually I guess that's because 2.6.28 XFS running on
Forget this. There's no misinterpretation. Should be "Replay arm eabi log on ia32 machine success (mount successfully returns). 2.6.28 XFS running on ia32 interprets unpacked xfs_inode_log_forma
... Well, to be honest, the restriction of a clean log when migrating an xfs filesystem between systems of differing endianness was design decision. And xfs userspace can't replay any log, native end