Filesystems are occasionally being shut down with this error: xfs_trans_ail_delete_bulk: attempting to delete a log item that is not in the AIL. It was diagnosed to be related to the EFI/EFD commit o
Filesystems are occasionally being shut down with this error: xfs_trans_ail_delete_bulk: attempting to delete a log item that is not in the AIL. It was diagnosed to be related to the EFI/EFD commit o
Well hang on, not everything can be cool in EFI-land here, if nothing else this: __xfs_efi_release( struct xfs_efi_log_item *efip) { struct xfs_ail *ailp = efip->efi_item.li_ailp; if (!test_and_clear
Or are you saying that it's working as designed, so that the first caller finds it set, clears it and does nothing, and the 2nd caller finds it unset, and frees it? help me out here ;) -Eric
On 04/03/13 16:02, Eric Sandeen wrote: On 4/3/13 2:46 PM, Eric Sandeen wrote: On 4/3/13 2:12 PM, Mark Tinguely wrote: On 04/02/13 22:09, Dave Chinner wrote: From: Dave Chinner<dchinner@xxxxxxxxxx> Fi
Speaking as the person whose designed and wrote the current EFI/EFD code in question here, I think it is broken and only working by chance. Hint: there is no comment anywhere (code or commit message)
Filesystems are occasionally being shut down with this error: xfs_trans_ail_delete_bulk: attempting to delete a log item that is not in the AIL. It was diagnosed to be related to the EFI/EFD commit o
I'm not sure that can actually happen - callback processing is supposed to be serialised by xlog_state_do_callback().... Thank you. I don't think the callbacks are being attached out of order. Like I