Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: amd64 -current crashs at boot

Andrew Doran wrote:
> Apologies for not returning to this sooner. I and my family have been quite
> sick over the last week.
> On Wed, Dec 24, 2008 at 01:43:37PM +0000, Patrick Welche wrote:
>> FSYNC_VFS seems to be used as
>>   vp->v_type == VBLK && vp->v_specmountpoint != NULL
>> Is there more to it? (eg was it meant to mean "this fs implements 
>> vop_fsync"?)
>> (ffs_update doesn't know about specnodes, but why should it?)
> There is more to it than that. FSYNC_VFS indicates that whole file system
> metadata needs flushing. It is passed only when the root vnode for the file
> system came from a non-ffs file system (zb tmpfs /dev). If the root vnode is
> on an ffs file system, ffs_full_fsync() flushes both the device vnode itself
> and metadata associated with the mounted-on file system.

sys/ufs/ffs/ffs_vnops.c , rev. 1.108 from christos fixed it for me.


Home | Main Index | Thread Index | Old Index