On 4/19/13 9:03 PM, 符永涛 wrote: > Hi Eric, > I will enable them and run test again. I can only reproduce it with > glusterfs rebalance. Glusterfs uses a mechanism it called syncop to > unlink file. For rebalance it uses > syncop_unlink(glusterfs/libglusterfs/src/syncop.c). In the glusterfs > sync_task framework(glusterfs/libglusterfs/src/syncop.c) it uses > "makecontext/swapcontext" > <http://www.opengroup.org/onlinepubs/009695399/functions/makecontext.html>. > Does it leads to racing unlink from different CPU core? Yep, I understand that it's rebalance. It dies when rebalance finishes because an open but unlinked file trips over the corrupted list from earlier, it seems. I don't know why makecontext would matter... Just to be sure, you are definitely loading the xfs module from the kernel you built, right, and you don't have a "priority" module getting loaded from elsewhere? Seems unlikely, but just to be sure. > Thank you. You could also add this patch to the xfs tracepoints to print more information about the inodes - the mode & flags. -Eric diff --git a/fs/xfs/linux-2.6/xfs_trace.h b/fs/xfs/linux-2.6/xfs_trace.h index e8ce644..c314b87 100644 --- a/fs/xfs/linux-2.6/xfs_trace.h +++ b/fs/xfs/linux-2.6/xfs_trace.h @@ -544,14 +544,18 @@ DECLARE_EVENT_CLASS(xfs_inode_class, TP_STRUCT__entry( __field(dev_t, dev) __field(xfs_ino_t, ino) + __field(__u16, mode) + __field(unsigned long, flags) ), TP_fast_assign( __entry->dev = VFS_I(ip)->i_sb->s_dev; __entry->ino = ip->i_ino; + __entry->mode = VFS_I(ip)->i_mode; + __entry->flags = ip->i_flags; ), - TP_printk("dev %d:%d ino 0x%llx", + TP_printk("dev %d:%d ino 0x%llx mode 0%o, flags 0x%lx", MAJOR(__entry->dev), MINOR(__entry->dev), - __entry->ino) + __entry->ino, __entry->mode, __entry->flags) ) #define DEFINE_INODE_EVENT(name) \ _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
