c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2026-05-22 18:40:43 +03:00
Code Issues Packages Projects Releases Wiki Activity

deadlock between mm_sem and tx assign in zfs_write() and page fault

Browse Source 
The bug time sequence:
1. thread #1, `zfs_write` assign a txg "n".
2. In a same process, thread #2, mmap page fault (which means the
   `mm_sem` is hold) occurred, `zfs_dirty_inode` open a txg failed,
   and wait previous txg "n" completed.
3. thread #1 call `uiomove` to write, however page fault is occurred
   in `uiomove`, which means it need `mm_sem`, but `mm_sem` is hold by
   thread #2, so it stuck and can't complete,  then txg "n" will
   not complete.

So thread #1 and thread #2 are deadlocked.

Reviewed-by: Chunwei Chen <tuxoko@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Signed-off-by: Grady Wong <grady.w@xtaotech.com>
Closes #7939
This commit is contained in:
ilbsmart
2018-10-17 02:11:24 +08:00
committed by Matthew Ahrens
parent b2030e5d51
commit 779a6c0bf6
6 changed files with 155 additions and 57 deletions
Download Patch File Download Diff File Expand all files Collapse all files
module/zfs/zfs_vnops.c
+21 -3
View File
@@ -650,7 +650,10 @@ zfs_write(struct inode *ip, uio_t *uio, int ioflag, cred_t *cr)
xuio = (xuio_t *)uio;
else
#endif
uio_prefaultpages(MIN(n, max_blksz), uio);
if (uio_prefaultpages(MIN(n, max_blksz), uio)) {
ZFS_EXIT(zfsvfs);
return (SET_ERROR(EFAULT));
}
/*
* If in append mode, set the io offset pointer to eof.
@@ -808,8 +811,19 @@ zfs_write(struct inode *ip, uio_t *uio, int ioflag, cred_t *cr)
ssize_t tx_bytes;
if (abuf == NULL) {
tx_bytes = uio->uio_resid;
uio->uio_fault_disable = B_TRUE;
error = dmu_write_uio_dbuf(sa_get_db(zp->z_sa_hdl),
uio, nbytes, tx);
if (error == EFAULT) {
dmu_tx_commit(tx);
if (uio_prefaultpages(MIN(n, max_blksz), uio)) {
break;
}
continue;
} else if (error != 0) {
dmu_tx_commit(tx);
break;
}
tx_bytes -= uio->uio_resid;
} else {
tx_bytes = nbytes;
@@ -909,8 +923,12 @@ zfs_write(struct inode *ip, uio_t *uio, int ioflag, cred_t *cr)
ASSERT(tx_bytes == nbytes);
n -= nbytes;
if (!xuio && n > 0)
uio_prefaultpages(MIN(n, max_blksz), uio);
if (!xuio && n > 0) {
if (uio_prefaultpages(MIN(n, max_blksz), uio)) {
error = EFAULT;
break;
}
}
}
zfs_inode_update(zp);