nfs: add 'noalignwrite' option for lock-less 'lost writes' prevention
Subsystems (2)
PATCH 1/1 nfs: add 'noalignwrite' option for lock-less 'lost writes' prevention
There are some applications that write to predefined non-overlapping file offsets from multiple clients and therefore don't need to rely on file locking. However, if these applications want non-aligned offsets and sizes they need to either use locks or risk data corruption, as the NFS client defaults to extending writes to whole pages.
This commit adds a new mount option noalignwrite, which allows to turn that off and avoid the need of locking, as long as these applications don't overlap on offsets.
--- fs/nfs/fs_context.c | 8 ++++++++ fs/nfs/super.c | 3 +++ fs/nfs/write.c | 3 +++ include/linux/nfs_fs_sb.h | 1 + 4 files changed, 15 insertions(+)
fs/nfs/fs_context.c +8 -0
@@ -49,6 +49,7 @@ enum nfs_param { Opt_bsize, Opt_clientaddr, Opt_cto, + Opt_alignwrite, Opt_fg, Opt_fscache, Opt_fscache_flag,
@@ -149,6 +150,7 @@ static const struct fs_parameter_spec nfs_fs_parameters[] = {fsparam_u32 ("bsize" , Opt_bsize),fsparam_string ("clientaddr" , Opt_clientaddr),fsparam_flag_no ("cto" , Opt_cto), +fsparam_flag_no ("alignwrite" , Opt_alignwrite),fsparam_flag ("fg" , Opt_fg),fsparam_flag_no ("fsc" , Opt_fscache_flag), fsparam_string("fsc", Opt_fscache),
@@ -592,6 +594,12 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,else ctx->flags |=NFS_MOUNT_TRUNK_DISCOVERY ;break ; +case Opt_alignwrite: +if (result.negated) + ctx->flags |=NFS_MOUNT_NO_ALIGNWRITE ; +else + ctx->flags &= ~NFS_MOUNT_NO_ALIGNWRITE ; +break ;case Opt_ac:if (result.negated) ctx->flags |=NFS_MOUNT_NOAC ;
fs/nfs/super.c +3 -0
@@ -549,6 +549,9 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss,else seq_puts (m,",local_lock=posix" ); +if (nfss->flags &NFS_MOUNT_NO_ALIGNWRITE ) +seq_puts (m,",noalignwrite" ); +if (nfss->flags &NFS_MOUNT_WRITE_EAGER ) {if (nfss->flags &NFS_MOUNT_WRITE_WAIT )seq_puts (m,",write=wait" );
fs/nfs/write.c +3 -0
@@ -1315,7 +1315,10 @@ static int nfs_can_extend_write(struct file *file, struct folio *folio,struct file_lock_context *flctx =locks_inode_context (inode);struct file_lock *fl;int ret; +unsigned int mntflags =NFS_SERVER (inode)->flags; +if (mntflags &NFS_MOUNT_NO_ALIGNWRITE ) +return 0 ;if (file->f_flags &O_DSYNC )return 0 ;if (!nfs_folio_write_uptodate (folio, pagelen))
include/linux/nfs_fs_sb.h +1 -0
@@ -157,6 +157,7 @@ struct nfs_server {#define NFS_MOUNT_WRITE_WAIT 0x02000000#define NFS_MOUNT_TRUNK_DISCOVERY 0x04000000#define NFS_MOUNT_SHUTDOWN 0x08000000 +#define NFS_MOUNT_NO_ALIGNWRITE 0x10000000unsigned int fattr_valid;/* Valid attributes */ unsigned int caps;/* server capabilities */
Signature · 1 line
2.39.3
Quoted 97 lines · Dan Aloni
There are some applications that write to predefined non-overlapping file offsets from multiple clients and therefore don't need to rely on
file locking. However, if these applications want non-aligned offsets and sizes they need to either use locks or risk data corruption, as the
NFS client defaults to extending writes to whole pages.This commit adds a new mount option
noalignwrite, which allows to turn
that off and avoid the need of locking, as long as these applications don't overlap on offsets.Signed-off-by Dan Aloni <[email protected]>---
fs/nfs/fs_context.c | 8 ++++++++
fs/nfs/super.c | 3 +++
fs/nfs/write.c | 3 +++
include/linux/nfs_fs_sb.h | 1 +4 files changed, 15 insertions(+)fs/nfs/fs_context.c +2
-0@@ -49,6 +49,7 @@ enum nfs_param { Opt_bsize, Opt_clientaddr, Opt_cto, + Opt_alignwrite, Opt_fg, Opt_fscache, Opt_fscache_flag,@@ -149,6 +150,7 @@ static const struct fs_parameter_spec nfs_fs_parameters[] = { fsparam_u32 ("bsize", Opt_bsize), fsparam_string("clientaddr", Opt_clientaddr), fsparam_flag_no("cto", Opt_cto), + fsparam_flag_no("alignwrite", Opt_alignwrite), fsparam_flag ("fg", Opt_fg), fsparam_flag_no("fsc", Opt_fscache_flag),
fsparam_string("fsc", Opt_fscache),fs/nfs/fs_context.c +0
-0@@ -592,6 +594,12 @@ static int nfs_fs_context_parse_param(struct fs_context *fc,else
ctx->flags|=NFS_MOUNT_TRUNK_DISCOVERY; break;fs/nfs/fs_context.c +6
-0+case Opt_alignwrite: +if (result.negated) + ctx->flags |=NFS_MOUNT_NO_ALIGNWRITE ; +else + ctx->flags &= ~NFS_MOUNT_NO_ALIGNWRITE ; +break ;case
Opt_ac:
if (result.negated)ctx->flags|=NFS_MOUNT_NOAC;fs/nfs/super.c +0
-0@@ -549,6 +549,9 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss,else
seq_puts(m, ",local_lock=posix");fs/nfs/super.c +3
-0+if (nfss->flags &NFS_MOUNT_NO_ALIGNWRITE ) +seq_puts (m,",noalignwrite" ); +if (
nfss->flags&NFS_MOUNT_WRITE_EAGER) { if (nfss->flags&NFS_MOUNT_WRITE_WAIT)seq_puts(m, ",write=wait");fs/nfs/write.c +0
-0@@ -1315,7 +1315,10 @@ static int nfs_can_extend_write(struct file *file, struct folio *folio,struct
file_lock_context*flctx =locks_inode_context(inode);
structfile_lock*fl;
int ret;fs/nfs/write.c +1
-0+unsigned int mntflags =NFS_SERVER (inode)->flags;fs/nfs/write.c +2
-0+if (mntflags &NFS_MOUNT_NO_ALIGNWRITE ) +return 0 ;if (
file->f_flags&O_DSYNC)
return 0;
if (!nfs_folio_write_uptodate(folio, pagelen))include/linux/nfs_fs_sb.h +1
-0@@ -157,6 +157,7 @@ struct nfs_server { #define NFS_MOUNT_WRITE_WAIT 0x02000000 #define NFS_MOUNT_TRUNK_DISCOVERY 0x04000000 #define NFS_MOUNT_SHUTDOWN 0x08000000 +#define NFS_MOUNT_NO_ALIGNWRITE 0x10000000
Quoted 5 lines
unsigned intfattr_valid; /* Valid attributes
*/
unsigned int caps; /* server
capabilities */
While I hate that we need a new mount option for this, I do understand the need. I do wonder if we'd be better served by a new fadvise64() flag or something. Either way, this would probably be needed in the interim, so...
Quoted 11 lines · Dan Aloni
There are some applications that write to predefined non-overlapping file offsets from multiple clients and therefore don't need to rely on file locking. However, if these applications want non-aligned offsets and sizes they need to either use locks or risk data corruption, as the NFS client defaults to extending writes to whole pages.
This commit adds a new mount option
noalignwrite, which allows to turn that off and avoid the need of locking, as long as these applications don't overlap on offsets.Signed-off-by Dan Aloni <[email protected]>
The new name seems a better choice.
I'm also on Christoph's camp that having this behavior be the default makes sense given that we have large folios.
Trond, Anna,
What are your thoughts on this?
Quoted 19 lines · Sagi Grimberg
Quoted 11 lines · Dan Aloni
There are some applications that write to predefined non-overlapping file offsets from multiple clients and therefore don't need to rely on file locking. However, if these applications want non-aligned offsets and sizes they need to either use locks or risk data corruption, as the NFS client defaults to extending writes to whole pages.
This commit adds a new mount option
noalignwrite, which allows to turn that off and avoid the need of locking, as long as these applications don't overlap on offsets.Signed-off-by Dan Aloni <[email protected]>The new name seems a better choice.
I'm also on Christoph's camp that having this behavior be the default makes sense given that we have large folios.
Quoted 5 lines
Reviewed-by Sagi Grimberg <[email protected]>Trond, Anna,
What are your thoughts on this?
Pinging. Do we need to do anything further? I can prepare the nfs-utils patch too.
Signature · 1 line
Dan Aloni