Discard is a very big hammer. If it succeeds, there's no undo
or going back; on the other hand, on some devices the giant discard
issued at mkfs time makes them go offline. :(
Although I'm reluctant to waffle back and forth on default behavior,
I think this disabling it by default the prudent thing to do.
This patch also adds the discard option into the default mke2fs.conf
file; although it could be set, it was not in the template.
Signed-off-by: Eric Sandeen <[email protected]>
---
If disabling by default isn't desired, we should at least include
the mke2fs.conf.in change (set to 1).
Thanks,
-Eric
diff --git a/misc/mke2fs.8.in b/misc/mke2fs.8.in
index 023ba49..a005eca 100644
--- a/misc/mke2fs.8.in
+++ b/misc/mke2fs.8.in
@@ -277,11 +277,11 @@ Attempt to discard blocks at mkfs time (discarding blocks initially is useful
on solid state devices and sparse / thin-provisioned storage). When the device
advertises that discard also zeroes data (any subsequent read after the discard
and before write returns zero), then mark all not-yet-zeroed inode tables as
-zeroed. This significantly speeds up filesystem initialization. This is set
-as default.
+zeroed. This significantly speeds up filesystem initialization. This is disabled
+by default.
.TP
.BI nodiscard
-Do not attempt to discard blocks at mkfs time.
+Do not attempt to discard blocks at mkfs time. This is the default.
.TP
.BI quotatype
Specify which quota type ('usr' or 'grp') is to be initialized. This
diff --git a/misc/mke2fs.c b/misc/mke2fs.c
index bbf477a..d723ee7 100644
--- a/misc/mke2fs.c
+++ b/misc/mke2fs.c
@@ -84,7 +84,7 @@ int cflag;
int verbose;
int quiet;
int super_only;
-int discard = 1; /* attempt to discard device before fs creation */
+int discard; /* attempt to discard device before fs creation? */
int direct_io;
int force;
int noaction;
diff --git a/misc/mke2fs.conf.in b/misc/mke2fs.conf.in
index 178733f..db6decb 100644
--- a/misc/mke2fs.conf.in
+++ b/misc/mke2fs.conf.in
@@ -2,6 +2,7 @@
base_features = sparse_super,filetype,resize_inode,dir_index,ext_attr
default_mntopts = acl,user_xattr
enable_periodic_fsck = 0
+ discard = 0
blocksize = 4096
inode_size = 256
inode_ratio = 16384
On 4/17/13 3:04 PM, Eric Sandeen wrote:
> Discard is a very big hammer. If it succeeds, there's no undo
> or going back; on the other hand, on some devices the giant discard
> issued at mkfs time makes them go offline. :(
>
> Although I'm reluctant to waffle back and forth on default behavior,
> I think this disabling it by default the prudent thing to do.
>
> This patch also adds the discard option into the default mke2fs.conf
> file; although it could be set, it was not in the template.
>
> Signed-off-by: Eric Sandeen <[email protected]>
Sorry, scratch this; it won't work if, say, discard=1 in the conf
file and the user specifies -E nodiscard or -K; it'll turn it back
on after those options are processed, I think. I'll send another
patch later when I'm not firing it off too quickly from a conference. ;)
-Eric
> ---
>
> If disabling by default isn't desired, we should at least include
> the mke2fs.conf.in change (set to 1).
>
> Thanks,
> -Eric
>
> diff --git a/misc/mke2fs.8.in b/misc/mke2fs.8.in
> index 023ba49..a005eca 100644
> --- a/misc/mke2fs.8.in
> +++ b/misc/mke2fs.8.in
> @@ -277,11 +277,11 @@ Attempt to discard blocks at mkfs time (discarding blocks initially is useful
> on solid state devices and sparse / thin-provisioned storage). When the device
> advertises that discard also zeroes data (any subsequent read after the discard
> and before write returns zero), then mark all not-yet-zeroed inode tables as
> -zeroed. This significantly speeds up filesystem initialization. This is set
> -as default.
> +zeroed. This significantly speeds up filesystem initialization. This is disabled
> +by default.
> .TP
> .BI nodiscard
> -Do not attempt to discard blocks at mkfs time.
> +Do not attempt to discard blocks at mkfs time. This is the default.
> .TP
> .BI quotatype
> Specify which quota type ('usr' or 'grp') is to be initialized. This
> diff --git a/misc/mke2fs.c b/misc/mke2fs.c
> index bbf477a..d723ee7 100644
> --- a/misc/mke2fs.c
> +++ b/misc/mke2fs.c
> @@ -84,7 +84,7 @@ int cflag;
> int verbose;
> int quiet;
> int super_only;
> -int discard = 1; /* attempt to discard device before fs creation */
> +int discard; /* attempt to discard device before fs creation? */
> int direct_io;
> int force;
> int noaction;
> diff --git a/misc/mke2fs.conf.in b/misc/mke2fs.conf.in
> index 178733f..db6decb 100644
> --- a/misc/mke2fs.conf.in
> +++ b/misc/mke2fs.conf.in
> @@ -2,6 +2,7 @@
> base_features = sparse_super,filetype,resize_inode,dir_index,ext_attr
> default_mntopts = acl,user_xattr
> enable_periodic_fsck = 0
> + discard = 0
> blocksize = 4096
> inode_size = 256
> inode_ratio = 16384
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
Discard is a very big hammer. If it succeeds, there's no undo
or going back; on the other hand, on some devices the giant discard
issued at mkfs time makes them go offline.
Although I'm reluctant to waffle back and forth on default behavior,
I think this disabling it by default the prudent thing to do.
This patch also adds the discard option into the default mke2fs.conf
file; although it could be set, it was not in the template.
Signed-off-by: Eric Sandeen <[email protected]>
---
V2: handle discard disable via "-K" since it's parsed *before* we get
the default. The extended options are parsed later, and will override
the mke2fs.conf default without problems.
diff --git a/misc/mke2fs.8.in b/misc/mke2fs.8.in
index 023ba49..a005eca 100644
--- a/misc/mke2fs.8.in
+++ b/misc/mke2fs.8.in
@@ -277,11 +277,11 @@ Attempt to discard blocks at mkfs time (discarding blocks initially is useful
on solid state devices and sparse / thin-provisioned storage). When the device
advertises that discard also zeroes data (any subsequent read after the discard
and before write returns zero), then mark all not-yet-zeroed inode tables as
-zeroed. This significantly speeds up filesystem initialization. This is set
-as default.
+zeroed. This significantly speeds up filesystem initialization. This is disabled
+by default.
.TP
.BI nodiscard
-Do not attempt to discard blocks at mkfs time.
+Do not attempt to discard blocks at mkfs time. This is the default.
.TP
.BI quotatype
Specify which quota type ('usr' or 'grp') is to be initialized. This
diff --git a/misc/mke2fs.c b/misc/mke2fs.c
index bbf477a..fa11da6 100644
--- a/misc/mke2fs.c
+++ b/misc/mke2fs.c
@@ -84,7 +84,7 @@ int cflag;
int verbose;
int quiet;
int super_only;
-int discard = 1; /* attempt to discard device before fs creation */
+int discard; /* attempt to discard device before fs creation? */
int direct_io;
int force;
int noaction;
@@ -1266,6 +1266,7 @@ static void PRS(int argc, char *argv[])
double reserved_ratio = -1.0;
int lsector_size = 0, psector_size = 0;
int show_version_only = 0;
+ int nodiscard = 0;
unsigned long long num_inodes = 0; /* unsigned long long to catch too-large input */
errcode_t retval;
char * oldpath = getenv("PATH");
@@ -1444,7 +1445,7 @@ profile_error:
"should not be used anymore. Use "
"\'-E nodiscard\' extended option "
"instead!\n"));
- discard = 0;
+ nodiscard = 1;
break;
case 'j':
if (!journal_size)
@@ -1985,7 +1986,8 @@ profile_error:
lazy_itable_init = get_bool_from_profile(fs_types,
"lazy_itable_init",
lazy_itable_init);
- discard = get_bool_from_profile(fs_types, "discard" , discard);
+ /* nodiscard is set if -K was used, extended discard opts parsed later */
+ discard = nodiscard ? 0 : get_bool_from_profile(fs_types, "discard" , 0);
journal_flags |= get_bool_from_profile(fs_types,
"lazy_journal_init", 0) ?
EXT2_MKJOURNAL_LAZYINIT : 0;
diff --git a/misc/mke2fs.conf.in b/misc/mke2fs.conf.in
index 178733f..db6decb 100644
--- a/misc/mke2fs.conf.in
+++ b/misc/mke2fs.conf.in
@@ -2,6 +2,7 @@
base_features = sparse_super,filetype,resize_inode,dir_index,ext_attr
default_mntopts = acl,user_xattr
enable_periodic_fsck = 0
+ discard = 0
blocksize = 4096
inode_size = 256
inode_ratio = 16384
On Apr 17, 2013, at 4:19 PM, Eric Sandeen <[email protected]> wrote:
> Discard is a very big hammer. If it succeeds, there's no undo
> or going back; on the other hand, on some devices the giant discard
> issued at mkfs time makes them go offline.
>
> Although I'm reluctant to waffle back and forth on default behavior,
> I think this disabling it by default the prudent thing to do.
>
Ok, self-nak on this one, the folks in the hallway track convinced me that it was wrong. I'll send another to handle it properly in mke2fs.conf, though.
Thanks,
Eric
> This patch also adds the discard option into the default mke2fs.conf
> file; although it could be set, it was not in the template.
>
> Signed-off-by: Eric Sandeen <[email protected]>
> ---
>
> V2: handle discard disable via "-K" since it's parsed *before* we get
> the default. The extended options are parsed later, and will override
> the mke2fs.conf default without problems.
>
> diff --git a/misc/mke2fs.8.in b/misc/mke2fs.8.in
> index 023ba49..a005eca 100644
> --- a/misc/mke2fs.8.in
> +++ b/misc/mke2fs.8.in
> @@ -277,11 +277,11 @@ Attempt to discard blocks at mkfs time (discarding blocks initially is useful
> on solid state devices and sparse / thin-provisioned storage). When the device
> advertises that discard also zeroes data (any subsequent read after the discard
> and before write returns zero), then mark all not-yet-zeroed inode tables as
> -zeroed. This significantly speeds up filesystem initialization. This is set
> -as default.
> +zeroed. This significantly speeds up filesystem initialization. This is disabled
> +by default.
> .TP
> .BI nodiscard
> -Do not attempt to discard blocks at mkfs time.
> +Do not attempt to discard blocks at mkfs time. This is the default.
> .TP
> .BI quotatype
> Specify which quota type ('usr' or 'grp') is to be initialized. This
> diff --git a/misc/mke2fs.c b/misc/mke2fs.c
> index bbf477a..fa11da6 100644
> --- a/misc/mke2fs.c
> +++ b/misc/mke2fs.c
> @@ -84,7 +84,7 @@ int cflag;
> int verbose;
> int quiet;
> int super_only;
> -int discard = 1; /* attempt to discard device before fs creation */
> +int discard; /* attempt to discard device before fs creation? */
> int direct_io;
> int force;
> int noaction;
> @@ -1266,6 +1266,7 @@ static void PRS(int argc, char *argv[])
> double reserved_ratio = -1.0;
> int lsector_size = 0, psector_size = 0;
> int show_version_only = 0;
> + int nodiscard = 0;
> unsigned long long num_inodes = 0; /* unsigned long long to catch too-large input */
> errcode_t retval;
> char * oldpath = getenv("PATH");
> @@ -1444,7 +1445,7 @@ profile_error:
> "should not be used anymore. Use "
> "\'-E nodiscard\' extended option "
> "instead!\n"));
> - discard = 0;
> + nodiscard = 1;
> break;
> case 'j':
> if (!journal_size)
> @@ -1985,7 +1986,8 @@ profile_error:
> lazy_itable_init = get_bool_from_profile(fs_types,
> "lazy_itable_init",
> lazy_itable_init);
> - discard = get_bool_from_profile(fs_types, "discard" , discard);
> + /* nodiscard is set if -K was used, extended discard opts parsed later */
> + discard = nodiscard ? 0 : get_bool_from_profile(fs_types, "discard" , 0);
> journal_flags |= get_bool_from_profile(fs_types,
> "lazy_journal_init", 0) ?
> EXT2_MKJOURNAL_LAZYINIT : 0;
> diff --git a/misc/mke2fs.conf.in b/misc/mke2fs.conf.in
> index 178733f..db6decb 100644
> --- a/misc/mke2fs.conf.in
> +++ b/misc/mke2fs.conf.in
> @@ -2,6 +2,7 @@
> base_features = sparse_super,filetype,resize_inode,dir_index,ext_attr
> default_mntopts = acl,user_xattr
> enable_periodic_fsck = 0
> + discard = 0
> blocksize = 4096
> inode_size = 256
> inode_ratio = 16384
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html