a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
/* $NetBSD: ffs_extern.h,v 1.21 2001/09/15 20:36:42 chs Exp $ */
|
1994-06-29 10:39:25 +04:00
|
|
|
|
1994-06-08 15:41:58 +04:00
|
|
|
/*-
|
|
|
|
|
* Copyright (c) 1991, 1993, 1994
|
|
|
|
|
* The Regents of the University of California. All rights reserved.
|
|
|
|
|
*
|
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
|
* are met:
|
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
|
* 3. All advertising materials mentioning features or use of this software
|
|
|
|
|
* must display the following acknowledgement:
|
|
|
|
|
* This product includes software developed by the University of
|
|
|
|
|
* California, Berkeley and its contributors.
|
|
|
|
|
* 4. Neither the name of the University nor the names of its contributors
|
|
|
|
|
* may be used to endorse or promote products derived from this software
|
|
|
|
|
* without specific prior written permission.
|
|
|
|
|
*
|
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
|
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
|
*
|
1998-03-01 05:20:01 +03:00
|
|
|
* @(#)ffs_extern.h 8.6 (Berkeley) 3/30/95
|
1994-06-08 15:41:58 +04:00
|
|
|
*/
|
|
|
|
|
|
1998-03-01 05:20:01 +03:00
|
|
|
/*
|
|
|
|
|
* Sysctl values for the fast filesystem.
|
|
|
|
|
*/
|
|
|
|
|
#define FFS_CLUSTERREAD 1 /* cluster reading enabled */
|
|
|
|
|
#define FFS_CLUSTERWRITE 2 /* cluster writing enabled */
|
|
|
|
|
#define FFS_REALLOCBLKS 3 /* block reallocation enabled */
|
|
|
|
|
#define FFS_ASYNCFREE 4 /* asynchronous block freeing enabled */
|
2000-04-04 13:23:20 +04:00
|
|
|
#define FFS_LOG_CHANGEOPT 5 /* log optimalization strategy change */
|
|
|
|
|
#define FFS_MAXID 6 /* number of valid ffs ids */
|
1998-03-01 05:20:01 +03:00
|
|
|
|
|
|
|
|
#define FFS_NAMES { \
|
|
|
|
|
{ 0, 0 }, \
|
|
|
|
|
{ "doclusterread", CTLTYPE_INT }, \
|
|
|
|
|
{ "doclusterwrite", CTLTYPE_INT }, \
|
|
|
|
|
{ "doreallocblks", CTLTYPE_INT }, \
|
|
|
|
|
{ "doasyncfree", CTLTYPE_INT }, \
|
2000-04-04 13:23:20 +04:00
|
|
|
{ "log_changeopt", CTLTYPE_INT }, \
|
1998-03-01 05:20:01 +03:00
|
|
|
}
|
|
|
|
|
|
1994-06-08 15:41:58 +04:00
|
|
|
struct buf;
|
|
|
|
|
struct fid;
|
|
|
|
|
struct fs;
|
|
|
|
|
struct inode;
|
1998-03-18 18:57:26 +03:00
|
|
|
struct dinode;
|
1994-06-08 15:41:58 +04:00
|
|
|
struct mount;
|
|
|
|
|
struct nameidata;
|
|
|
|
|
struct proc;
|
|
|
|
|
struct statfs;
|
|
|
|
|
struct timeval;
|
|
|
|
|
struct ucred;
|
1996-02-10 01:22:18 +03:00
|
|
|
struct ufsmount;
|
1994-06-08 15:41:58 +04:00
|
|
|
struct uio;
|
|
|
|
|
struct vnode;
|
|
|
|
|
struct mbuf;
|
1996-02-10 01:22:18 +03:00
|
|
|
struct cg;
|
1994-06-08 15:41:58 +04:00
|
|
|
|
1998-09-01 07:11:08 +04:00
|
|
|
extern struct pool ffs_inode_pool; /* memory pool for inodes */
|
|
|
|
|
|
1994-06-08 15:41:58 +04:00
|
|
|
__BEGIN_DECLS
|
|
|
|
|
|
1996-02-10 01:22:18 +03:00
|
|
|
/* ffs_alloc.c */
|
1998-03-01 05:20:01 +03:00
|
|
|
int ffs_alloc __P((struct inode *, ufs_daddr_t, ufs_daddr_t , int, struct ucred *,
|
|
|
|
|
ufs_daddr_t *));
|
|
|
|
|
int ffs_realloccg __P((struct inode *, ufs_daddr_t, ufs_daddr_t, int, int ,
|
2000-11-27 11:39:39 +03:00
|
|
|
struct ucred *, struct buf **, ufs_daddr_t *));
|
1996-02-10 01:22:18 +03:00
|
|
|
int ffs_reallocblks __P((void *));
|
|
|
|
|
int ffs_valloc __P((void *));
|
1998-03-01 05:20:01 +03:00
|
|
|
ufs_daddr_t ffs_blkpref __P((struct inode *, ufs_daddr_t, int, ufs_daddr_t *));
|
|
|
|
|
void ffs_blkfree __P((struct inode *, ufs_daddr_t, long));
|
1996-02-10 01:22:18 +03:00
|
|
|
int ffs_vfree __P((void *));
|
1999-11-15 21:49:07 +03:00
|
|
|
void ffs_clusteracct __P((struct fs *, struct cg *, ufs_daddr_t, int));
|
1994-06-08 15:41:58 +04:00
|
|
|
|
1996-02-10 01:22:18 +03:00
|
|
|
/* ffs_balloc.c */
|
1999-11-15 21:49:07 +03:00
|
|
|
int ffs_balloc __P((void *));
|
a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
int ffs_gop_alloc __P((struct vnode *, off_t, off_t, int, struct ucred *));
|
1996-02-10 01:22:18 +03:00
|
|
|
|
1998-03-18 18:57:26 +03:00
|
|
|
/* ffs_bswap.c */
|
2001-08-17 06:18:46 +04:00
|
|
|
void ffs_sb_swap __P((struct fs*, struct fs *));
|
1998-03-18 18:57:26 +03:00
|
|
|
void ffs_dinode_swap __P((struct dinode *, struct dinode *));
|
|
|
|
|
void ffs_csum_swap __P((struct csum *, struct csum *, int));
|
|
|
|
|
|
1996-02-10 01:22:18 +03:00
|
|
|
/* ffs_inode.c */
|
|
|
|
|
int ffs_update __P((void *));
|
|
|
|
|
int ffs_truncate __P((void *));
|
|
|
|
|
|
|
|
|
|
/* ffs_subr.c */
|
|
|
|
|
int ffs_blkatoff __P((void *));
|
1999-11-15 21:49:07 +03:00
|
|
|
int ffs_freefile __P((void *));
|
1998-03-18 18:57:26 +03:00
|
|
|
void ffs_fragacct __P((struct fs *, int, int32_t[], int, int));
|
1994-06-08 15:41:58 +04:00
|
|
|
#ifdef DIAGNOSTIC
|
|
|
|
|
void ffs_checkoverlap __P((struct buf *, struct inode *));
|
|
|
|
|
#endif
|
2001-08-09 12:15:26 +04:00
|
|
|
int ffs_isblock __P((struct fs *, u_char *, ufs_daddr_t));
|
|
|
|
|
int ffs_isfreeblock __P((struct fs *, u_char *, ufs_daddr_t));
|
1998-03-01 05:20:01 +03:00
|
|
|
void ffs_clrblock __P((struct fs *, u_char *, ufs_daddr_t));
|
2001-08-09 12:15:26 +04:00
|
|
|
void ffs_setblock __P((struct fs *, u_char *, ufs_daddr_t));
|
1996-02-10 01:22:18 +03:00
|
|
|
|
|
|
|
|
/* ffs_vfsops.c */
|
2000-03-16 21:20:06 +03:00
|
|
|
void ffs_init __P((void));
|
2001-09-15 20:12:54 +04:00
|
|
|
void ffs_reinit __P((void));
|
2000-03-16 21:20:06 +03:00
|
|
|
void ffs_done __P((void));
|
1996-02-10 01:22:18 +03:00
|
|
|
int ffs_mountroot __P((void));
|
1996-12-22 13:10:12 +03:00
|
|
|
int ffs_mount __P((struct mount *, const char *, void *, struct nameidata *,
|
1996-02-10 01:22:18 +03:00
|
|
|
struct proc *));
|
|
|
|
|
int ffs_reload __P((struct mount *, struct ucred *, struct proc *));
|
|
|
|
|
int ffs_mountfs __P((struct vnode *, struct mount *, struct proc *));
|
|
|
|
|
int ffs_oldfscompat __P((struct fs *));
|
|
|
|
|
int ffs_unmount __P((struct mount *, int, struct proc *));
|
|
|
|
|
int ffs_flushfiles __P((struct mount *, int, struct proc *));
|
|
|
|
|
int ffs_statfs __P((struct mount *, struct statfs *, struct proc *));
|
|
|
|
|
int ffs_sync __P((struct mount *, int, struct ucred *, struct proc *));
|
|
|
|
|
int ffs_vget __P((struct mount *, ino_t, struct vnode **));
|
1999-02-27 02:44:43 +03:00
|
|
|
int ffs_fhtovp __P((struct mount *, struct fid *, struct vnode **));
|
1996-02-10 01:22:18 +03:00
|
|
|
int ffs_vptofh __P((struct vnode *, struct fid *));
|
1998-03-01 05:20:01 +03:00
|
|
|
int ffs_sysctl __P((int *, u_int, void *, size_t *, void *, size_t,
|
|
|
|
|
struct proc *));
|
1996-02-10 01:22:18 +03:00
|
|
|
int ffs_sbupdate __P((struct ufsmount *, int));
|
|
|
|
|
int ffs_cgupdate __P((struct ufsmount *, int));
|
|
|
|
|
|
|
|
|
|
/* ffs_vnops.c */
|
|
|
|
|
int ffs_read __P((void *));
|
|
|
|
|
int ffs_write __P((void *));
|
1999-11-15 21:49:07 +03:00
|
|
|
int ffs_fsync __P((void *));
|
1996-02-10 01:22:18 +03:00
|
|
|
int ffs_reclaim __P((void *));
|
a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
int ffs_getpages __P((void *));
|
|
|
|
|
void ffs_gop_size __P((struct vnode *, off_t, off_t *));
|
1994-06-08 15:41:58 +04:00
|
|
|
__END_DECLS
|
|
|
|
|
|
1999-11-15 21:49:07 +03:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Soft dependency function prototypes.
|
|
|
|
|
*/
|
|
|
|
|
void softdep_initialize __P((void));
|
2001-09-15 20:12:54 +04:00
|
|
|
void softdep_reinitialize __P((void));
|
1999-11-15 21:49:07 +03:00
|
|
|
int softdep_mount __P((struct vnode *, struct mount *, struct fs *,
|
|
|
|
|
struct ucred *));
|
|
|
|
|
int softdep_flushfiles __P((struct mount *, int, struct proc *));
|
|
|
|
|
void softdep_update_inodeblock __P((struct inode *, struct buf *, int));
|
|
|
|
|
void softdep_load_inodeblock __P((struct inode *));
|
|
|
|
|
void softdep_freefile __P((void *));
|
|
|
|
|
void softdep_setup_freeblocks __P((struct inode *, off_t));
|
|
|
|
|
void softdep_setup_inomapdep __P((struct buf *, struct inode *, ino_t));
|
|
|
|
|
void softdep_setup_blkmapdep __P((struct buf *, struct fs *, ufs_daddr_t));
|
|
|
|
|
void softdep_setup_allocdirect __P((struct inode *, ufs_lbn_t, ufs_daddr_t,
|
|
|
|
|
ufs_daddr_t, long, long, struct buf *));
|
|
|
|
|
void softdep_setup_allocindir_meta __P((struct buf *, struct inode *,
|
|
|
|
|
struct buf *, int, ufs_daddr_t));
|
|
|
|
|
void softdep_setup_allocindir_page __P((struct inode *, ufs_lbn_t,
|
|
|
|
|
struct buf *, int, ufs_daddr_t, ufs_daddr_t, struct buf *));
|
|
|
|
|
void softdep_fsync_mountdev __P((struct vnode *));
|
|
|
|
|
int softdep_sync_metadata __P((void *));
|
|
|
|
|
|
1996-02-10 01:22:18 +03:00
|
|
|
extern int (**ffs_vnodeop_p) __P((void *));
|
|
|
|
|
extern int (**ffs_specop_p) __P((void *));
|
|
|
|
|
extern int (**ffs_fifoop_p) __P((void *));
|
1998-06-25 00:58:44 +04:00
|
|
|
|
|
|
|
|
|