Requested by yamt:

- In getpages don't allocate if we are not locked
- Use kmem_alloc instead of malloc and don't sleep

Also provide a 64 entry stack array so we don't have to allocate in the
common case.
This commit is contained in:
christos 2008-10-17 14:24:43 +00:00
parent 86bba1afff
commit 02934a6249

View File

@ -1,4 +1,4 @@
/* $NetBSD: nfs_bio.c,v 1.178 2008/10/17 06:40:21 dogcow Exp $ */
/* $NetBSD: nfs_bio.c,v 1.179 2008/10/17 14:24:43 christos Exp $ */
/*
* Copyright (c) 1989, 1993
@ -35,7 +35,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: nfs_bio.c,v 1.178 2008/10/17 06:40:21 dogcow Exp $");
__KERNEL_RCSID(0, "$NetBSD: nfs_bio.c,v 1.179 2008/10/17 14:24:43 christos Exp $");
#include "opt_nfs.h"
#include "opt_ddb.h"
@ -921,7 +921,7 @@ nfs_doio_write(struct buf *bp, struct uio *uiop)
int iomode;
bool stalewriteverf = false;
int i, npages = (bp->b_bcount + PAGE_SIZE - 1) >> PAGE_SHIFT;
struct vm_page **pgs;
struct vm_page **pgs, *spgs[64];
#ifndef NFS_V2_ONLY
bool needcommit = true; /* need only COMMIT RPC */
#else
@ -932,14 +932,20 @@ nfs_doio_write(struct buf *bp, struct uio *uiop)
int error;
off_t off, cnt;
if (npages < __arraycount(spgs))
pgs = spgs;
else {
if ((pgs = kmem_alloc(sizeof(*pgs) * npages, KM_NOSLEEP)) ==
NULL)
return ENOMEM;
}
if ((bp->b_flags & B_ASYNC) != 0 && NFS_ISV3(vp)) {
iomode = NFSV3WRITE_UNSTABLE;
} else {
iomode = NFSV3WRITE_FILESYNC;
}
pgs = malloc(sizeof(*pgs) * npages, M_TEMP, M_WAITOK);
#ifndef NFS_V2_ONLY
again:
#endif
@ -1121,7 +1127,8 @@ again:
#ifndef NFS_V2_ONLY
out:
#endif
free(pgs, M_TEMP);
if (pgs != spgs)
kmem_free(pgs, sizeof(*pgs) * npages);
return error;
}
@ -1217,19 +1224,29 @@ nfs_getpages(void *v)
struct uvm_object *uobj = &vp->v_uobj;
struct nfsnode *np = VTONFS(vp);
const int npages = *ap->a_count;
struct vm_page *pg, **pgs, **opgs;
struct vm_page *pg, **pgs, **opgs, *spgs[64];
off_t origoffset, len;
int i, error;
bool v3 = NFS_ISV3(vp);
bool write = (ap->a_access_type & VM_PROT_WRITE) != 0;
bool locked = (ap->a_flags & PGO_LOCKED) != 0;
/*
* If we are not locked we are not really using opgs,
* so just initialize it
*/
if (!locked || npages < __arraycount(spgs))
opgs = spgs;
else {
if ((opgs = kmem_alloc(npages * sizeof(*opgs), KM_NOSLEEP)) ==
NULL)
return ENOMEM;
}
/*
* call the genfs code to get the pages. `pgs' may be NULL
* when doing read-ahead.
*/
opgs = malloc(npages * sizeof(*opgs), M_TEMP, M_WAITOK);
pgs = ap->a_m;
if (write && locked && v3) {
KASSERT(pgs != NULL);
@ -1325,6 +1342,7 @@ nfs_getpages(void *v)
mutex_exit(&np->n_commitlock);
}
out:
free(opgs, M_TEMP);
if (opgs != spgs)
kmem_free(opgs, sizeof(*opgs) * npages);
return error;
}