/*
* This is now called from local media FS's to operate against their
* own vnodes if they fail to implement VOP_GETPAGES.
*/
int
vnode_pager_generic_getpages(struct vnode *vp, vm_page_t *m, int count,
int *a_rbehind, int *a_rahead, vop_getpages_iodone_t iodone, void *arg)
{
vm_object_t object;
struct bufobj *bo;
struct buf *bp;
off_t foff;
#ifdef INVARIANTS
off_t blkno0;
#endif
int bsize, pagesperblock, *freecnt;
int error, before, after, rbehind, rahead, poff, i;
int bytecount, secmask;
KASSERT(vp->v_type != VCHR && vp->v_type != VBLK,
("%s does not support devices", __func__));
if (vp->v_iflag & VI_DOOMED)
return (VM_PAGER_BAD);
object = vp->v_object;
foff = IDX_TO_OFF(m[0]->pindex);
bsize = vp->v_mount->mnt_stat.f_iosize;
pagesperblock = bsize / PAGE_SIZE;
KASSERT(foff < object->un_pager.vnp.vnp_size,
("%s: page %p offset beyond vp %p size", __func__, m[0], vp));
KASSERT(count <= sizeof(bp->b_pages),
("%s: requested %d pages", __func__, count));
/*
* The last page has valid blocks. Invalid part can only
* exist at the end of file, and the page is made fully valid
* by zeroing in vm_pager_get_pages().
*/
if (m[count - 1]->valid != 0 && --count == 0) {
if (iodone != NULL)
iodone(arg, m, 1, 0);
return (VM_PAGER_OK);
}
/*
* Synchronous and asynchronous paging operations use different
* free pbuf counters. This is done to avoid asynchronous requests
* to consume all pbufs.
* Allocate the pbuf at the very beginning of the function, so that
* if we are low on certain kind of pbufs don't even proceed to BMAP,
* but sleep.
*/
freecnt = iodone != NULL ?
&vnode_async_pbuf_freecnt : &vnode_pbuf_freecnt;
bp = getpbuf(freecnt);
/*
* Get the underlying device blocks for the file with VOP_BMAP().
* If the file system doesn't support VOP_BMAP, use old way of
* getting pages via VOP_READ.
*/
error = VOP_BMAP(vp, foff / bsize, &bo, &bp->b_blkno, &after, &before);
if (error == EOPNOTSUPP) {
relpbuf(bp, freecnt);
VM_OBJECT_WLOCK(object);
for (i = 0; i < count; i++) {
PCPU_INC(cnt.v_vnodein);
PCPU_INC(cnt.v_vnodepgsin);
error = vnode_pager_input_old(object, m[i]);
if (error)
break;
}
VM_OBJECT_WUNLOCK(object);
return (error);
} else if (error != 0) {
relpbuf(bp, freecnt);
return (VM_PAGER_ERROR);
}
/*
* If the file system supports BMAP, but blocksize is smaller
* than a page size, then use special small filesystem code.
*/
if (pagesperblock == 0) {
relpbuf(bp, freecnt);
for (i = 0; i < count; i++) {
PCPU_INC(cnt.v_vnodein);
PCPU_INC(cnt.v_vnodepgsin);
error = vnode_pager_input_smlfs(object, m[i]);
if (error)
break;
}
return (error);
}
/*
* A sparse file can be encountered only for a single page request,
* which may not be preceded by call to vm_pager_haspage().
*/
if (bp->b_blkno == -1) {
KASSERT(count == 1,
("%s: array[%d] request to a sparse file %p", __func__,
count, vp));
relpbuf(bp, freecnt);
pmap_zero_page(m[0]);
KASSERT(m[0]->dirty == 0, ("%s: page %p is dirty",
__func__, m[0]));
VM_OBJECT_WLOCK(object);
m[0]->valid = VM_PAGE_BITS_ALL;
VM_OBJECT_WUNLOCK(object);
return (VM_PAGER_OK);
}
#ifdef INVARIANTS
blkno0 = bp->b_blkno;
#endif
bp->b_blkno += (foff % bsize) / DEV_BSIZE;
/* Recalculate blocks available after/before to pages. */
poff = (foff % bsize) / PAGE_SIZE;
before *= pagesperblock;
before += poff;
after *= pagesperblock;
after += pagesperblock - (poff + 1);
if (m[0]->pindex + after >= object->size)
after = object->size - 1 - m[0]->pindex;
KASSERT(count <= after + 1, ("%s: %d pages asked, can do only %d",
__func__, count, after + 1));
after -= count - 1;
/* Trim requested rbehind/rahead to possible values. */
rbehind = a_rbehind ? *a_rbehind : 0;
rahead = a_rahead ? *a_rahead : 0;
rbehind = min(rbehind, before);
rbehind = min(rbehind, m[0]->pindex);
rahead = min(rahead, after);
rahead = min(rahead, object->size - m[count - 1]->pindex);
/*
* Check that total amount of pages fit into buf. Trim rbehind and
* rahead evenly if not.
*/
if (rbehind + rahead + count > nitems(bp->b_pages)) {
int trim, sum;
trim = rbehind + rahead + count - nitems(bp->b_pages) + 1;
sum = rbehind + rahead;
if (rbehind == before) {
/* Roundup rbehind trim to block size. */
rbehind -= roundup(trim * rbehind / sum, pagesperblock);
if (rbehind < 0)
rbehind = 0;
} else
rbehind -= trim * rbehind / sum;
rahead -= trim * rahead / sum;
}
KASSERT(rbehind + rahead + count <= nitems(bp->b_pages),
("%s: behind %d ahead %d count %d", __func__,
rbehind, rahead, count));
/*
* Fill in the bp->b_pages[] array with requested and optional
* read behind or read ahead pages. Read behind pages are looked
* up in a backward direction, down to a first cached page. Same
* for read ahead pages, but there is no need to shift the array
* in case of encountering a cached page.
*/
i = bp->b_npages = 0;
if (rbehind) {
vm_pindex_t startpindex, tpindex;
vm_page_t p;
VM_OBJECT_WLOCK(object);
startpindex = m[0]->pindex - rbehind;
if ((p = TAILQ_PREV(m[0], pglist, listq)) != NULL &&
p->pindex >= startpindex)
startpindex = p->pindex + 1;
/* tpindex is unsigned; beware of numeric underflow. */
for (tpindex = m[0]->pindex - 1;
tpindex >= startpindex && tpindex < m[0]->pindex;
tpindex--, i++) {
p = vm_page_alloc(object, tpindex, VM_ALLOC_NORMAL);
if (p == NULL) {
/* Shift the array. */
for (int j = 0; j < i; j++)
bp->b_pages[j] = bp->b_pages[j +
tpindex + 1 - startpindex];
break;
}
bp->b_pages[tpindex - startpindex] = p;
}
bp->b_pgbefore = i;
bp->b_npages += i;
bp->b_blkno -= IDX_TO_OFF(i) / DEV_BSIZE;
} else
bp->b_pgbefore = 0;
/* Requested pages. */
for (int j = 0; j < count; j++, i++)
bp->b_pages[i] = m[j];
bp->b_npages += count;
if (rahead) {
vm_pindex_t endpindex, tpindex;
vm_page_t p;
if (!VM_OBJECT_WOWNED(object))
VM_OBJECT_WLOCK(object);
endpindex = m[count - 1]->pindex + rahead + 1;
if ((p = TAILQ_NEXT(m[count - 1], listq)) != NULL &&
p->pindex < endpindex)
endpindex = p->pindex;
if (endpindex > object->size)
endpindex = object->size;
for (tpindex = m[count - 1]->pindex + 1;
tpindex < endpindex; i++, tpindex++) {
p = vm_page_alloc(object, tpindex, VM_ALLOC_NORMAL);
if (p == NULL)
break;
bp->b_pages[i] = p;
}
bp->b_pgafter = i - bp->b_npages;
bp->b_npages = i;
} else
bp->b_pgafter = 0;
if (VM_OBJECT_WOWNED(object))
VM_OBJECT_WUNLOCK(object);
/* Report back actual behind/ahead read. */
if (a_rbehind)
*a_rbehind = bp->b_pgbefore;
if (a_rahead)
*a_rahead = bp->b_pgafter;
#ifdef INVARIANTS
KASSERT(bp->b_npages <= nitems(bp->b_pages),
("%s: buf %p overflowed", __func__, bp));
for (int j = 1; j < bp->b_npages; j++)
KASSERT(bp->b_pages[j]->pindex - 1 ==
bp->b_pages[j - 1]->pindex,
("%s: pages array not consecutive, bp %p", __func__, bp));
#endif
/*
* Recalculate first offset and bytecount with regards to read behind.
* Truncate bytecount to vnode real size and round up physical size
* for real devices.
*/
foff = IDX_TO_OFF(bp->b_pages[0]->pindex);
bytecount = bp->b_npages << PAGE_SHIFT;
if ((foff + bytecount) > object->un_pager.vnp.vnp_size)
bytecount = object->un_pager.vnp.vnp_size - foff;
secmask = bo->bo_bsize - 1;
KASSERT(secmask < PAGE_SIZE && secmask > 0,
("%s: sector size %d too large", __func__, secmask + 1));
bytecount = (bytecount + secmask) & ~secmask;
/*
* And map the pages to be read into the kva, if the filesystem
* requires mapped buffers.
*/
if ((vp->v_mount->mnt_kern_flag & MNTK_UNMAPPED_BUFS) != 0 &&
unmapped_buf_allowed) {
bp->b_data = unmapped_buf;
bp->b_offset = 0;
} else {
bp->b_data = bp->b_kvabase;
pmap_qenter((vm_offset_t)bp->b_data, bp->b_pages, bp->b_npages);
}
/* Build a minimal buffer header. */
bp->b_iocmd = BIO_READ;
KASSERT(bp->b_rcred == NOCRED, ("leaking read ucred"));
KASSERT(bp->b_wcred == NOCRED, ("leaking write ucred"));
bp->b_rcred = crhold(curthread->td_ucred);
bp->b_wcred = crhold(curthread->td_ucred);
pbgetbo(bo, bp);
bp->b_vp = vp;
bp->b_bcount = bp->b_bufsize = bp->b_runningbufspace = bytecount;
bp->b_iooffset = dbtob(bp->b_blkno);
KASSERT(IDX_TO_OFF(m[0]->pindex - bp->b_pages[0]->pindex) ==
(blkno0 - bp->b_blkno) * DEV_BSIZE +
IDX_TO_OFF(m[0]->pindex) % bsize,
("wrong offsets bsize %d m[0] %ju b_pages[0] %ju "
"blkno0 %ju b_blkno %ju", bsize,
(uintmax_t)m[0]->pindex, (uintmax_t)bp->b_pages[0]->pindex,
(uintmax_t)blkno0, (uintmax_t)bp->b_blkno));
atomic_add_long(&runningbufspace, bp->b_runningbufspace);
PCPU_INC(cnt.v_vnodein);
PCPU_ADD(cnt.v_vnodepgsin, bp->b_npages);
if (iodone != NULL) { /* async */
bp->b_pgiodone = iodone;
bp->b_caller1 = arg;
bp->b_iodone = vnode_pager_generic_getpages_done_async;
bp->b_flags |= B_ASYNC;
BUF_KERNPROC(bp);
bstrategy(bp);
return (VM_PAGER_OK);
} else {
bp->b_iodone = bdone;
bstrategy(bp);
bwait(bp, PVM, "vnread");
error = vnode_pager_generic_getpages_done(bp);
for (i = 0; i < bp->b_npages; i++)
bp->b_pages[i] = NULL;
bp->b_vp = NULL;
pbrelbo(bp);
relpbuf(bp, &vnode_pbuf_freecnt);
return (error != 0 ? VM_PAGER_ERROR : VM_PAGER_OK);
}
}
/*
* This is now called from local media FS's to operate against their
* own vnodes if they fail to implement VOP_PUTPAGES.
*
* This is typically called indirectly via the pageout daemon and
* clustering has already typically occurred, so in general we ask the
* underlying filesystem to write the data out asynchronously rather
* then delayed.
*/
int
vnode_pager_generic_putpages(struct vnode *vp, vm_page_t *ma, int bytecount,
int flags, int *rtvals)
{
int i;
vm_object_t object;
vm_page_t m;
int count;
int maxsize, ncount;
vm_ooffset_t poffset;
struct uio auio;
struct iovec aiov;
int error;
int ioflags;
int ppscheck = 0;
static struct timeval lastfail;
static int curfail;
object = vp->v_object;
count = bytecount / PAGE_SIZE;
for (i = 0; i < count; i++)
rtvals[i] = VM_PAGER_ERROR;
if ((int64_t)ma[0]->pindex < 0) {
printf("vnode_pager_putpages: attempt to write meta-data!!! -- 0x%lx(%lx)\n",
(long)ma[0]->pindex, (u_long)ma[0]->dirty);
rtvals[0] = VM_PAGER_BAD;
return VM_PAGER_BAD;
}
maxsize = count * PAGE_SIZE;
ncount = count;
poffset = IDX_TO_OFF(ma[0]->pindex);
/*
* If the page-aligned write is larger then the actual file we
* have to invalidate pages occurring beyond the file EOF. However,
* there is an edge case where a file may not be page-aligned where
* the last page is partially invalid. In this case the filesystem
* may not properly clear the dirty bits for the entire page (which
* could be VM_PAGE_BITS_ALL due to the page having been mmap()d).
* With the page locked we are free to fix-up the dirty bits here.
*
* We do not under any circumstances truncate the valid bits, as
* this will screw up bogus page replacement.
*/
VM_OBJECT_WLOCK(object);
if (maxsize + poffset > object->un_pager.vnp.vnp_size) {
if (object->un_pager.vnp.vnp_size > poffset) {
int pgoff;
maxsize = object->un_pager.vnp.vnp_size - poffset;
ncount = btoc(maxsize);
if ((pgoff = (int)maxsize & PAGE_MASK) != 0) {
/*
* If the object is locked and the following
* conditions hold, then the page's dirty
* field cannot be concurrently changed by a
* pmap operation.
*/
m = ma[ncount - 1];
vm_page_assert_sbusied(m);
KASSERT(!pmap_page_is_write_mapped(m),
("vnode_pager_generic_putpages: page %p is not read-only", m));
vm_page_clear_dirty(m, pgoff, PAGE_SIZE -
pgoff);
}
} else {
maxsize = 0;
ncount = 0;
}
if (ncount < count) {
for (i = ncount; i < count; i++) {
rtvals[i] = VM_PAGER_BAD;
}
}
}
VM_OBJECT_WUNLOCK(object);
/*
* pageouts are already clustered, use IO_ASYNC to force a bawrite()
* rather then a bdwrite() to prevent paging I/O from saturating
* the buffer cache. Dummy-up the sequential heuristic to cause
* large ranges to cluster. If neither IO_SYNC or IO_ASYNC is set,
* the system decides how to cluster.
*/
ioflags = IO_VMIO;
if (flags & (VM_PAGER_PUT_SYNC | VM_PAGER_PUT_INVAL))
ioflags |= IO_SYNC;
else if ((flags & VM_PAGER_CLUSTER_OK) == 0)
ioflags |= IO_ASYNC;
ioflags |= (flags & VM_PAGER_PUT_INVAL) ? IO_INVAL: 0;
ioflags |= (flags & VM_PAGER_PUT_NOREUSE) ? IO_NOREUSE : 0;
ioflags |= IO_SEQMAX << IO_SEQSHIFT;
aiov.iov_base = (caddr_t) 0;
aiov.iov_len = maxsize;
auio.uio_iov = &aiov;
auio.uio_iovcnt = 1;
auio.uio_offset = poffset;
auio.uio_segflg = UIO_NOCOPY;
auio.uio_rw = UIO_WRITE;
auio.uio_resid = maxsize;
auio.uio_td = (struct thread *) 0;
error = VOP_WRITE(vp, &auio, ioflags, curthread->td_ucred);
PCPU_INC(cnt.v_vnodeout);
PCPU_ADD(cnt.v_vnodepgsout, ncount);
if (error) {
if ((ppscheck = ppsratecheck(&lastfail, &curfail, 1)))
printf("vnode_pager_putpages: I/O error %d\n", error);
}
if (auio.uio_resid) {
if (ppscheck || ppsratecheck(&lastfail, &curfail, 1))
printf("vnode_pager_putpages: residual I/O %zd at %lu\n",
auio.uio_resid, (u_long)ma[0]->pindex);
}
for (i = 0; i < ncount; i++) {
rtvals[i] = VM_PAGER_OK;
}
return rtvals[0];
}
static void
do_fork(struct thread *td, struct fork_req *fr, struct proc *p2, struct thread *td2,
struct vmspace *vm2, struct file *fp_procdesc)
{
struct proc *p1, *pptr;
int trypid;
struct filedesc *fd;
struct filedesc_to_leader *fdtol;
struct sigacts *newsigacts;
sx_assert(&proctree_lock, SX_SLOCKED);
sx_assert(&allproc_lock, SX_XLOCKED);
p1 = td->td_proc;
trypid = fork_findpid(fr->fr_flags);
sx_sunlock(&proctree_lock);
p2->p_state = PRS_NEW; /* protect against others */
p2->p_pid = trypid;
AUDIT_ARG_PID(p2->p_pid);
LIST_INSERT_HEAD(&allproc, p2, p_list);
allproc_gen++;
LIST_INSERT_HEAD(PIDHASH(p2->p_pid), p2, p_hash);
tidhash_add(td2);
PROC_LOCK(p2);
PROC_LOCK(p1);
sx_xunlock(&allproc_lock);
bcopy(&p1->p_startcopy, &p2->p_startcopy,
__rangeof(struct proc, p_startcopy, p_endcopy));
pargs_hold(p2->p_args);
PROC_UNLOCK(p1);
bzero(&p2->p_startzero,
__rangeof(struct proc, p_startzero, p_endzero));
/* Tell the prison that we exist. */
prison_proc_hold(p2->p_ucred->cr_prison);
PROC_UNLOCK(p2);
/*
* Malloc things while we don't hold any locks.
*/
if (fr->fr_flags & RFSIGSHARE)
newsigacts = NULL;
else
newsigacts = sigacts_alloc();
/*
* Copy filedesc.
*/
if (fr->fr_flags & RFCFDG) {
fd = fdinit(p1->p_fd, false);
fdtol = NULL;
} else if (fr->fr_flags & RFFDG) {
fd = fdcopy(p1->p_fd);
fdtol = NULL;
} else {
fd = fdshare(p1->p_fd);
if (p1->p_fdtol == NULL)
p1->p_fdtol = filedesc_to_leader_alloc(NULL, NULL,
p1->p_leader);
if ((fr->fr_flags & RFTHREAD) != 0) {
/*
* Shared file descriptor table, and shared
* process leaders.
*/
fdtol = p1->p_fdtol;
FILEDESC_XLOCK(p1->p_fd);
fdtol->fdl_refcount++;
FILEDESC_XUNLOCK(p1->p_fd);
} else {
/*
* Shared file descriptor table, and different
* process leaders.
*/
fdtol = filedesc_to_leader_alloc(p1->p_fdtol,
p1->p_fd, p2);
}
}
/*
* Make a proc table entry for the new process.
* Start by zeroing the section of proc that is zero-initialized,
* then copy the section that is copied directly from the parent.
*/
PROC_LOCK(p2);
PROC_LOCK(p1);
bzero(&td2->td_startzero,
__rangeof(struct thread, td_startzero, td_endzero));
bcopy(&td->td_startcopy, &td2->td_startcopy,
__rangeof(struct thread, td_startcopy, td_endcopy));
bcopy(&p2->p_comm, &td2->td_name, sizeof(td2->td_name));
td2->td_sigstk = td->td_sigstk;
td2->td_flags = TDF_INMEM;
td2->td_lend_user_pri = PRI_MAX;
#ifdef VIMAGE
td2->td_vnet = NULL;
td2->td_vnet_lpush = NULL;
#endif
/*
* Allow the scheduler to initialize the child.
*/
thread_lock(td);
sched_fork(td, td2);
thread_unlock(td);
/*
* Duplicate sub-structures as needed.
* Increase reference counts on shared objects.
*/
p2->p_flag = P_INMEM;
p2->p_flag2 = p1->p_flag2 & (P2_NOTRACE | P2_NOTRACE_EXEC | P2_TRAPCAP);
p2->p_swtick = ticks;
if (p1->p_flag & P_PROFIL)
startprofclock(p2);
/*
* Whilst the proc lock is held, copy the VM domain data out
* using the VM domain method.
*/
vm_domain_policy_init(&p2->p_vm_dom_policy);
vm_domain_policy_localcopy(&p2->p_vm_dom_policy,
&p1->p_vm_dom_policy);
if (fr->fr_flags & RFSIGSHARE) {
p2->p_sigacts = sigacts_hold(p1->p_sigacts);
} else {
sigacts_copy(newsigacts, p1->p_sigacts);
p2->p_sigacts = newsigacts;
}
if (fr->fr_flags & RFTSIGZMB)
p2->p_sigparent = RFTSIGNUM(fr->fr_flags);
else if (fr->fr_flags & RFLINUXTHPN)
p2->p_sigparent = SIGUSR1;
else
p2->p_sigparent = SIGCHLD;
p2->p_textvp = p1->p_textvp;
p2->p_fd = fd;
p2->p_fdtol = fdtol;
if (p1->p_flag2 & P2_INHERIT_PROTECTED) {
p2->p_flag |= P_PROTECTED;
p2->p_flag2 |= P2_INHERIT_PROTECTED;
}
/*
* p_limit is copy-on-write. Bump its refcount.
*/
lim_fork(p1, p2);
thread_cow_get_proc(td2, p2);
pstats_fork(p1->p_stats, p2->p_stats);
PROC_UNLOCK(p1);
PROC_UNLOCK(p2);
/* Bump references to the text vnode (for procfs). */
if (p2->p_textvp)
vref(p2->p_textvp);
/*
* Set up linkage for kernel based threading.
*/
if ((fr->fr_flags & RFTHREAD) != 0) {
mtx_lock(&ppeers_lock);
p2->p_peers = p1->p_peers;
p1->p_peers = p2;
p2->p_leader = p1->p_leader;
mtx_unlock(&ppeers_lock);
PROC_LOCK(p1->p_leader);
if ((p1->p_leader->p_flag & P_WEXIT) != 0) {
PROC_UNLOCK(p1->p_leader);
/*
* The task leader is exiting, so process p1 is
* going to be killed shortly. Since p1 obviously
* isn't dead yet, we know that the leader is either
* sending SIGKILL's to all the processes in this
* task or is sleeping waiting for all the peers to
* exit. We let p1 complete the fork, but we need
* to go ahead and kill the new process p2 since
* the task leader may not get a chance to send
* SIGKILL to it. We leave it on the list so that
* the task leader will wait for this new process
* to commit suicide.
*/
PROC_LOCK(p2);
kern_psignal(p2, SIGKILL);
PROC_UNLOCK(p2);
} else
PROC_UNLOCK(p1->p_leader);
} else {
p2->p_peers = NULL;
p2->p_leader = p2;
}
sx_xlock(&proctree_lock);
PGRP_LOCK(p1->p_pgrp);
PROC_LOCK(p2);
PROC_LOCK(p1);
/*
* Preserve some more flags in subprocess. P_PROFIL has already
* been preserved.
*/
p2->p_flag |= p1->p_flag & P_SUGID;
td2->td_pflags |= (td->td_pflags & TDP_ALTSTACK) | TDP_FORKING;
SESS_LOCK(p1->p_session);
if (p1->p_session->s_ttyvp != NULL && p1->p_flag & P_CONTROLT)
p2->p_flag |= P_CONTROLT;
SESS_UNLOCK(p1->p_session);
if (fr->fr_flags & RFPPWAIT)
p2->p_flag |= P_PPWAIT;
p2->p_pgrp = p1->p_pgrp;
LIST_INSERT_AFTER(p1, p2, p_pglist);
PGRP_UNLOCK(p1->p_pgrp);
LIST_INIT(&p2->p_children);
LIST_INIT(&p2->p_orphans);
callout_init_mtx(&p2->p_itcallout, &p2->p_mtx, 0);
/*
* If PF_FORK is set, the child process inherits the
* procfs ioctl flags from its parent.
*/
if (p1->p_pfsflags & PF_FORK) {
p2->p_stops = p1->p_stops;
p2->p_pfsflags = p1->p_pfsflags;
}
/*
* This begins the section where we must prevent the parent
* from being swapped.
*/
_PHOLD(p1);
PROC_UNLOCK(p1);
/*
* Attach the new process to its parent.
*
* If RFNOWAIT is set, the newly created process becomes a child
* of init. This effectively disassociates the child from the
* parent.
*/
if ((fr->fr_flags & RFNOWAIT) != 0) {
pptr = p1->p_reaper;
p2->p_reaper = pptr;
} else {
p2->p_reaper = (p1->p_treeflag & P_TREE_REAPER) != 0 ?
p1 : p1->p_reaper;
pptr = p1;
}
p2->p_pptr = pptr;
LIST_INSERT_HEAD(&pptr->p_children, p2, p_sibling);
LIST_INIT(&p2->p_reaplist);
LIST_INSERT_HEAD(&p2->p_reaper->p_reaplist, p2, p_reapsibling);
if (p2->p_reaper == p1)
p2->p_reapsubtree = p2->p_pid;
sx_xunlock(&proctree_lock);
/* Inform accounting that we have forked. */
p2->p_acflag = AFORK;
PROC_UNLOCK(p2);
#ifdef KTRACE
ktrprocfork(p1, p2);
#endif
/*
* Finish creating the child process. It will return via a different
* execution path later. (ie: directly into user mode)
*/
vm_forkproc(td, p2, td2, vm2, fr->fr_flags);
if (fr->fr_flags == (RFFDG | RFPROC)) {
PCPU_INC(cnt.v_forks);
PCPU_ADD(cnt.v_forkpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
} else if (fr->fr_flags == (RFFDG | RFPROC | RFPPWAIT | RFMEM)) {
PCPU_INC(cnt.v_vforks);
PCPU_ADD(cnt.v_vforkpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
} else if (p1 == &proc0) {
PCPU_INC(cnt.v_kthreads);
PCPU_ADD(cnt.v_kthreadpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
} else {
PCPU_INC(cnt.v_rforks);
PCPU_ADD(cnt.v_rforkpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
}
/*
* Associate the process descriptor with the process before anything
* can happen that might cause that process to need the descriptor.
* However, don't do this until after fork(2) can no longer fail.
*/
if (fr->fr_flags & RFPROCDESC)
procdesc_new(p2, fr->fr_pd_flags);
/*
* Both processes are set up, now check if any loadable modules want
* to adjust anything.
*/
EVENTHANDLER_INVOKE(process_fork, p1, p2, fr->fr_flags);
/*
* Set the child start time and mark the process as being complete.
*/
PROC_LOCK(p2);
PROC_LOCK(p1);
microuptime(&p2->p_stats->p_start);
PROC_SLOCK(p2);
p2->p_state = PRS_NORMAL;
PROC_SUNLOCK(p2);
#ifdef KDTRACE_HOOKS
/*
* Tell the DTrace fasttrap provider about the new process so that any
* tracepoints inherited from the parent can be removed. We have to do
* this only after p_state is PRS_NORMAL since the fasttrap module will
* use pfind() later on.
*/
if ((fr->fr_flags & RFMEM) == 0 && dtrace_fasttrap_fork)
dtrace_fasttrap_fork(p1, p2);
#endif
/*
* Hold the process so that it cannot exit after we make it runnable,
* but before we wait for the debugger.
*/
_PHOLD(p2);
if (p1->p_ptevents & PTRACE_FORK) {
/*
* Arrange for debugger to receive the fork event.
*
* We can report PL_FLAG_FORKED regardless of
* P_FOLLOWFORK settings, but it does not make a sense
* for runaway child.
*/
td->td_dbgflags |= TDB_FORK;
td->td_dbg_forked = p2->p_pid;
td2->td_dbgflags |= TDB_STOPATFORK;
}
if (fr->fr_flags & RFPPWAIT) {
td->td_pflags |= TDP_RFPPWAIT;
td->td_rfppwait_p = p2;
td->td_dbgflags |= TDB_VFORK;
}
PROC_UNLOCK(p2);
/*
* Now can be swapped.
*/
_PRELE(p1);
PROC_UNLOCK(p1);
/*
* Tell any interested parties about the new process.
*/
knote_fork(p1->p_klist, p2->p_pid);
SDT_PROBE3(proc, , , create, p2, p1, fr->fr_flags);
if (fr->fr_flags & RFPROCDESC) {
procdesc_finit(p2->p_procdesc, fp_procdesc);
fdrop(fp_procdesc, td);
}
if ((fr->fr_flags & RFSTOPPED) == 0) {
/*
* If RFSTOPPED not requested, make child runnable and
* add to run queue.
*/
thread_lock(td2);
TD_SET_CAN_RUN(td2);
sched_add(td2, SRQ_BORING);
thread_unlock(td2);
if (fr->fr_pidp != NULL)
*fr->fr_pidp = p2->p_pid;
} else {
*fr->fr_procp = p2;
}
PROC_LOCK(p2);
/*
* Wait until debugger is attached to child.
*/
while (td2->td_proc == p2 && (td2->td_dbgflags & TDB_STOPATFORK) != 0)
cv_wait(&p2->p_dbgwait, &p2->p_mtx);
_PRELE(p2);
racct_proc_fork_done(p2);
PROC_UNLOCK(p2);
}
/*
struct vnop_putpages_args {
struct vnode *a_vp;
vm_page_t *a_m;
int a_count;
int a_sync;
int *a_rtvals;
vm_ooffset_t a_offset;
};
*/
static int
fuse_vnop_putpages(struct vop_putpages_args *ap)
{
struct uio uio;
struct iovec iov;
vm_offset_t kva;
struct buf *bp;
int i, error, npages, count;
off_t offset;
int *rtvals;
struct vnode *vp;
struct thread *td;
struct ucred *cred;
vm_page_t *pages;
vm_ooffset_t fsize;
FS_DEBUG2G("heh\n");
vp = ap->a_vp;
KASSERT(vp->v_object, ("objectless vp passed to putpages"));
fsize = vp->v_object->un_pager.vnp.vnp_size;
td = curthread; /* XXX */
cred = curthread->td_ucred; /* XXX */
pages = ap->a_m;
count = ap->a_count;
rtvals = ap->a_rtvals;
npages = btoc(count);
offset = IDX_TO_OFF(pages[0]->pindex);
if (!fsess_opt_mmap(vnode_mount(vp))) {
FS_DEBUG("called on non-cacheable vnode??\n");
}
for (i = 0; i < npages; i++)
rtvals[i] = VM_PAGER_AGAIN;
/*
* When putting pages, do not extend file past EOF.
*/
if (offset + count > fsize) {
count = fsize - offset;
if (count < 0)
count = 0;
}
/*
* We use only the kva address for the buffer, but this is extremely
* convienient and fast.
*/
bp = getpbuf(&fuse_pbuf_freecnt);
kva = (vm_offset_t)bp->b_data;
pmap_qenter(kva, pages, npages);
PCPU_INC(cnt.v_vnodeout);
PCPU_ADD(cnt.v_vnodepgsout, count);
iov.iov_base = (caddr_t)kva;
iov.iov_len = count;
uio.uio_iov = &iov;
uio.uio_iovcnt = 1;
uio.uio_offset = offset;
uio.uio_resid = count;
uio.uio_segflg = UIO_SYSSPACE;
uio.uio_rw = UIO_WRITE;
uio.uio_td = td;
error = fuse_io_dispatch(vp, &uio, IO_DIRECT, cred);
pmap_qremove(kva, npages);
relpbuf(bp, &fuse_pbuf_freecnt);
if (!error) {
int nwritten = round_page(count - uio.uio_resid) / PAGE_SIZE;
for (i = 0; i < nwritten; i++) {
rtvals[i] = VM_PAGER_OK;
VM_OBJECT_WLOCK(pages[i]->object);
vm_page_undirty(pages[i]);
VM_OBJECT_WUNLOCK(pages[i]->object);
}
}
return rtvals[0];
}
/*
struct vnop_getpages_args {
struct vnode *a_vp;
vm_page_t *a_m;
int a_count;
int a_reqpage;
vm_ooffset_t a_offset;
};
*/
static int
fuse_vnop_getpages(struct vop_getpages_args *ap)
{
int i, error, nextoff, size, toff, count, npages;
struct uio uio;
struct iovec iov;
vm_offset_t kva;
struct buf *bp;
struct vnode *vp;
struct thread *td;
struct ucred *cred;
vm_page_t *pages;
FS_DEBUG2G("heh\n");
vp = ap->a_vp;
KASSERT(vp->v_object, ("objectless vp passed to getpages"));
td = curthread; /* XXX */
cred = curthread->td_ucred; /* XXX */
pages = ap->a_m;
count = ap->a_count;
if (!fsess_opt_mmap(vnode_mount(vp))) {
FS_DEBUG("called on non-cacheable vnode??\n");
return (VM_PAGER_ERROR);
}
npages = btoc(count);
/*
* If the requested page is partially valid, just return it and
* allow the pager to zero-out the blanks. Partially valid pages
* can only occur at the file EOF.
*/
VM_OBJECT_WLOCK(vp->v_object);
fuse_vm_page_lock_queues();
if (pages[ap->a_reqpage]->valid != 0) {
for (i = 0; i < npages; ++i) {
if (i != ap->a_reqpage) {
fuse_vm_page_lock(pages[i]);
vm_page_free(pages[i]);
fuse_vm_page_unlock(pages[i]);
}
}
fuse_vm_page_unlock_queues();
VM_OBJECT_WUNLOCK(vp->v_object);
return 0;
}
fuse_vm_page_unlock_queues();
VM_OBJECT_WUNLOCK(vp->v_object);
/*
* We use only the kva address for the buffer, but this is extremely
* convienient and fast.
*/
bp = getpbuf(&fuse_pbuf_freecnt);
kva = (vm_offset_t)bp->b_data;
pmap_qenter(kva, pages, npages);
PCPU_INC(cnt.v_vnodein);
PCPU_ADD(cnt.v_vnodepgsin, npages);
iov.iov_base = (caddr_t)kva;
iov.iov_len = count;
uio.uio_iov = &iov;
uio.uio_iovcnt = 1;
uio.uio_offset = IDX_TO_OFF(pages[0]->pindex);
uio.uio_resid = count;
uio.uio_segflg = UIO_SYSSPACE;
uio.uio_rw = UIO_READ;
uio.uio_td = td;
error = fuse_io_dispatch(vp, &uio, IO_DIRECT, cred);
pmap_qremove(kva, npages);
relpbuf(bp, &fuse_pbuf_freecnt);
if (error && (uio.uio_resid == count)) {
FS_DEBUG("error %d\n", error);
VM_OBJECT_WLOCK(vp->v_object);
fuse_vm_page_lock_queues();
for (i = 0; i < npages; ++i) {
if (i != ap->a_reqpage) {
fuse_vm_page_lock(pages[i]);
vm_page_free(pages[i]);
fuse_vm_page_unlock(pages[i]);
}
}
fuse_vm_page_unlock_queues();
VM_OBJECT_WUNLOCK(vp->v_object);
return VM_PAGER_ERROR;
}
/*
* Calculate the number of bytes read and validate only that number
* of bytes. Note that due to pending writes, size may be 0. This
* does not mean that the remaining data is invalid!
*/
size = count - uio.uio_resid;
VM_OBJECT_WLOCK(vp->v_object);
fuse_vm_page_lock_queues();
for (i = 0, toff = 0; i < npages; i++, toff = nextoff) {
vm_page_t m;
nextoff = toff + PAGE_SIZE;
m = pages[i];
if (nextoff <= size) {
/*
* Read operation filled an entire page
*/
m->valid = VM_PAGE_BITS_ALL;
KASSERT(m->dirty == 0,
("fuse_getpages: page %p is dirty", m));
} else if (size > toff) {
/*
* Read operation filled a partial page.
*/
m->valid = 0;
vm_page_set_valid_range(m, 0, size - toff);
KASSERT(m->dirty == 0,
("fuse_getpages: page %p is dirty", m));
} else {
/*
* Read operation was short. If no error occured
* we may have hit a zero-fill section. We simply
* leave valid set to 0.
*/
;
}
if (i != ap->a_reqpage)
vm_page_readahead_finish(m);
}
fuse_vm_page_unlock_queues();
VM_OBJECT_WUNLOCK(vp->v_object);
return 0;
}
/*
* This is now called from local media FS's to operate against their
* own vnodes if they fail to implement VOP_GETPAGES.
*/
int
vnode_pager_generic_getpages(struct vnode *vp, vm_page_t *m, int bytecount,
int reqpage, vop_getpages_iodone_t iodone, void *arg)
{
vm_object_t object;
struct bufobj *bo;
struct buf *bp;
daddr_t firstaddr, reqblock;
off_t foff, pib;
int pbefore, pafter, i, size, bsize, first, last, *freecnt;
int count, error, before, after, secmask;
KASSERT(vp->v_type != VCHR && vp->v_type != VBLK,
("vnode_pager_generic_getpages does not support devices"));
if (vp->v_iflag & VI_DOOMED)
return (VM_PAGER_BAD);
object = vp->v_object;
count = bytecount / PAGE_SIZE;
bsize = vp->v_mount->mnt_stat.f_iosize;
/*
* Synchronous and asynchronous paging operations use different
* free pbuf counters. This is done to avoid asynchronous requests
* to consume all pbufs.
* Allocate the pbuf at the very beginning of the function, so that
* if we are low on certain kind of pbufs don't even proceed to BMAP,
* but sleep.
*/
freecnt = iodone != NULL ?
&vnode_async_pbuf_freecnt : &vnode_pbuf_freecnt;
bp = getpbuf(freecnt);
/*
* Get the underlying device blocks for the file with VOP_BMAP().
* If the file system doesn't support VOP_BMAP, use old way of
* getting pages via VOP_READ.
*/
error = VOP_BMAP(vp, IDX_TO_OFF(m[reqpage]->pindex) / bsize, &bo,
&reqblock, &after, &before);
if (error == EOPNOTSUPP) {
relpbuf(bp, freecnt);
VM_OBJECT_WLOCK(object);
for (i = 0; i < count; i++)
if (i != reqpage) {
vm_page_lock(m[i]);
vm_page_free(m[i]);
vm_page_unlock(m[i]);
}
PCPU_INC(cnt.v_vnodein);
PCPU_INC(cnt.v_vnodepgsin);
error = vnode_pager_input_old(object, m[reqpage]);
VM_OBJECT_WUNLOCK(object);
return (error);
} else if (error != 0) {
relpbuf(bp, freecnt);
vm_pager_free_nonreq(object, m, reqpage, count, FALSE);
return (VM_PAGER_ERROR);
/*
* If the blocksize is smaller than a page size, then use
* special small filesystem code.
*/
} else if ((PAGE_SIZE / bsize) > 1) {
relpbuf(bp, freecnt);
vm_pager_free_nonreq(object, m, reqpage, count, FALSE);
PCPU_INC(cnt.v_vnodein);
PCPU_INC(cnt.v_vnodepgsin);
return (vnode_pager_input_smlfs(object, m[reqpage]));
}
/*
* Since the caller has busied the requested page, that page's valid
* field will not be changed by other threads.
*/
vm_page_assert_xbusied(m[reqpage]);
/*
* If we have a completely valid page available to us, we can
* clean up and return. Otherwise we have to re-read the
* media.
*/
if (m[reqpage]->valid == VM_PAGE_BITS_ALL) {
relpbuf(bp, freecnt);
vm_pager_free_nonreq(object, m, reqpage, count, FALSE);
return (VM_PAGER_OK);
} else if (reqblock == -1) {
relpbuf(bp, freecnt);
pmap_zero_page(m[reqpage]);
KASSERT(m[reqpage]->dirty == 0,
("vnode_pager_generic_getpages: page %p is dirty", m));
VM_OBJECT_WLOCK(object);
m[reqpage]->valid = VM_PAGE_BITS_ALL;
vm_pager_free_nonreq(object, m, reqpage, count, TRUE);
VM_OBJECT_WUNLOCK(object);
return (VM_PAGER_OK);
} else if (m[reqpage]->valid != 0) {
VM_OBJECT_WLOCK(object);
m[reqpage]->valid = 0;
VM_OBJECT_WUNLOCK(object);
}
pib = IDX_TO_OFF(m[reqpage]->pindex) % bsize;
pbefore = ((daddr_t)before * bsize + pib) / PAGE_SIZE;
pafter = ((daddr_t)(after + 1) * bsize - pib) / PAGE_SIZE - 1;
first = reqpage < pbefore ? 0 : reqpage - pbefore;
last = reqpage + pafter >= count ? count - 1 : reqpage + pafter;
if (first > 0 || last + 1 < count) {
VM_OBJECT_WLOCK(object);
for (i = 0; i < first; i++) {
vm_page_lock(m[i]);
vm_page_free(m[i]);
vm_page_unlock(m[i]);
}
for (i = last + 1; i < count; i++) {
vm_page_lock(m[i]);
vm_page_free(m[i]);
vm_page_unlock(m[i]);
}
VM_OBJECT_WUNLOCK(object);
}
/*
* here on direct device I/O
*/
firstaddr = reqblock;
firstaddr += pib / DEV_BSIZE;
firstaddr -= IDX_TO_OFF(reqpage - first) / DEV_BSIZE;
/*
* The first and last page have been calculated now, move
* input pages to be zero based, and adjust the count.
*/
m += first;
reqpage -= first;
count = last - first + 1;
/*
* calculate the file virtual address for the transfer
*/
foff = IDX_TO_OFF(m[0]->pindex);
/*
* calculate the size of the transfer
*/
size = count * PAGE_SIZE;
KASSERT(count > 0, ("zero count"));
if ((foff + size) > object->un_pager.vnp.vnp_size)
size = object->un_pager.vnp.vnp_size - foff;
KASSERT(size > 0, ("zero size"));
/*
* round up physical size for real devices.
*/
secmask = bo->bo_bsize - 1;
KASSERT(secmask < PAGE_SIZE && secmask > 0,
("vnode_pager_generic_getpages: sector size %d too large",
secmask + 1));
size = (size + secmask) & ~secmask;
/*
* and map the pages to be read into the kva, if the filesystem
* requires mapped buffers.
*/
if ((vp->v_mount->mnt_kern_flag & MNTK_UNMAPPED_BUFS) != 0 &&
unmapped_buf_allowed) {
bp->b_data = unmapped_buf;
bp->b_offset = 0;
} else {
bp->b_data = bp->b_kvabase;
pmap_qenter((vm_offset_t)bp->b_data, m, count);
}
/* build a minimal buffer header */
bp->b_iocmd = BIO_READ;
KASSERT(bp->b_rcred == NOCRED, ("leaking read ucred"));
KASSERT(bp->b_wcred == NOCRED, ("leaking write ucred"));
bp->b_rcred = crhold(curthread->td_ucred);
bp->b_wcred = crhold(curthread->td_ucred);
bp->b_blkno = firstaddr;
pbgetbo(bo, bp);
bp->b_vp = vp;
bp->b_bcount = size;
bp->b_bufsize = size;
bp->b_runningbufspace = bp->b_bufsize;
for (i = 0; i < count; i++)
bp->b_pages[i] = m[i];
bp->b_npages = count;
bp->b_pager.pg_reqpage = reqpage;
atomic_add_long(&runningbufspace, bp->b_runningbufspace);
PCPU_INC(cnt.v_vnodein);
PCPU_ADD(cnt.v_vnodepgsin, count);
/* do the input */
bp->b_iooffset = dbtob(bp->b_blkno);
if (iodone != NULL) { /* async */
bp->b_pager.pg_iodone = iodone;
bp->b_caller1 = arg;
bp->b_iodone = vnode_pager_generic_getpages_done_async;
bp->b_flags |= B_ASYNC;
BUF_KERNPROC(bp);
bstrategy(bp);
/* Good bye! */
} else {
bp->b_iodone = bdone;
bstrategy(bp);
bwait(bp, PVM, "vnread");
error = vnode_pager_generic_getpages_done(bp);
for (i = 0; i < bp->b_npages; i++)
bp->b_pages[i] = NULL;
bp->b_vp = NULL;
pbrelbo(bp);
relpbuf(bp, &vnode_pbuf_freecnt);
}
return (error != 0 ? VM_PAGER_ERROR : VM_PAGER_OK);
}
