Hand merged.

arch/i386/kernel/cpu/mtrr/generic.c

@@ -69,7 +69,9 @@ void generic_set_mtrr(unsigned int reg, unsigned long base,
 {
 	u32 cr0, cr4 = 0;
 	u32 deftype_lo, deftype_hi;
+	static spinlock_t set_atomicity_lock = SPIN_LOCK_UNLOCKED;
 
+	spin_lock(&set_atomicity_lock);
 	/*  Save value of CR4 and clear Page Global Enable (bit 7)  */
 	if ( cpu_has_pge ) {
 		cr4 = read_cr4();
@@ -112,6 +114,7 @@ void generic_set_mtrr(unsigned int reg, unsigned long base,
 	/*  Restore value of CR4  */
 	if ( cpu_has_pge )
 		write_cr4(cr4);
+	spin_unlock(&set_atomicity_lock);
 }
 
 int generic_validate_add_page(unsigned long base, unsigned long size, unsigned int type)

arch/i386/mm/fault.c

@@ -184,7 +184,7 @@ asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long error_code)
 	 * If we're in an interrupt, have no user context or are running in an
 	 * atomic region then we must not take the fault..
 	 */
-	if (preempt_count() || !mm)
+	if (in_atomic() || !mm)
 		goto no_context;
 
 	down_read(&mm->mmap_sem);

drivers/md/linear.c

@@ -55,12 +55,12 @@ static int linear_run (mddev_t *mddev)
 		int j = rdev->raid_disk;
 		dev_info_t *disk = conf->disks + j;
 
-		if (j < 0 || j > mddev->raid_disks || disk->bdev) {
+		if (j < 0 || j > mddev->raid_disks || disk->rdev) {
 			printk("linear: disk numbering problem. Aborting!\n");
 			goto out;
 		}
 
-		disk->bdev = rdev->bdev;
+		disk->rdev = rdev;
 		disk->size = rdev->size;
 
 		if (!conf->smallest || (disk->size < conf->smallest->size))
@@ -153,11 +153,11 @@ static int linear_make_request (request_queue_t *q, struct bio *bio)
     
 	if (block >= (tmp_dev->size + tmp_dev->offset)
 				|| block < tmp_dev->offset) {
-		printk ("linear_make_request: Block %ld out of bounds on dev %s size %ld offset %ld\n", block, bdevname(tmp_dev->bdev), tmp_dev->size, tmp_dev->offset);
+		printk ("linear_make_request: Block %ld out of bounds on dev %s size %ld offset %ld\n", block, bdevname(tmp_dev->rdev->bdev), tmp_dev->size, tmp_dev->offset);
 		bio_io_error(bio);
 		return 0;
 	}
-	bio->bi_bdev = tmp_dev->bdev;
+	bio->bi_bdev = tmp_dev->rdev->bdev;
 	bio->bi_sector = bio->bi_sector - (tmp_dev->offset << 1);
 
 	return 1;
@@ -176,11 +176,11 @@ static int linear_status (char *page, mddev_t *mddev)
 	for (j = 0; j < conf->nr_zones; j++)
 	{
 		sz += sprintf(page+sz, "[%s",
-			bdev_partition_name(conf->hash_table[j].dev0->bdev));
+			bdev_partition_name(conf->hash_table[j].dev0->rdev->bdev));
 
 		if (conf->hash_table[j].dev1)
 			sz += sprintf(page+sz, "/%s] ",
-			  bdev_partition_name(conf->hash_table[j].dev1->bdev));
+			  bdev_partition_name(conf->hash_table[j].dev1->rdev->bdev));
 		else
 			sz += sprintf(page+sz, "] ");
 	}

drivers/md/multipath.c

@@ -70,7 +70,7 @@ static void mp_pool_free(void *mpb, void *data)
 	kfree(mpb);
 }
 
-static int multipath_map (mddev_t *mddev, struct block_device **bdev)
+static int multipath_map (mddev_t *mddev, mdk_rdev_t **rdevp)
 {
 	multipath_conf_t *conf = mddev_to_conf(mddev);
 	int i, disks = MD_SB_DISKS;
@@ -80,12 +80,17 @@ static int multipath_map (mddev_t *mddev, struct block_device **bdev)
 	 * now we use the first available disk.
 	 */
 
+	spin_lock_irq(&conf->device_lock);
 	for (i = 0; i < disks; i++) {
-		if (conf->multipaths[i].operational) {
-			*bdev = conf->multipaths[i].bdev;
-			return (0);
+		mdk_rdev_t *rdev = conf->multipaths[i].rdev;
+		if (rdev && rdev->in_sync) {
+			*rdevp = rdev;
+			atomic_inc(&rdev->nr_pending);
+			spin_unlock_irq(&conf->device_lock);
+			return 0;
 		}
 	}
+	spin_unlock_irq(&conf->device_lock);
 
 	printk (KERN_ERR "multipath_map(): no more operational IO paths?\n");
 	return (-1);
@@ -126,21 +131,21 @@ void multipath_end_request(struct bio *bio)
 {
 	int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
 	struct multipath_bh * mp_bh = (struct multipath_bh *)(bio->bi_private);
-	multipath_conf_t *conf;
-	struct block_device *bdev;
-	if (uptodate) {
+	multipath_conf_t *conf = mddev_to_conf(mp_bh->mddev);
+	mdk_rdev_t *rdev = conf->multipaths[mp_bh->path].rdev;
+
+	if (uptodate)
 		multipath_end_bh_io(mp_bh, uptodate);
-		return;
-	}
+	else {
 		/*
 		 * oops, IO error:
 		 */
-	conf = mddev_to_conf(mp_bh->mddev);
-	bdev = conf->multipaths[mp_bh->path].bdev;
-	md_error (mp_bh->mddev, bdev);
+		md_error (mp_bh->mddev, rdev);
 		printk(KERN_ERR "multipath: %s: rescheduling sector %lu\n", 
-		 bdev_partition_name(bdev), bio->bi_sector);
+		       bdev_partition_name(rdev->bdev), bio->bi_sector);
 		multipath_reschedule_retry(mp_bh);
+	}
+	atomic_dec(&rdev->nr_pending);
 	return;
 }
 
@@ -153,9 +158,11 @@ static int multipath_read_balance (multipath_conf_t *conf)
 {
 	int disk;
 
-	for (disk = 0; disk < MD_SB_DISKS; disk++)	
-		if (conf->multipaths[disk].operational)
+	for (disk = 0; disk < MD_SB_DISKS; disk++) {
+		mdk_rdev_t *rdev = conf->multipaths[disk].rdev;
+		if (rdev && rdev->in_sync)
 			return disk;
+	}
 	BUG();
 	return 0;
 }
@@ -175,11 +182,14 @@ static int multipath_make_request (request_queue_t *q, struct bio * bio)
 	/*
 	 * read balancing logic:
 	 */
+	spin_lock_irq(&conf->device_lock);
 	mp_bh->path = multipath_read_balance(conf);
 	multipath = conf->multipaths + mp_bh->path;
+	atomic_inc(&multipath->rdev->nr_pending);
+	spin_unlock_irq(&conf->device_lock);
 
 	mp_bh->bio = *bio;
-	mp_bh->bio.bi_bdev = multipath->bdev;
+	mp_bh->bio.bi_bdev = multipath->rdev->bdev;
 	mp_bh->bio.bi_end_io = multipath_end_request;
 	mp_bh->bio.bi_private = mp_bh;
 	generic_make_request(&mp_bh->bio);
@@ -195,7 +205,8 @@ static int multipath_status (char *page, mddev_t *mddev)
 						 conf->working_disks);
 	for (i = 0; i < conf->raid_disks; i++)
 		sz += sprintf (page+sz, "%s",
-			conf->multipaths[i].operational ? "U" : "_");
+			       conf->multipaths[i].rdev && 
+			       conf->multipaths[i].rdev->in_sync ? "U" : "_");
 	sz += sprintf (page+sz, "]");
 	return sz;
 }
@@ -210,28 +221,13 @@ static int multipath_status (char *page, mddev_t *mddev)
 "multipath: IO failure on %s, disabling IO path. \n" \
 "	Operation continuing on %d IO paths.\n"
 
-static void mark_disk_bad (mddev_t *mddev, int failed)
-{
-	multipath_conf_t *conf = mddev_to_conf(mddev);
-	struct multipath_info *multipath = conf->multipaths+failed;
-
-	multipath->operational = 0;
-	mddev->sb_dirty = 1;
-	conf->working_disks--;
-	printk (DISK_FAILED, bdev_partition_name (multipath->bdev),
-				 conf->working_disks);
-}
 
 /*
  * Careful, this can execute in IRQ contexts as well!
  */
-static int multipath_error (mddev_t *mddev, struct block_device *bdev)
+static void multipath_error (mddev_t *mddev, mdk_rdev_t *rdev)
 {
 	multipath_conf_t *conf = mddev_to_conf(mddev);
-	struct multipath_info * multipaths = conf->multipaths;
-	int disks = MD_SB_DISKS;
-	int i;
-
 
 	if (conf->working_disks <= 1) {
 		/*
@@ -239,24 +235,21 @@ static int multipath_error (mddev_t *mddev, struct block_device *bdev)
 		 * first check if this is a queued request for a device
 		 * which has just failed.
 		 */
-		for (i = 0; i < disks; i++) {
-			if (multipaths[i].bdev == bdev && !multipaths[i].operational)
-				return 0;
-		}
 		printk (LAST_DISK);
-		return 1; /* leave it active... it's all we have */
+		/* leave it active... it's all we have */
 	} else {
 		/*
 		 * Mark disk as unusable
 		 */
-		for (i = 0; i < disks; i++) {
-			if (multipaths[i].bdev == bdev && multipaths[i].operational) {
-				mark_disk_bad(mddev, i);
-				break;
-			}
+		if (!rdev->faulty) {
+			rdev->in_sync = 0;
+			rdev->faulty = 1;
+			mddev->sb_dirty = 1;
+			conf->working_disks--;
+			printk (DISK_FAILED, bdev_partition_name (rdev->bdev),
+				conf->working_disks);
 		}
 	}
-	return 0;
 }
 
 #undef LAST_DISK
@@ -279,11 +272,10 @@ static void print_multipath_conf (multipath_conf_t *conf)
 
 	for (i = 0; i < MD_SB_DISKS; i++) {
 		tmp = conf->multipaths + i;
-		if (tmp->operational || tmp->used_slot)
-			printk(" disk%d, o:%d, us:%d dev:%s\n",
-				i,tmp->operational,
-				tmp->used_slot,
-				bdev_partition_name(tmp->bdev));
+		if (tmp->rdev)
+			printk(" disk%d, o:%d, dev:%s\n",
+				i,!tmp->rdev->faulty,
+			       bdev_partition_name(tmp->rdev->bdev));
 	}
 }
 
@@ -291,24 +283,23 @@ static void print_multipath_conf (multipath_conf_t *conf)
 static int multipath_add_disk(mddev_t *mddev, mdk_rdev_t *rdev)
 {
 	multipath_conf_t *conf = mddev->private;
-	int err = 1;
-	struct multipath_info *p = conf->multipaths + rdev->raid_disk;
+	int found = 0;
+	int path;
+	struct multipath_info *p;
 
 	print_multipath_conf(conf);
 	spin_lock_irq(&conf->device_lock);
-	if (!p->used_slot) {
-		p->bdev = rdev->bdev;
-		p->operational = 1;
-		p->used_slot = 1;
+	for (path=0; path<mddev->raid_disks; path++) 
+		if ((p=conf->multipaths+path)->rdev == NULL) {
+			p->rdev = rdev;
 			conf->working_disks++;
-		err = 0;
+			rdev->raid_disk = path;
+			found = 1;
 		}
-	if (err)
-		MD_BUG();
 	spin_unlock_irq(&conf->device_lock);
 
 	print_multipath_conf(conf);
-	return err;
+	return found;
 }
 
 static int multipath_remove_disk(mddev_t *mddev, int number)
@@ -320,14 +311,14 @@ static int multipath_remove_disk(mddev_t *mddev, int number)
 	print_multipath_conf(conf);
 	spin_lock_irq(&conf->device_lock);
 
-	if (p->used_slot) {
-		if (p->operational) {
+	if (p->rdev) {
+		if (p->rdev->in_sync ||
+		    atomic_read(&p->rdev->nr_pending)) {
 			printk(KERN_ERR "hot-remove-disk, slot %d is identified but is still operational!\n", number);
 			err = -EBUSY;
 			goto abort;
 		}
-		p->bdev = NULL;
-		p->used_slot = 0;
+		p->rdev = NULL;
 		err = 0;
 	}
 	if (err)
@@ -359,7 +350,7 @@ static void multipathd (void *data)
 	struct bio *bio;
 	unsigned long flags;
 	mddev_t *mddev;
-	struct block_device *bdev;
+	mdk_rdev_t *rdev;
 
 	for (;;) {
 		spin_lock_irqsave(&retry_list_lock, flags);
@@ -372,16 +363,16 @@ static void multipathd (void *data)
 		mddev = mp_bh->mddev;
 		bio = &mp_bh->bio;
 		bio->bi_sector = mp_bh->master_bio->bi_sector;
-		bdev = bio->bi_bdev;
 		
-		multipath_map (mddev, &bio->bi_bdev);
-		if (bio->bi_bdev == bdev) {
+		rdev = NULL;
+		if (multipath_map (mddev, &rdev)<0) {
 			printk(IO_ERROR,
 				bdev_partition_name(bio->bi_bdev), bio->bi_sector);
 			multipath_end_bh_io(mp_bh, 0);
 		} else {
 			printk(REDIRECT_SECTOR,
 				bdev_partition_name(bio->bi_bdev), bio->bi_sector);
+			bio->bi_bdev = rdev->bdev;
 			generic_make_request(bio);
 		}
 	}
@@ -436,7 +427,6 @@ static int multipath_run (mddev_t *mddev)
 	struct multipath_info *disk;
 	mdk_rdev_t *rdev;
 	struct list_head *tmp;
-	int num_rdevs = 0;
 
 	MOD_INC_USE_COUNT;
 
@@ -458,40 +448,20 @@ static int multipath_run (mddev_t *mddev)
 	}
 	memset(conf, 0, sizeof(*conf));
 
+	conf->working_disks = 0;
 	ITERATE_RDEV(mddev,rdev,tmp) {
-		if (rdev->faulty) {
-			/* this is a "should never happen" case and if it */
-			/* ever does happen, a continue; won't help */
-			printk(ERRORS, bdev_partition_name(rdev->bdev));
-			continue;
-		} else {
-			/* this is a "should never happen" case and if it */
-			/* ever does happen, a continue; won't help */
-			if (!rdev->sb) {
-				MD_BUG();
-				continue;
-			}
-		}
-		if (rdev->desc_nr == -1) {
-			MD_BUG();
+		disk_idx = rdev->raid_disk;
+		if (disk_idx < 0 ||
+		    disk_idx >= mddev->raid_disks)
 			continue;
-		}
 
-		disk_idx = rdev->raid_disk;
 		disk = conf->multipaths + disk_idx;
-
-		/*
-		 * Mark all disks as active to start with, there are no
-		 * spares.  multipath_read_balance deals with choose
-		 * the "best" operational device.
-		 */
-		disk->bdev = rdev->bdev;
-		disk->operational = 1;
-		disk->used_slot = 1;
-		num_rdevs++;
+		disk->rdev = rdev;
+		if (!rdev->faulty) 
+			conf->working_disks++;
 	}
 
-	conf->raid_disks = mddev->raid_disks = num_rdevs;
+	conf->raid_disks = mddev->raid_disks;
 	mddev->sb_dirty = 1;
 	conf->mddev = mddev;
 	conf->device_lock = SPIN_LOCK_UNLOCKED;
@@ -500,6 +470,7 @@ static int multipath_run (mddev_t *mddev)
 		printk(NONE_OPERATIONAL, mdidx(mddev));
 		goto out_free_conf;
 	}
+	mddev->degraded = conf->raid_disks = conf->working_disks;
 
 	conf->pool = mempool_create(NR_RESERVED_BUFS,
 				    mp_pool_alloc, mp_pool_free,

drivers/md/raid0.c

@@ -87,7 +87,7 @@ static int create_strip_zones (mddev_t *mddev)
 	cnt = 0;
 	smallest = NULL;
 	ITERATE_RDEV(mddev, rdev1, tmp1) {
-		int j = rdev1->sb->this_disk.raid_disk;
+		int j = rdev1->raid_disk;
 
 		if (j < 0 || j >= mddev->raid_disks) {
 			printk("raid0: bad disk number %d - aborting!\n", j);

drivers/net/appletalk/Config.in

@@ -16,3 +16,4 @@ if [ "$CONFIG_DEV_APPLETALK" = "y" ]; then
       bool '    Appletalk-IP to IP Decapsulation support' CONFIG_IPDDP_DECAP
    fi
 fi
+

fs/nfsd/nfs3proc.c

@@ -134,7 +134,7 @@ nfsd3_proc_access(struct svc_rqst *rqstp, struct nfsd3_accessargs *argp,
 
 	fh_copy(&resp->fh, &argp->fh);
 	resp->access = argp->access;
-	nfserr = nfsd_access(rqstp, &resp->fh, &resp->access);
+	nfserr = nfsd_access(rqstp, &resp->fh, &resp->access, NULL);
 	RETURN_STATUS(nfserr);
 }
 
@@ -267,7 +267,7 @@ nfsd3_proc_create(struct svc_rqst *rqstp, struct nfsd3_createargs *argp,
 	/* Now create the file and set attributes */
 	nfserr = nfsd_create_v3(rqstp, dirfhp, argp->name, argp->len,
 				attr, newfhp,
-				argp->createmode, argp->verf);
+				argp->createmode, argp->verf, NULL);
 
 	RETURN_STATUS(nfserr);
 }
@@ -460,7 +460,7 @@ nfsd3_proc_readdir(struct svc_rqst *rqstp, struct nfsd3_readdirargs *argp,
 	fh_copy(&resp->fh, &argp->fh);
 	nfserr = nfsd_readdir(rqstp, &resp->fh, (loff_t) argp->cookie, 
 					nfs3svc_encode_entry,
-					buffer, &count, argp->verf);
+					buffer, &count, argp->verf, NULL);
 	memcpy(resp->verf, argp->verf, 8);
 	resp->count = count;
 
@@ -495,7 +495,7 @@ nfsd3_proc_readdirplus(struct svc_rqst *rqstp, struct nfsd3_readdirargs *argp,
 	fh_copy(&resp->fh, &argp->fh);
 	nfserr = nfsd_readdir(rqstp, &resp->fh, (loff_t) argp->cookie, 
 					nfs3svc_encode_entry_plus,
-					buffer, &count, argp->verf);
+					buffer, &count, argp->verf, NULL);
 	memcpy(resp->verf, argp->verf, 8);
 	resp->count = count;
 

fs/nfsd/nfsfh.c

@@ -107,8 +107,10 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access)
 		int fsid = 0;
 
 		error = nfserr_stale;
-		if (rqstp->rq_vers == 3)
+		if (rqstp->rq_vers > 2)
 			error = nfserr_badhandle;
+		if (rqstp->rq_vers == 4 && fh->fh_size == 0)
+			return nfserr_nofilehandle;
 
 		if (fh->fh_version == 1) {
 			datap = fh->fh_auth;
@@ -171,7 +173,7 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access)
 		 * Look up the dentry using the NFS file handle.
 		 */
 		error = nfserr_stale;
-		if (rqstp->rq_vers == 3)
+		if (rqstp->rq_vers > 2)
 			error = nfserr_badhandle;
 
 		if (fh->fh_version != 1) {
@@ -234,11 +236,23 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access)
 
 	/* Type can be negative when creating hardlinks - not to a dir */
 	if (type > 0 && (inode->i_mode & S_IFMT) != type) {
-		error = (type == S_IFDIR)? nfserr_notdir : nfserr_isdir;
+		if (rqstp->rq_vers == 4 && (inode->i_mode & S_IFMT) == S_IFLNK)
+			error = nfserr_symlink;
+		else if (type == S_IFDIR)
+			error = nfserr_notdir;
+		else if ((inode->i_mode & S_IFMT) == S_IFDIR)
+			error = nfserr_isdir;
+		else
+			error = nfserr_inval;
 		goto out;
 	}
 	if (type < 0 && (inode->i_mode & S_IFMT) == -type) {
-		error = (type == -S_IFDIR)? nfserr_notdir : nfserr_isdir;
+		if (rqstp->rq_vers == 4 && (inode->i_mode & S_IFMT) == S_IFLNK)
+			error = nfserr_symlink;
+		else if (type == -S_IFDIR)
+			error = nfserr_isdir;
+		else
+			error = nfserr_notdir;
 		goto out;
 	}
 
@@ -303,6 +317,8 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry, st
 	 *
 	 */
 
+	u8 ref_fh_version = 0;
+	u8 ref_fh_fsid_type = 0;
 	struct inode * inode = dentry->d_inode;
 	struct dentry *parent = dentry->d_parent;
 	__u32 *datap;
@@ -312,6 +328,13 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry, st
 		parent->d_name.name, dentry->d_name.name,
 		(inode ? inode->i_ino : 0));
 
+	if (ref_fh) {
+		ref_fh_version = ref_fh->fh_handle.fh_version;
+		ref_fh_fsid_type = ref_fh->fh_handle.fh_fsid_type;
+		if (ref_fh == fhp)
+			fh_put(ref_fh);
+	}
+
 	if (fhp->fh_locked || fhp->fh_dentry) {
 		printk(KERN_ERR "fh_compose: fh %s/%s not initialized!\n",
 			parent->d_name.name, dentry->d_name.name);
@@ -323,8 +346,7 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry, st
 	fhp->fh_dentry = dentry; /* our internal copy */
 	fhp->fh_export = exp;
 
-	if (ref_fh &&
-	    ref_fh->fh_handle.fh_version == 0xca) {
+	if (ref_fh_version == 0xca) {
 		/* old style filehandle please */
 		memset(&fhp->fh_handle.fh_base, 0, NFS_FHSIZE);
 		fhp->fh_handle.fh_size = NFS_FHSIZE;
@@ -340,7 +362,7 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry, st
 		fhp->fh_handle.fh_auth_type = 0;
 		datap = fhp->fh_handle.fh_auth+0;
 		if ((exp->ex_flags & NFSEXP_FSID) &&
-		    (!ref_fh || ref_fh->fh_handle.fh_fsid_type == 1)) {
+		    (ref_fh_fsid_type == 1)) {
 			fhp->fh_handle.fh_fsid_type = 1;
 			/* fsid_type 1 == 4 bytes filesystem id */
 			*datap++ = exp->ex_fsid;
@@ -424,6 +446,10 @@ fh_put(struct svc_fh *fhp)
 		fh_unlock(fhp);
 		fhp->fh_dentry = NULL;
 		dput(dentry);
+#ifdef CONFIG_NFSD_V3
+		fhp->fh_pre_saved = 0;
+		fhp->fh_post_saved = 0;
+#endif
 		nfsd_nr_put++;
 	}
 	return;

fs/nfsd/nfsproc.c

@@ -492,7 +492,7 @@ nfsd_proc_readdir(struct svc_rqst *rqstp, struct nfsd_readdirargs *argp,
 	/* Read directory and encode entries on the fly */
 	nfserr = nfsd_readdir(rqstp, &argp->fh, (loff_t) argp->cookie, 
 			      nfssvc_encode_entry,
-			      buffer, &count, NULL);
+			      buffer, &count, NULL, NULL);
 	resp->count = count;
 
 	fh_put(&argp->fh);

fs/nfsd/vfs.c

@@ -348,12 +348,12 @@ static struct accessmap	nfs3_anyaccess[] = {
 };
 
 int
-nfsd_access(struct svc_rqst *rqstp, struct svc_fh *fhp, u32 *access)
+nfsd_access(struct svc_rqst *rqstp, struct svc_fh *fhp, u32 *access, u32 *supported)
 {
 	struct accessmap	*map;
 	struct svc_export	*export;
 	struct dentry		*dentry;
-	u32			query, result = 0;
+	u32			query, result = 0, sresult = 0;
 	unsigned int		error;
 
 	error = fh_verify(rqstp, fhp, 0, MAY_NOP);
@@ -375,6 +375,9 @@ nfsd_access(struct svc_rqst *rqstp, struct svc_fh *fhp, u32 *access)
 	for  (; map->access; map++) {
 		if (map->access & query) {
 			unsigned int err2;
+
+			sresult |= map->access;
+
 			err2 = nfsd_permission(export, dentry, map->how);
 			switch (err2) {
 			case nfs_ok:
@@ -395,6 +398,8 @@ nfsd_access(struct svc_rqst *rqstp, struct svc_fh *fhp, u32 *access)
 		}
 	}
 	*access = result;
+	if (supported)
+		*supported = sresult;
 
  out:
 	return error;
@@ -756,6 +761,9 @@ nfsd_commit(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	struct file	file;
 	int		err;
 
+	if ((u64)count > ~(u64)offset)
+		return nfserr_inval;
+
 	if ((err = nfsd_open(rqstp, fhp, S_IFREG, MAY_WRITE, &file)) != 0)
 		return err;
 	if (EX_ISSYNC(fhp->fh_export)) {
@@ -904,7 +912,8 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
 int
 nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		char *fname, int flen, struct iattr *iap,
-		struct svc_fh *resfhp, int createmode, u32 *verifier)
+		struct svc_fh *resfhp, int createmode, u32 *verifier,
+	        int *truncp)
 {
 	struct dentry	*dentry, *dchild;
 	struct inode	*dirp;
@@ -966,6 +975,16 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp,
 		case NFS3_CREATE_UNCHECKED:
 			if (! S_ISREG(dchild->d_inode->i_mode))
 				err = nfserr_exist;
+			else if (truncp) {
+				/* in nfsv4, we need to treat this case a little
+				 * differently.  we don't want to truncate the
+				 * file now; this would be wrong if the OPEN
+				 * fails for some other reason.  furthermore,
+				 * if the size is nonzero, we should ignore it
+				 * according to spec!
+				 */
+				*truncp = (iap->ia_valid & ATTR_SIZE) && !iap->ia_size;
+			}
 			else {
 				iap->ia_valid &= ATTR_SIZE;
 				goto set_attr;
@@ -1326,6 +1345,9 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
 		goto out;
 	}
 
+	if (!type)
+		type = rdentry->d_inode->i_mode & S_IFMT;
+
 	if (type != S_IFDIR) { /* It's UNLINK */
 #ifdef MSNFS
 		if ((fhp->fh_export->ex_flags & NFSEXP_MSNFS) &&
@@ -1359,7 +1381,7 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
  */
 int
 nfsd_readdir(struct svc_rqst *rqstp, struct svc_fh *fhp, loff_t offset, 
-             encode_dent_fn func, u32 *buffer, int *countp, u32 *verf)
+             encode_dent_fn func, u32 *buffer, int *countp, u32 *verf, u32 *bmval)
 {
 	u32		*p;
 	int		oldlen, eof, err;
@@ -1380,6 +1402,10 @@ nfsd_readdir(struct svc_rqst *rqstp, struct svc_fh *fhp, loff_t offset,
 	cd.buffer = buffer;
 	cd.buflen = *countp; /* count of words */
 	cd.dirfh  = fhp;
+	if (bmval) {
+		cd.bmval[0] = bmval[0];
+		cd.bmval[1] = bmval[1];
+	}
 
 	/*
 	 * Read the directory entries. This silly loop is necessary because
@@ -1395,13 +1421,20 @@ nfsd_readdir(struct svc_rqst *rqstp, struct svc_fh *fhp, loff_t offset,
 		if (err < 0)
 			goto out_nfserr;
 
+		err = cd.nfserr;
+		if (err)
+			goto out_close;
 	} while (oldlen != cd.buflen && !cd.eob);
 
+	err = nfserr_readdir_nospc;
+	if (rqstp->rq_vers == 4 && cd.eob && cd.buffer == buffer)
+		goto out_close;
+
 	/* If we didn't fill the buffer completely, we're at EOF */
 	eof = !cd.eob;
 
 	if (cd.offset) {
-		if (rqstp->rq_vers == 3)
+		if (rqstp->rq_vers > 2)
 			(void)xdr_encode_hyper(cd.offset, file.f_pos);
 		else
 			*cd.offset = htonl(file.f_pos);

include/asm-i386/hardirq.h

@@ -77,8 +77,10 @@ typedef struct {
 #define irq_enter()		(preempt_count() += HARDIRQ_OFFSET)
 
 #if CONFIG_PREEMPT
+# define in_atomic()	(preempt_count() != kernel_locked())
 # define IRQ_EXIT_OFFSET (HARDIRQ_OFFSET-1)
 #else
+# define in_atomic()	(preempt_count() != 0)
 # define IRQ_EXIT_OFFSET HARDIRQ_OFFSET
 #endif
 #define irq_exit()							\

include/linux/nfs.h

@@ -39,39 +39,68 @@
  * standard, but seem to be widely used nevertheless.
  */
  enum nfs_stat {
-	NFS_OK = 0,			/* v2 v3 */
-	NFSERR_PERM = 1,		/* v2 v3 */
-	NFSERR_NOENT = 2,		/* v2 v3 */
-	NFSERR_IO = 5,			/* v2 v3 */
-	NFSERR_NXIO = 6,		/* v2 v3 */
+	NFS_OK = 0,			/* v2 v3 v4 */
+	NFSERR_PERM = 1,		/* v2 v3 v4 */
+	NFSERR_NOENT = 2,		/* v2 v3 v4 */
+	NFSERR_IO = 5,			/* v2 v3 v4 */
+	NFSERR_NXIO = 6,		/* v2 v3 v4 */
 	NFSERR_EAGAIN = 11,		/* v2 v3 */
-	NFSERR_ACCES = 13,		/* v2 v3 */
-	NFSERR_EXIST = 17,		/* v2 v3 */
-	NFSERR_XDEV = 18,		/*    v3 */
-	NFSERR_NODEV = 19,		/* v2 v3 */
-	NFSERR_NOTDIR = 20,		/* v2 v3 */
-	NFSERR_ISDIR = 21,		/* v2 v3 */
-	NFSERR_INVAL = 22,		/* v2 v3 that Sun forgot */
-	NFSERR_FBIG = 27,		/* v2 v3 */
-	NFSERR_NOSPC = 28,		/* v2 v3 */
-	NFSERR_ROFS = 30,		/* v2 v3 */
-	NFSERR_MLINK = 31,		/*    v3 */
+	NFSERR_ACCES = 13,		/* v2 v3 v4 */
+	NFSERR_EXIST = 17,		/* v2 v3 v4 */
+	NFSERR_XDEV = 18,		/*    v3 v4 */
+	NFSERR_NODEV = 19,		/* v2 v3 v4 */
+	NFSERR_NOTDIR = 20,		/* v2 v3 v4 */
+	NFSERR_ISDIR = 21,		/* v2 v3 v4 */
+	NFSERR_INVAL = 22,		/* v2 v3 v4 */
+	NFSERR_FBIG = 27,		/* v2 v3 v4 */
+	NFSERR_NOSPC = 28,		/* v2 v3 v4 */
+	NFSERR_ROFS = 30,		/* v2 v3 v4 */
+	NFSERR_MLINK = 31,		/*    v3 v4 */
 	NFSERR_OPNOTSUPP = 45,		/* v2 v3 */
-	NFSERR_NAMETOOLONG = 63,	/* v2 v3 */
-	NFSERR_NOTEMPTY = 66,		/* v2 v3 */
-	NFSERR_DQUOT = 69,		/* v2 v3 */
-	NFSERR_STALE = 70,		/* v2 v3 */
+	NFSERR_NAMETOOLONG = 63,	/* v2 v3 v4 */
+	NFSERR_NOTEMPTY = 66,		/* v2 v3 v4 */
+	NFSERR_DQUOT = 69,		/* v2 v3 v4 */
+	NFSERR_STALE = 70,		/* v2 v3 v4 */
 	NFSERR_REMOTE = 71,		/* v2 v3 */
 	NFSERR_WFLUSH = 99,		/* v2    */
-	NFSERR_BADHANDLE = 10001,	/*    v3 */
+	NFSERR_BADHANDLE = 10001,	/*    v3 v4 */
 	NFSERR_NOT_SYNC = 10002,	/*    v3 */
-	NFSERR_BAD_COOKIE = 10003,	/*    v3 */
-	NFSERR_NOTSUPP = 10004,		/*    v3 */
-	NFSERR_TOOSMALL = 10005,	/*    v3 */
-	NFSERR_SERVERFAULT = 10006,	/*    v3 */
-	NFSERR_BADTYPE = 10007,		/*    v3 */
-	NFSERR_JUKEBOX = 10008		/*    v3 */
- };
+	NFSERR_BAD_COOKIE = 10003,	/*    v3 v4 */
+	NFSERR_NOTSUPP = 10004,		/*    v3 v4 */
+	NFSERR_TOOSMALL = 10005,	/*    v3 v4 */
+	NFSERR_SERVERFAULT = 10006,	/*    v3 v4 */
+	NFSERR_BADTYPE = 10007,		/*    v3 v4 */
+	NFSERR_JUKEBOX = 10008,		/*    v3 v4 */
+	NFSERR_SAME = 10009,		/*       v4 */
+	NFSERR_DENIED = 10010,		/*       v4 */
+	NFSERR_EXPIRED = 10011,		/*       v4 */
+	NFSERR_LOCKED = 10012,		/*       v4 */
+	NFSERR_GRACE = 10013,		/*       v4 */
+	NFSERR_FHEXPIRED = 10014,	/*       v4 */
+	NFSERR_SHARE_DENIED = 10015,	/*       v4 */
+	NFSERR_WRONGSEC = 10016,	/*       v4 */
+	NFSERR_CLID_INUSE = 10017,	/*       v4 */
+	NFSERR_RESOURCE = 10018,	/*       v4 */
+	NFSERR_MOVED = 10019,		/*       v4 */
+	NFSERR_NOFILEHANDLE = 10020,	/*       v4 */
+	NFSERR_MINOR_VERS_MISMATCH = 10021,   /* v4 */
+	NFSERR_STALE_CLIENTID = 10022,	/*       v4 */
+	NFSERR_STALE_STATEID = 10023,	/*       v4 */
+	NFSERR_OLD_STATEID = 10024,	/*       v4 */
+	NFSERR_BAD_STATEID = 10025,	/*       v4 */
+	NFSERR_BAD_SEQID = 10026,	/*       v4 */
+	NFSERR_NOT_SAME = 10027,	/*       v4 */
+	NFSERR_LOCK_RANGE = 10028,	/*       v4 */
+	NFSERR_SYMLINK = 10029,		/*       v4 */
+	NFSERR_READDIR_NOSPC = 10030,	/*       v4 */
+	NFSERR_LEASE_MOVED = 10031,	/*       v4 */
+	NFSERR_ATTRNOTSUPP = 10032,	/*       v4 */
+	NFSERR_NO_GRACE = 10033,	/*       v4 */
+	NFSERR_RECLAIM_BAD = 10034,	/*       v4 */
+	NFSERR_RECLAIM_CONFLICT = 10035,/*       v4 */
+	NFSERR_BAD_XDR = 10036,		/*       v4 */
+	NFSERR_LOCKS_HELD = 10037	/*       v4 */
+};
 
 /* NFSv2 file types - beware, these are not the same in NFSv3 */
 

include/linux/nfsd/nfsd.h

@@ -55,6 +55,8 @@ struct readdir_cd {
 	char			plus;		/* readdirplus */
 	char			eob;		/* end of buffer */
 	char			dotonly;
+	int			nfserr;		/* v4 only */
+	u32			bmval[2];	/* v4 only */
 };
 typedef int		(*encode_dent_fn)(struct readdir_cd *, const char *,
 						int, loff_t, ino_t, unsigned int);
@@ -86,11 +88,11 @@ int		nfsd_create(struct svc_rqst *, struct svc_fh *,
 				char *name, int len, struct iattr *attrs,
 				int type, dev_t rdev, struct svc_fh *res);
 #ifdef CONFIG_NFSD_V3
-int		nfsd_access(struct svc_rqst *, struct svc_fh *, u32 *);
+int		nfsd_access(struct svc_rqst *, struct svc_fh *, u32 *, u32 *);
 int		nfsd_create_v3(struct svc_rqst *, struct svc_fh *,
 				char *name, int len, struct iattr *attrs,
 				struct svc_fh *res, int createmode,
-				u32 *verifier);
+				u32 *verifier, int *truncp);
 int		nfsd_commit(struct svc_rqst *, struct svc_fh *,
 				off_t, unsigned long);
 #endif /* CONFIG_NFSD_V3 */
@@ -119,7 +121,8 @@ int		nfsd_truncate(struct svc_rqst *, struct svc_fh *,
 				unsigned long size);
 int		nfsd_readdir(struct svc_rqst *, struct svc_fh *,
 				loff_t, encode_dent_fn,
-				u32 *buffer, int *countp, u32 *verf);
+				u32 *buffer, int *countp, u32 *verf,
+				u32 *bmval);
 int		nfsd_statfs(struct svc_rqst *, struct svc_fh *,
 				struct statfs *);
 
@@ -170,6 +173,16 @@ void		nfsd_lockd_unexport(struct svc_client *);
 #define	nfserr_serverfault	__constant_htonl(NFSERR_SERVERFAULT)
 #define	nfserr_badtype		__constant_htonl(NFSERR_BADTYPE)
 #define	nfserr_jukebox		__constant_htonl(NFSERR_JUKEBOX)
+#define	nfserr_bad_cookie	__constant_htonl(NFSERR_BAD_COOKIE)
+#define	nfserr_same		__constant_htonl(NFSERR_SAME)
+#define	nfserr_clid_inuse	__constant_htonl(NFSERR_CLID_INUSE)
+#define	nfserr_resource		__constant_htonl(NFSERR_RESOURCE)
+#define	nfserr_nofilehandle	__constant_htonl(NFSERR_NOFILEHANDLE)
+#define	nfserr_minor_vers_mismatch	__constant_htonl(NFSERR_MINOR_VERS_MISMATCH)
+#define	nfserr_symlink		__constant_htonl(NFSERR_SYMLINK)
+#define	nfserr_not_same		__constant_htonl(NFSERR_NOT_SAME)
+#define	nfserr_readdir_nospc	__constant_htonl(NFSERR_READDIR_NOSPC)
+#define	nfserr_bad_xdr		__constant_htonl(NFSERR_BAD_XDR)
 
 /* error code for internal use - if a request fails due to
  * kmalloc failure, it gets dropped.  Client should resend eventually

include/linux/raid/linear.h

@@ -4,7 +4,7 @@
 #include <linux/raid/md.h>
 
 struct dev_info {
-	struct block_device *bdev;
+	mdk_rdev_t	*rdev;
 	unsigned long	size;
 	unsigned long	offset;
 };

include/linux/raid/md.h

@@ -76,9 +76,8 @@ extern void md_unregister_thread (mdk_thread_t *thread);
 extern void md_wakeup_thread(mdk_thread_t *thread);
 extern void md_interrupt_thread (mdk_thread_t *thread);
 extern void md_done_sync(mddev_t *mddev, int blocks, int ok);
-extern void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors);
-extern int md_error (mddev_t *mddev, struct block_device *bdev);
-extern int md_run_setup(void);
+extern void md_sync_acct(mdk_rdev_t *rdev, unsigned long nr_sectors);
+extern void md_error (mddev_t *mddev, mdk_rdev_t *rdev);
 
 extern void md_print_devices (void);
 

include/linux/raid/md_k.h

@@ -154,12 +154,26 @@ struct mdk_rdev_s
 	mdp_super_t	*sb;
 	unsigned long	sb_offset;
 
-	int alias_device;		/* device alias to the same disk */
+	/* A device can be in one of three states based on two flags:
+	 * Not working:   faulty==1 in_sync==0
+	 * Fully working: faulty==0 in_sync==1
+	 * Working, but not
+	 * in sync with array
+	 *                faulty==0 in_sync==0
+	 *
+	 * It can never have faulty==1, in_sync==1
+	 * This reduces the burden of testing multiple flags in many cases
+	 */
 	int faulty;			/* if faulty do not issue IO requests */
 	int in_sync;			/* device is a full member of the array */
 
 	int desc_nr;			/* descriptor index in the superblock */
 	int raid_disk;			/* role of device in array */
+
+	atomic_t	nr_pending;	/* number of pending requests.
+					 * only maintained for arrays that
+					 * support hot removal
+					 */
 };
 
 typedef struct mdk_personality_s mdk_personality_t;
@@ -202,7 +216,7 @@ struct mddev_s
 	int				in_sync;	/* know to not need resync */
 	struct semaphore		reconfig_sem;
 	atomic_t			active;
-	mdk_rdev_t			*spare;
+	int				spares;
 
 	int				degraded;	/* whether md should consider
 							 * adding a spare
@@ -223,11 +237,12 @@ struct mdk_personality_s
 	int (*run)(mddev_t *mddev);
 	int (*stop)(mddev_t *mddev);
 	int (*status)(char *page, mddev_t *mddev);
-	int (*error_handler)(mddev_t *mddev, struct block_device *bdev);
+	/* error_handler must set ->faulty and clear ->in_sync
+	 * if appropriate, and should abort recovery if needed 
+	 */
+	void (*error_handler)(mddev_t *mddev, mdk_rdev_t *rdev);
 	int (*hot_add_disk) (mddev_t *mddev, mdk_rdev_t *rdev);
 	int (*hot_remove_disk) (mddev_t *mddev, int number);
-	int (*spare_write) (mddev_t *mddev);
-	int (*spare_inactive) (mddev_t *mddev);
 	int (*spare_active) (mddev_t *mddev);
 	int (*sync_request)(mddev_t *mddev, sector_t sector_nr, int go_faster);
 };
@@ -272,9 +287,6 @@ extern mdk_rdev_t * find_rdev_nr(mddev_t *mddev, int nr);
 #define ITERATE_RDEV_PENDING(rdev,tmp)					\
 	ITERATE_RDEV_GENERIC(pending_raid_disks,rdev,tmp)
 
-#define xchg_values(x,y) do { __typeof__(x) __tmp = x; \
-				x = y; y = __tmp; } while (0)
-
 typedef struct mdk_thread_s {
 	void			(*run) (void *data);
 	void			*data;

include/linux/raid/multipath.h

@@ -5,14 +5,7 @@
 #include <linux/bio.h>
 
 struct multipath_info {
-	struct block_device *bdev;
-
-	/*
-	 * State bits:
-	 */
-	int		operational;
-
-	int		used_slot;
+	mdk_rdev_t	*rdev;
 };
 
 struct multipath_private_data {

include/linux/raid/raid1.h

@@ -6,18 +6,8 @@
 typedef struct mirror_info mirror_info_t;
 
 struct mirror_info {
-	struct block_device *bdev;
+	mdk_rdev_t	*rdev;
 	sector_t	head_position;
-	atomic_t	nr_pending;
-
-	/*
-	 * State bits:
-	 */
-	int		operational;
-	int		write_only;
-	int		spare;
-
-	int		used_slot;
 };
 
 typedef struct r1bio_s r1bio_t;
@@ -30,7 +20,6 @@ struct r1_private_data_s {
 	int			last_used;
 	sector_t		next_seq_sect;
 	mdk_thread_t		*thread;
-	mirror_info_t		*spare;
 	spinlock_t		device_lock;
 
 	/* for use when syncing mirrors: */

include/linux/raid/raid5.h

@@ -148,6 +148,11 @@ struct stripe_head {
 #define	R5_UPTODATE	0	/* page contains current data */
 #define	R5_LOCKED	1	/* IO has been submitted on "req" */
 #define	R5_OVERWRITE	2	/* towrite covers whole page */
+/* and some that are internal to handle_stripe */
+#define	R5_Insync	3	/* rdev && rdev->in_sync at start */
+#define	R5_Wantread	4	/* want to schedule a read */
+#define	R5_Wantwrite	5
+#define	R5_Syncio	6	/* this io need to be accounted as resync io */
 
 /*
  * Write method
@@ -192,11 +197,7 @@ struct stripe_head {
  
 
 struct disk_info {
-	struct block_device *bdev;
-	int	operational;
-	int	write_only;
-	int	spare;
-	int	used_slot;
+	mdk_rdev_t	*rdev;
 };
 
 struct raid5_private_data {

mm/swap.c

@@ -76,9 +76,8 @@ void lru_add_drain(void)
  */
 void __page_cache_release(struct page *page)
 {
-	BUG_ON(page_count(page) != 0);
-	if (PageLRU(page)) {
 	unsigned long flags;
+	BUG_ON(page_count(page) != 0);
 
 	spin_lock_irqsave(&_pagemap_lru_lock, flags);
 	if (TestClearPageLRU(page)) {
@@ -90,7 +89,6 @@ void __page_cache_release(struct page *page)
 	if (page_count(page) != 0)
 		page = NULL;
 	spin_unlock_irqrestore(&_pagemap_lru_lock, flags);
-	}
 	if (page)
 		__free_pages_ok(page, 0);
 }

mm/vmalloc.c

@@ -153,15 +153,20 @@ int map_vm_area(struct vm_struct *area, pgprot_t prot, struct page ***pages)
 	unsigned long address = VMALLOC_VMADDR(area->addr);
 	unsigned long end = address + (area->size-PAGE_SIZE);
 	pgd_t *dir;
+	int err = 0;
 
 	dir = pgd_offset_k(address);
 	spin_lock(&init_mm.page_table_lock);
 	do {
 		pmd_t *pmd = pmd_alloc(&init_mm, dir, address);
-		if (!pmd)
-			return -ENOMEM;
-		if (map_area_pmd(pmd, address, end - address, prot, pages))
-			return -ENOMEM;
+		if (!pmd) {
+			err = -ENOMEM;
+			break;
+		}
+		if (map_area_pmd(pmd, address, end - address, prot, pages)) {
+			err = -ENOMEM;
+			break;
+		}
 
 		address = (address + PGDIR_SIZE) & PGDIR_MASK;
 		dir++;
@@ -169,7 +174,7 @@ int map_vm_area(struct vm_struct *area, pgprot_t prot, struct page ***pages)
 
 	spin_unlock(&init_mm.page_table_lock);
 	flush_cache_all();
-	return 0;
+	return err;
 }
 
 
@@ -379,15 +384,21 @@ void *__vmalloc(unsigned long size, int gfp_mask, pgprot_t prot)
 
 	area->nr_pages = nr_pages;
 	area->pages = pages = kmalloc(array_size, (gfp_mask & ~__GFP_HIGHMEM));
-	if (!area->pages)
+	if (!area->pages) {
+		remove_vm_area(area->addr);
+		kfree(area);
 		return NULL;
+	}
 	memset(area->pages, 0, array_size);
 
 	for (i = 0; i < area->nr_pages; i++) {
 		area->pages[i] = alloc_page(gfp_mask);
-		if (unlikely(!area->pages[i]))
+		if (unlikely(!area->pages[i])) {
+			/* Successfully allocated i pages, free them in __vunmap() */
+			area->nr_pages = i;
 			goto fail;
 		}
+	}
 	
 	if (map_vm_area(area, prot, &pages))
 		goto fail;

net/sunrpc/svcsock.c

@@ -140,7 +140,7 @@ svc_sock_enqueue(struct svc_sock *svsk)
 	    && !test_bit(SK_CLOSE, &svsk->sk_flags)
 	    && !test_bit(SK_CONN, &svsk->sk_flags)) {
 		/* Don't enqueue while not enough space for reply */
-		dprintk("svc: socket %p  no space, %d > %ld, not enqueued\n",
+		dprintk("svc: socket %p  no space, %d*2 > %ld, not enqueued\n",
 			svsk->sk_sk, svsk->sk_reserved+serv->sv_bufsz,
 			sock_wspace(svsk->sk_sk));
 		goto out_unlock;
@@ -574,6 +574,14 @@ svc_udp_init(struct svc_sock *svsk)
 	svsk->sk_recvfrom = svc_udp_recvfrom;
 	svsk->sk_sendto = svc_udp_sendto;
 
+	/* initialise setting must have enough space to
+	 * receive and respond to one request.  
+	 * svc_udp_recvfrom will re-adjust if necessary
+	 */
+	svc_sock_setbufsize(svsk->sk_sock,
+			    3 * svsk->sk_server->sv_bufsz,
+			    3 * svsk->sk_server->sv_bufsz);
+
 	set_bit(SK_CHNGBUF, &svsk->sk_flags);
 
 	return 0;
@@ -679,6 +687,8 @@ svc_tcp_accept(struct svc_sock *svsk)
 		goto failed;		/* aborted connection or whatever */
 	}
 	set_bit(SK_CONN, &svsk->sk_flags);
+	svc_sock_enqueue(svsk);
+
 	slen = sizeof(sin);
 	err = ops->getname(newsock, (struct sockaddr *) &sin, &slen, 1);
 	if (err < 0) {
@@ -944,6 +954,14 @@ svc_tcp_init(struct svc_sock *svsk)
 		svsk->sk_reclen = 0;
 		svsk->sk_tcplen = 0;
 
+		/* initialise setting must have enough space to
+		 * receive and respond to one request.  
+		 * svc_tcp_recvfrom will re-adjust if necessary
+		 */
+		svc_sock_setbufsize(svsk->sk_sock,
+				    3 * svsk->sk_server->sv_bufsz,
+				    3 * svsk->sk_server->sv_bufsz);
+
 		set_bit(SK_CHNGBUF, &svsk->sk_flags);
 	}
 
@@ -1220,7 +1238,7 @@ svc_create_socket(struct svc_serv *serv, int protocol, struct sockaddr_in *sin)
 	}
 
 	if (protocol == IPPROTO_TCP) {
-		if ((error = sock->ops->listen(sock, 5)) < 0)
+		if ((error = sock->ops->listen(sock, 64)) < 0)
 			goto bummer;
 	}