https://github.com/torvalds/linux
Revision 4fca50d440cc5d4dc570ad5484cc0b70b381bc2a authored by Jan Kara on 08 September 2022, 09:21:24 UTC, committed by Theodore Ts'o on 22 September 2022, 02:11:34 UTC
One of the side-effects of mb_optimize_scan was that the optimized
functions to select next group to try were called even before we tried
the goal group. As a result we no longer allocate files close to
corresponding inodes as well as we don't try to expand currently
allocated extent in the same group. This results in reaim regression
with workfile.disk workload of upto 8% with many clients on my test
machine:

                     baseline               mb_optimize_scan
Hmean     disk-1       2114.16 (   0.00%)     2099.37 (  -0.70%)
Hmean     disk-41     87794.43 (   0.00%)    83787.47 *  -4.56%*
Hmean     disk-81    148170.73 (   0.00%)   135527.05 *  -8.53%*
Hmean     disk-121   177506.11 (   0.00%)   166284.93 *  -6.32%*
Hmean     disk-161   220951.51 (   0.00%)   207563.39 *  -6.06%*
Hmean     disk-201   208722.74 (   0.00%)   203235.59 (  -2.63%)
Hmean     disk-241   222051.60 (   0.00%)   217705.51 (  -1.96%)
Hmean     disk-281   252244.17 (   0.00%)   241132.72 *  -4.41%*
Hmean     disk-321   255844.84 (   0.00%)   245412.84 *  -4.08%*

Also this is causing huge regression (time increased by a factor of 5 or
so) when untarring archive with lots of small files on some eMMC storage
cards.

Fix the problem by making sure we try goal group first.

Fixes: 196e402adf2e ("ext4: improve cr 0 / cr 1 group scanning")
CC: stable@kernel.org
Reported-and-tested-by: Stefan Wahren <stefan.wahren@i2se.com>
Tested-by: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/all/20220727105123.ckwrhbilzrxqpt24@quack3/
Link: https://lore.kernel.org/all/0d81a7c2-46b7-6010-62a4-3e6cfc1628d6@i2se.com/
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20220908092136.11770-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
1 parent 7e18e42
Raw File
Tip revision: 4fca50d440cc5d4dc570ad5484cc0b70b381bc2a authored by Jan Kara on 08 September 2022, 09:21:24 UTC
ext4: make mballoc try target group first even with mb_optimize_scan
Tip revision: 4fca50d
crypto_user_base.c
// SPDX-License-Identifier: GPL-2.0-only
/*
 * Crypto user configuration API.
 *
 * Copyright (C) 2011 secunet Security Networks AG
 * Copyright (C) 2011 Steffen Klassert <steffen.klassert@secunet.com>
 */

#include <linux/module.h>
#include <linux/crypto.h>
#include <linux/cryptouser.h>
#include <linux/sched.h>
#include <linux/security.h>
#include <net/netlink.h>
#include <net/net_namespace.h>
#include <net/sock.h>
#include <crypto/internal/skcipher.h>
#include <crypto/internal/rng.h>
#include <crypto/akcipher.h>
#include <crypto/kpp.h>
#include <crypto/internal/cryptouser.h>

#include "internal.h"

#define null_terminated(x)	(strnlen(x, sizeof(x)) < sizeof(x))

static DEFINE_MUTEX(crypto_cfg_mutex);

struct crypto_dump_info {
	struct sk_buff *in_skb;
	struct sk_buff *out_skb;
	u32 nlmsg_seq;
	u16 nlmsg_flags;
};

struct crypto_alg *crypto_alg_match(struct crypto_user_alg *p, int exact)
{
	struct crypto_alg *q, *alg = NULL;

	down_read(&crypto_alg_sem);

	list_for_each_entry(q, &crypto_alg_list, cra_list) {
		int match = 0;

		if (crypto_is_larval(q))
			continue;

		if ((q->cra_flags ^ p->cru_type) & p->cru_mask)
			continue;

		if (strlen(p->cru_driver_name))
			match = !strcmp(q->cra_driver_name,
					p->cru_driver_name);
		else if (!exact)
			match = !strcmp(q->cra_name, p->cru_name);

		if (!match)
			continue;

		if (unlikely(!crypto_mod_get(q)))
			continue;

		alg = q;
		break;
	}

	up_read(&crypto_alg_sem);

	return alg;
}

static int crypto_report_cipher(struct sk_buff *skb, struct crypto_alg *alg)
{
	struct crypto_report_cipher rcipher;

	memset(&rcipher, 0, sizeof(rcipher));

	strscpy(rcipher.type, "cipher", sizeof(rcipher.type));

	rcipher.blocksize = alg->cra_blocksize;
	rcipher.min_keysize = alg->cra_cipher.cia_min_keysize;
	rcipher.max_keysize = alg->cra_cipher.cia_max_keysize;

	return nla_put(skb, CRYPTOCFGA_REPORT_CIPHER,
		       sizeof(rcipher), &rcipher);
}

static int crypto_report_comp(struct sk_buff *skb, struct crypto_alg *alg)
{
	struct crypto_report_comp rcomp;

	memset(&rcomp, 0, sizeof(rcomp));

	strscpy(rcomp.type, "compression", sizeof(rcomp.type));

	return nla_put(skb, CRYPTOCFGA_REPORT_COMPRESS, sizeof(rcomp), &rcomp);
}

static int crypto_report_one(struct crypto_alg *alg,
			     struct crypto_user_alg *ualg, struct sk_buff *skb)
{
	memset(ualg, 0, sizeof(*ualg));

	strscpy(ualg->cru_name, alg->cra_name, sizeof(ualg->cru_name));
	strscpy(ualg->cru_driver_name, alg->cra_driver_name,
		sizeof(ualg->cru_driver_name));
	strscpy(ualg->cru_module_name, module_name(alg->cra_module),
		sizeof(ualg->cru_module_name));

	ualg->cru_type = 0;
	ualg->cru_mask = 0;
	ualg->cru_flags = alg->cra_flags;
	ualg->cru_refcnt = refcount_read(&alg->cra_refcnt);

	if (nla_put_u32(skb, CRYPTOCFGA_PRIORITY_VAL, alg->cra_priority))
		goto nla_put_failure;
	if (alg->cra_flags & CRYPTO_ALG_LARVAL) {
		struct crypto_report_larval rl;

		memset(&rl, 0, sizeof(rl));
		strscpy(rl.type, "larval", sizeof(rl.type));
		if (nla_put(skb, CRYPTOCFGA_REPORT_LARVAL, sizeof(rl), &rl))
			goto nla_put_failure;
		goto out;
	}

	if (alg->cra_type && alg->cra_type->report) {
		if (alg->cra_type->report(skb, alg))
			goto nla_put_failure;

		goto out;
	}

	switch (alg->cra_flags & (CRYPTO_ALG_TYPE_MASK | CRYPTO_ALG_LARVAL)) {
	case CRYPTO_ALG_TYPE_CIPHER:
		if (crypto_report_cipher(skb, alg))
			goto nla_put_failure;

		break;
	case CRYPTO_ALG_TYPE_COMPRESS:
		if (crypto_report_comp(skb, alg))
			goto nla_put_failure;

		break;
	}

out:
	return 0;

nla_put_failure:
	return -EMSGSIZE;
}

static int crypto_report_alg(struct crypto_alg *alg,
			     struct crypto_dump_info *info)
{
	struct sk_buff *in_skb = info->in_skb;
	struct sk_buff *skb = info->out_skb;
	struct nlmsghdr *nlh;
	struct crypto_user_alg *ualg;
	int err = 0;

	nlh = nlmsg_put(skb, NETLINK_CB(in_skb).portid, info->nlmsg_seq,
			CRYPTO_MSG_GETALG, sizeof(*ualg), info->nlmsg_flags);
	if (!nlh) {
		err = -EMSGSIZE;
		goto out;
	}

	ualg = nlmsg_data(nlh);

	err = crypto_report_one(alg, ualg, skb);
	if (err) {
		nlmsg_cancel(skb, nlh);
		goto out;
	}

	nlmsg_end(skb, nlh);

out:
	return err;
}

static int crypto_report(struct sk_buff *in_skb, struct nlmsghdr *in_nlh,
			 struct nlattr **attrs)
{
	struct net *net = sock_net(in_skb->sk);
	struct crypto_user_alg *p = nlmsg_data(in_nlh);
	struct crypto_alg *alg;
	struct sk_buff *skb;
	struct crypto_dump_info info;
	int err;

	if (!null_terminated(p->cru_name) || !null_terminated(p->cru_driver_name))
		return -EINVAL;

	alg = crypto_alg_match(p, 0);
	if (!alg)
		return -ENOENT;

	err = -ENOMEM;
	skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
	if (!skb)
		goto drop_alg;

	info.in_skb = in_skb;
	info.out_skb = skb;
	info.nlmsg_seq = in_nlh->nlmsg_seq;
	info.nlmsg_flags = 0;

	err = crypto_report_alg(alg, &info);

drop_alg:
	crypto_mod_put(alg);

	if (err) {
		kfree_skb(skb);
		return err;
	}

	return nlmsg_unicast(net->crypto_nlsk, skb, NETLINK_CB(in_skb).portid);
}

static int crypto_dump_report(struct sk_buff *skb, struct netlink_callback *cb)
{
	const size_t start_pos = cb->args[0];
	size_t pos = 0;
	struct crypto_dump_info info;
	struct crypto_alg *alg;
	int res;

	info.in_skb = cb->skb;
	info.out_skb = skb;
	info.nlmsg_seq = cb->nlh->nlmsg_seq;
	info.nlmsg_flags = NLM_F_MULTI;

	down_read(&crypto_alg_sem);
	list_for_each_entry(alg, &crypto_alg_list, cra_list) {
		if (pos >= start_pos) {
			res = crypto_report_alg(alg, &info);
			if (res == -EMSGSIZE)
				break;
			if (res)
				goto out;
		}
		pos++;
	}
	cb->args[0] = pos;
	res = skb->len;
out:
	up_read(&crypto_alg_sem);
	return res;
}

static int crypto_dump_report_done(struct netlink_callback *cb)
{
	return 0;
}

static int crypto_update_alg(struct sk_buff *skb, struct nlmsghdr *nlh,
			     struct nlattr **attrs)
{
	struct crypto_alg *alg;
	struct crypto_user_alg *p = nlmsg_data(nlh);
	struct nlattr *priority = attrs[CRYPTOCFGA_PRIORITY_VAL];
	LIST_HEAD(list);

	if (!netlink_capable(skb, CAP_NET_ADMIN))
		return -EPERM;

	if (!null_terminated(p->cru_name) || !null_terminated(p->cru_driver_name))
		return -EINVAL;

	if (priority && !strlen(p->cru_driver_name))
		return -EINVAL;

	alg = crypto_alg_match(p, 1);
	if (!alg)
		return -ENOENT;

	down_write(&crypto_alg_sem);

	crypto_remove_spawns(alg, &list, NULL);

	if (priority)
		alg->cra_priority = nla_get_u32(priority);

	up_write(&crypto_alg_sem);

	crypto_mod_put(alg);
	crypto_remove_final(&list);

	return 0;
}

static int crypto_del_alg(struct sk_buff *skb, struct nlmsghdr *nlh,
			  struct nlattr **attrs)
{
	struct crypto_alg *alg;
	struct crypto_user_alg *p = nlmsg_data(nlh);
	int err;

	if (!netlink_capable(skb, CAP_NET_ADMIN))
		return -EPERM;

	if (!null_terminated(p->cru_name) || !null_terminated(p->cru_driver_name))
		return -EINVAL;

	alg = crypto_alg_match(p, 1);
	if (!alg)
		return -ENOENT;

	/* We can not unregister core algorithms such as aes-generic.
	 * We would loose the reference in the crypto_alg_list to this algorithm
	 * if we try to unregister. Unregistering such an algorithm without
	 * removing the module is not possible, so we restrict to crypto
	 * instances that are build from templates. */
	err = -EINVAL;
	if (!(alg->cra_flags & CRYPTO_ALG_INSTANCE))
		goto drop_alg;

	err = -EBUSY;
	if (refcount_read(&alg->cra_refcnt) > 2)
		goto drop_alg;

	crypto_unregister_instance((struct crypto_instance *)alg);
	err = 0;

drop_alg:
	crypto_mod_put(alg);
	return err;
}

static int crypto_add_alg(struct sk_buff *skb, struct nlmsghdr *nlh,
			  struct nlattr **attrs)
{
	int exact = 0;
	const char *name;
	struct crypto_alg *alg;
	struct crypto_user_alg *p = nlmsg_data(nlh);
	struct nlattr *priority = attrs[CRYPTOCFGA_PRIORITY_VAL];

	if (!netlink_capable(skb, CAP_NET_ADMIN))
		return -EPERM;

	if (!null_terminated(p->cru_name) || !null_terminated(p->cru_driver_name))
		return -EINVAL;

	if (strlen(p->cru_driver_name))
		exact = 1;

	if (priority && !exact)
		return -EINVAL;

	alg = crypto_alg_match(p, exact);
	if (alg) {
		crypto_mod_put(alg);
		return -EEXIST;
	}

	if (strlen(p->cru_driver_name))
		name = p->cru_driver_name;
	else
		name = p->cru_name;

	alg = crypto_alg_mod_lookup(name, p->cru_type, p->cru_mask);
	if (IS_ERR(alg))
		return PTR_ERR(alg);

	down_write(&crypto_alg_sem);

	if (priority)
		alg->cra_priority = nla_get_u32(priority);

	up_write(&crypto_alg_sem);

	crypto_mod_put(alg);

	return 0;
}

static int crypto_del_rng(struct sk_buff *skb, struct nlmsghdr *nlh,
			  struct nlattr **attrs)
{
	if (!netlink_capable(skb, CAP_NET_ADMIN))
		return -EPERM;
	return crypto_del_default_rng();
}

#define MSGSIZE(type) sizeof(struct type)

static const int crypto_msg_min[CRYPTO_NR_MSGTYPES] = {
	[CRYPTO_MSG_NEWALG	- CRYPTO_MSG_BASE] = MSGSIZE(crypto_user_alg),
	[CRYPTO_MSG_DELALG	- CRYPTO_MSG_BASE] = MSGSIZE(crypto_user_alg),
	[CRYPTO_MSG_UPDATEALG	- CRYPTO_MSG_BASE] = MSGSIZE(crypto_user_alg),
	[CRYPTO_MSG_GETALG	- CRYPTO_MSG_BASE] = MSGSIZE(crypto_user_alg),
	[CRYPTO_MSG_DELRNG	- CRYPTO_MSG_BASE] = 0,
	[CRYPTO_MSG_GETSTAT	- CRYPTO_MSG_BASE] = MSGSIZE(crypto_user_alg),
};

static const struct nla_policy crypto_policy[CRYPTOCFGA_MAX+1] = {
	[CRYPTOCFGA_PRIORITY_VAL]   = { .type = NLA_U32},
};

#undef MSGSIZE

static const struct crypto_link {
	int (*doit)(struct sk_buff *, struct nlmsghdr *, struct nlattr **);
	int (*dump)(struct sk_buff *, struct netlink_callback *);
	int (*done)(struct netlink_callback *);
} crypto_dispatch[CRYPTO_NR_MSGTYPES] = {
	[CRYPTO_MSG_NEWALG	- CRYPTO_MSG_BASE] = { .doit = crypto_add_alg},
	[CRYPTO_MSG_DELALG	- CRYPTO_MSG_BASE] = { .doit = crypto_del_alg},
	[CRYPTO_MSG_UPDATEALG	- CRYPTO_MSG_BASE] = { .doit = crypto_update_alg},
	[CRYPTO_MSG_GETALG	- CRYPTO_MSG_BASE] = { .doit = crypto_report,
						       .dump = crypto_dump_report,
						       .done = crypto_dump_report_done},
	[CRYPTO_MSG_DELRNG	- CRYPTO_MSG_BASE] = { .doit = crypto_del_rng },
	[CRYPTO_MSG_GETSTAT	- CRYPTO_MSG_BASE] = { .doit = crypto_reportstat},
};

static int crypto_user_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh,
			       struct netlink_ext_ack *extack)
{
	struct net *net = sock_net(skb->sk);
	struct nlattr *attrs[CRYPTOCFGA_MAX+1];
	const struct crypto_link *link;
	int type, err;

	type = nlh->nlmsg_type;
	if (type > CRYPTO_MSG_MAX)
		return -EINVAL;

	type -= CRYPTO_MSG_BASE;
	link = &crypto_dispatch[type];

	if ((type == (CRYPTO_MSG_GETALG - CRYPTO_MSG_BASE) &&
	    (nlh->nlmsg_flags & NLM_F_DUMP))) {
		struct crypto_alg *alg;
		unsigned long dump_alloc = 0;

		if (link->dump == NULL)
			return -EINVAL;

		down_read(&crypto_alg_sem);
		list_for_each_entry(alg, &crypto_alg_list, cra_list)
			dump_alloc += CRYPTO_REPORT_MAXSIZE;
		up_read(&crypto_alg_sem);

		{
			struct netlink_dump_control c = {
				.dump = link->dump,
				.done = link->done,
				.min_dump_alloc = min(dump_alloc, 65535UL),
			};
			err = netlink_dump_start(net->crypto_nlsk, skb, nlh, &c);
		}

		return err;
	}

	err = nlmsg_parse_deprecated(nlh, crypto_msg_min[type], attrs,
				     CRYPTOCFGA_MAX, crypto_policy, extack);
	if (err < 0)
		return err;

	if (link->doit == NULL)
		return -EINVAL;

	return link->doit(skb, nlh, attrs);
}

static void crypto_netlink_rcv(struct sk_buff *skb)
{
	mutex_lock(&crypto_cfg_mutex);
	netlink_rcv_skb(skb, &crypto_user_rcv_msg);
	mutex_unlock(&crypto_cfg_mutex);
}

static int __net_init crypto_netlink_init(struct net *net)
{
	struct netlink_kernel_cfg cfg = {
		.input	= crypto_netlink_rcv,
	};

	net->crypto_nlsk = netlink_kernel_create(net, NETLINK_CRYPTO, &cfg);
	return net->crypto_nlsk == NULL ? -ENOMEM : 0;
}

static void __net_exit crypto_netlink_exit(struct net *net)
{
	netlink_kernel_release(net->crypto_nlsk);
	net->crypto_nlsk = NULL;
}

static struct pernet_operations crypto_netlink_net_ops = {
	.init = crypto_netlink_init,
	.exit = crypto_netlink_exit,
};

static int __init crypto_user_init(void)
{
	return register_pernet_subsys(&crypto_netlink_net_ops);
}

static void __exit crypto_user_exit(void)
{
	unregister_pernet_subsys(&crypto_netlink_net_ops);
}

module_init(crypto_user_init);
module_exit(crypto_user_exit);
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Steffen Klassert <steffen.klassert@secunet.com>");
MODULE_DESCRIPTION("Crypto userspace configuration API");
MODULE_ALIAS("net-pf-16-proto-21");
back to top