* [RFC] xfrm: x86_64 CONFIG_COMPAT support
From: Florian Westphal @ 2010-02-05 1:47 UTC (permalink / raw)
To: netdev
At the moment, it is not possible to use the xfrm netlink interface on
x86_64 with a 32bit userland.
The problem is due to a few structures, notably struct xfrm_usersa_info,
which have different sizes in user/kernelspace (3 byte padding on x86, 7
byte on x86_64). Thus, when the kernel skips the family specific header,
it skips 4 additional bytes; netlink verification thus fails.
What follows is a (incomplete, untested) draft patch that tries
to solve this by introducing a delta table. With the changes in place
so far its possible to run
"ip xfrm state add ..." with a 32bit ip binary.
Unfortunately, the MSG_COMPAT flag seems to get lost along
the way, so this uses x86_64 is_compat_task() for now.
The delta table is not sufficient, as struct xfrm_usersa_info
can be embedded inside other structs, so when accessing data
after xfrm_usersa_info, one has to use an appropriate compat
structure definition.
Before I spend any more time on this, my questions are:
- does this approach look somewhat sane to you?
- do you have any suggestions how xfrm CONFIG_COMPAT support should
look like?
- are you aware of pitfalls other than xfrm structure padding?
- are there platforms other than x86_64 that have this problem?
Thanks in advance, Florian
diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c
index d5a7129..e8747fb 100644
--- a/net/xfrm/xfrm_user.c
+++ b/net/xfrm/xfrm_user.c
@@ -11,6 +11,7 @@
*/
#include <linux/crypto.h>
+#include <linux/compat.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/types.h>
@@ -31,6 +32,35 @@
#include <linux/in6.h>
#endif
+#if defined(CONFIG_X86_64) && defined(CONFIG_COMPAT)
+#define XFRM_COMPAT_TEST is_compat_task()
+/* userspace size decreases by this amount due to padding */
+static const u8 xfrm_msg_min_compat_pad[XFRM_NR_MSGTYPES] = {
+ [XFRM_MSG_NEWSA - XFRM_MSG_BASE] = 4,
+ [XFRM_MSG_NEWPOLICY - XFRM_MSG_BASE] = 4,
+
+ [XFRM_MSG_ALLOCSPI - XFRM_MSG_BASE] = 4, /* need converter */
+ [XFRM_MSG_ACQUIRE - XFRM_MSG_BASE] = 4, /* need converter */
+ [XFRM_MSG_EXPIRE - XFRM_MSG_BASE] = 4, /* need converter */
+ [XFRM_MSG_UPDPOLICY - XFRM_MSG_BASE] = 4,
+ [XFRM_MSG_UPDSA - XFRM_MSG_BASE] = 4,
+ [XFRM_MSG_POLEXPIRE - XFRM_MSG_BASE] = 4, /* need converter */
+
+ [XFRM_MSG_NEWAE - XFRM_MSG_BASE] = 4,
+ [XFRM_MSG_GETAE - XFRM_MSG_BASE] = 4,
+};
+
+#else
+#define XFRM_COMPAT_TEST 0
+#endif /* defined(CONFIG_X86_64) && defined(CONFIG_COMPAT) */
+
+struct xfrm_compat_userspi_info {
+ __u32 xfrm_usersa_info[55];
+ __u32 min, max;
+};
+
+static int xfrm_msg_hdrlen(int type);
+
static inline int aead_len(struct xfrm_algo_aead *alg)
{
return sizeof(*alg) + ((alg->alg_key_len + 7) / 8);
@@ -700,9 +730,10 @@ static int dump_one_state(struct xfrm_state *x, int count, void *ptr)
struct xfrm_usersa_info *p;
struct nlmsghdr *nlh;
int err;
+ int type = XFRM_MSG_NEWSA - XFRM_MSG_BASE;
nlh = nlmsg_put(skb, NETLINK_CB(in_skb).pid, sp->nlmsg_seq,
- XFRM_MSG_NEWSA, sizeof(*p), sp->nlmsg_flags);
+ XFRM_MSG_NEWSA, xfrm_msg_hdrlen(type), sp->nlmsg_flags);
if (nlh == NULL)
return -EMSGSIZE;
@@ -916,6 +947,30 @@ out_noput:
static int verify_userspi_info(struct xfrm_userspi_info *p)
{
+ if (XFRM_COMPAT_TEST) {
+ struct xfrm_compat_userspi_info *compat = (struct xfrm_compat_userspi_info *) p;
+
+ switch (p->info.id.proto) {
+ case IPPROTO_AH:
+ case IPPROTO_ESP:
+ break;
+
+ case IPPROTO_COMP:
+ /* IPCOMP spi is 16-bits. */
+ if (compat->max >= 0x10000)
+ return -EINVAL;
+ break;
+
+ default:
+ return -EINVAL;
+ }
+
+ if (compat->min > compat->max)
+ return -EINVAL;
+
+ return 0;
+ }
+
switch (p->info.id.proto) {
case IPPROTO_AH:
case IPPROTO_ESP:
@@ -974,7 +1029,12 @@ static int xfrm_alloc_userspi(struct sk_buff *skb, struct nlmsghdr *nlh,
if (x == NULL)
goto out_noput;
- err = xfrm_alloc_spi(x, p->min, p->max);
+ if (XFRM_COMPAT_TEST) {
+ struct xfrm_compat_userspi_info *compat = (struct xfrm_compat_userspi_info *) p;
+ err = xfrm_alloc_spi(x, compat->min, compat->max);
+ } else {
+ err = xfrm_alloc_spi(x, p->min, p->max);
+ }
if (err)
goto out;
@@ -2102,6 +2162,13 @@ static struct xfrm_link {
[XFRM_MSG_GETSPDINFO - XFRM_MSG_BASE] = { .doit = xfrm_get_spdinfo },
};
+static int xfrm_msg_hdrlen(int type)
+{
+ if (unlikely(XFRM_COMPAT_TEST))
+ return xfrm_msg_min[type] - xfrm_msg_min_compat_pad[type];
+ return xfrm_msg_min[type];
+}
+
static int xfrm_user_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
{
struct net *net = sock_net(skb->sk);
@@ -2129,7 +2196,7 @@ static int xfrm_user_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
return netlink_dump_start(net->xfrm.nlsk, skb, nlh, link->dump, link->done);
}
- err = nlmsg_parse(nlh, xfrm_msg_min[type], attrs, XFRMA_MAX,
+ err = nlmsg_parse(nlh, xfrm_msg_hdrlen(type), attrs, XFRMA_MAX,
xfrma_policy);
if (err < 0)
return err;
^ permalink raw reply related
* [PATCH 13/13] net: atm: use seq_list_foo() helpers
From: Li Zefan @ 2010-02-05 1:47 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/atm/proc.c | 2 +-
net/atm/resources.c | 18 ++----------------
2 files changed, 3 insertions(+), 17 deletions(-)
diff --git a/net/atm/proc.c b/net/atm/proc.c
index ab8419a..c0c3a79 100644
--- a/net/atm/proc.c
+++ b/net/atm/proc.c
@@ -236,7 +236,7 @@ static int atm_dev_seq_show(struct seq_file *seq, void *v)
"Itf Type ESI/\"MAC\"addr "
"AAL(TX,err,RX,err,drop) ... [refcnt]\n";
- if (v == SEQ_START_TOKEN)
+ if (v == &atm_devs)
seq_puts(seq, atm_dev_banner);
else {
struct atm_dev *dev = list_entry(v, struct atm_dev, dev_list);
diff --git a/net/atm/resources.c b/net/atm/resources.c
index f4091d6..bbd01b0 100644
--- a/net/atm/resources.c
+++ b/net/atm/resources.c
@@ -454,21 +454,10 @@ done:
return error;
}
-static inline void *dev_get_idx(loff_t left)
-{
- struct list_head *p;
-
- list_for_each(p, &atm_devs) {
- if (!--left)
- break;
- }
- return (p != &atm_devs) ? p : NULL;
-}
-
void *atm_dev_seq_start(struct seq_file *seq, loff_t *pos)
{
mutex_lock(&atm_dev_mutex);
- return *pos ? dev_get_idx(*pos) : SEQ_START_TOKEN;
+ return seq_list_start_head(&atm_devs, *pos);
}
void atm_dev_seq_stop(struct seq_file *seq, void *v)
@@ -478,8 +467,5 @@ void atm_dev_seq_stop(struct seq_file *seq, void *v)
void *atm_dev_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
- ++*pos;
- v = (v == SEQ_START_TOKEN)
- ? atm_devs.next : ((struct list_head *)v)->next;
- return (v == &atm_devs) ? NULL : v;
+ return seq_list_next(v, &atm_devs, pos);
}
--
1.6.3
^ permalink raw reply related
* [PATCH 12/13] net: ipx: use seq_list_foo() helpers
From: Li Zefan @ 2010-02-05 1:47 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/ipx/ipx_proc.c | 89 +++++++---------------------------------------------
1 files changed, 12 insertions(+), 77 deletions(-)
diff --git a/net/ipx/ipx_proc.c b/net/ipx/ipx_proc.c
index 5761784..cd44358 100644
--- a/net/ipx/ipx_proc.c
+++ b/net/ipx/ipx_proc.c
@@ -13,45 +13,15 @@
#include <net/tcp_states.h>
#include <net/ipx.h>
-static __inline__ struct ipx_interface *ipx_get_interface_idx(loff_t pos)
-{
- struct ipx_interface *i;
-
- list_for_each_entry(i, &ipx_interfaces, node)
- if (!pos--)
- goto out;
- i = NULL;
-out:
- return i;
-}
-
-static struct ipx_interface *ipx_interfaces_next(struct ipx_interface *i)
-{
- struct ipx_interface *rc = NULL;
-
- if (i->node.next != &ipx_interfaces)
- rc = list_entry(i->node.next, struct ipx_interface, node);
- return rc;
-}
-
static void *ipx_seq_interface_start(struct seq_file *seq, loff_t *pos)
{
- loff_t l = *pos;
-
spin_lock_bh(&ipx_interfaces_lock);
- return l ? ipx_get_interface_idx(--l) : SEQ_START_TOKEN;
+ return seq_list_start_head(&ipx_interfaces, *pos);
}
static void *ipx_seq_interface_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct ipx_interface *i;
-
- ++*pos;
- if (v == SEQ_START_TOKEN)
- i = ipx_interfaces_head();
- else
- i = ipx_interfaces_next(v);
- return i;
+ return seq_list_next(v, &ipx_interfaces, pos);
}
static void ipx_seq_interface_stop(struct seq_file *seq, void *v)
@@ -63,7 +33,7 @@ static int ipx_seq_interface_show(struct seq_file *seq, void *v)
{
struct ipx_interface *i;
- if (v == SEQ_START_TOKEN) {
+ if (v == &ipx_interfaces) {
seq_puts(seq, "Network Node_Address Primary Device "
"Frame_Type");
#ifdef IPX_REFCNT_DEBUG
@@ -73,7 +43,7 @@ static int ipx_seq_interface_show(struct seq_file *seq, void *v)
goto out;
}
- i = v;
+ i = list_entry(v, struct ipx_interface, node);
seq_printf(seq, "%08lX ", (unsigned long int)ntohl(i->if_netnum));
seq_printf(seq, "%02X%02X%02X%02X%02X%02X ",
i->if_node[0], i->if_node[1], i->if_node[2],
@@ -89,53 +59,16 @@ out:
return 0;
}
-static struct ipx_route *ipx_routes_head(void)
-{
- struct ipx_route *rc = NULL;
-
- if (!list_empty(&ipx_routes))
- rc = list_entry(ipx_routes.next, struct ipx_route, node);
- return rc;
-}
-
-static struct ipx_route *ipx_routes_next(struct ipx_route *r)
-{
- struct ipx_route *rc = NULL;
-
- if (r->node.next != &ipx_routes)
- rc = list_entry(r->node.next, struct ipx_route, node);
- return rc;
-}
-
-static __inline__ struct ipx_route *ipx_get_route_idx(loff_t pos)
-{
- struct ipx_route *r;
-
- list_for_each_entry(r, &ipx_routes, node)
- if (!pos--)
- goto out;
- r = NULL;
-out:
- return r;
-}
-
static void *ipx_seq_route_start(struct seq_file *seq, loff_t *pos)
{
- loff_t l = *pos;
read_lock_bh(&ipx_routes_lock);
- return l ? ipx_get_route_idx(--l) : SEQ_START_TOKEN;
+ return seq_list_start_head(&ipx_routes, *pos);
}
static void *ipx_seq_route_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct ipx_route *r;
-
++*pos;
- if (v == SEQ_START_TOKEN)
- r = ipx_routes_head();
- else
- r = ipx_routes_next(v);
- return r;
+ return seq_list_next(v, &ipx_routes, pos);
}
static void ipx_seq_route_stop(struct seq_file *seq, void *v)
@@ -147,11 +80,13 @@ static int ipx_seq_route_show(struct seq_file *seq, void *v)
{
struct ipx_route *rt;
- if (v == SEQ_START_TOKEN) {
+ if (v == &ipx_routes) {
seq_puts(seq, "Network Router_Net Router_Node\n");
goto out;
}
- rt = v;
+
+ rt = list_entry(v, struct ipx_route, node);
+
seq_printf(seq, "%08lX ", (unsigned long int)ntohl(rt->ir_net));
if (rt->ir_routed)
seq_printf(seq, "%08lX %02X%02X%02X%02X%02X%02X\n",
@@ -226,9 +161,9 @@ static void *ipx_seq_socket_next(struct seq_file *seq, void *v, loff_t *pos)
spin_unlock_bh(&i->if_sklist_lock);
sk = NULL;
for (;;) {
- i = ipx_interfaces_next(i);
- if (!i)
+ if (i->node.next == &ipx_interfaces)
break;
+ i = list_entry(i->node.next, struct ipx_interface, node);
spin_lock_bh(&i->if_sklist_lock);
if (!hlist_empty(&i->if_sklist)) {
sk = sk_head(&i->if_sklist);
--
1.6.3
^ permalink raw reply related
* [PATCH 11/13] net: irda: use seq_list_foo() helpers
From: Li Zefan @ 2010-02-05 1:47 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/irda/irlan/irlan_common.c | 26 ++++----------------------
1 files changed, 4 insertions(+), 22 deletions(-)
diff --git a/net/irda/irlan/irlan_common.c b/net/irda/irlan/irlan_common.c
index 315ead3..2141593 100644
--- a/net/irda/irlan/irlan_common.c
+++ b/net/irda/irlan/irlan_common.c
@@ -1128,19 +1128,8 @@ int irlan_extract_param(__u8 *buf, char *name, char *value, __u16 *len)
*/
static void *irlan_seq_start(struct seq_file *seq, loff_t *pos)
{
- int i = 1;
- struct irlan_cb *self;
-
rcu_read_lock();
- if (*pos == 0)
- return SEQ_START_TOKEN;
-
- list_for_each_entry(self, &irlans, dev_list) {
- if (*pos == i)
- return self;
- ++i;
- }
- return NULL;
+ return seq_list_start_head(&irlans, *pos);
}
/* Return entry after v, and increment pos */
@@ -1148,14 +1137,7 @@ static void *irlan_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
struct list_head *nxt;
- ++*pos;
- if (v == SEQ_START_TOKEN)
- nxt = irlans.next;
- else
- nxt = ((struct irlan_cb *)v)->dev_list.next;
-
- return (nxt == &irlans) ? NULL
- : list_entry(nxt, struct irlan_cb, dev_list);
+ return seq_list_next(v, &irlans, pos);
}
/* End of reading /proc file */
@@ -1170,10 +1152,10 @@ static void irlan_seq_stop(struct seq_file *seq, void *v)
*/
static int irlan_seq_show(struct seq_file *seq, void *v)
{
- if (v == SEQ_START_TOKEN)
+ if (v == &irlans)
seq_puts(seq, "IrLAN instances:\n");
else {
- struct irlan_cb *self = v;
+ struct irlan_cb *self = list_entry(v, struct irlan_cb, list);
IRDA_ASSERT(self != NULL, return -1;);
IRDA_ASSERT(self->magic == IRLAN_MAGIC, return -1;);
--
1.6.3
^ permalink raw reply related
* [PATCH 09/13] net: af_key: use seq_hlist_foo() helpers
From: Li Zefan @ 2010-02-05 1:46 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/key/af_key.c | 20 +++-----------------
1 files changed, 3 insertions(+), 17 deletions(-)
diff --git a/net/key/af_key.c b/net/key/af_key.c
index 76fa6fe..cb006b6 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -3655,9 +3655,8 @@ static const struct net_proto_family pfkey_family_ops = {
#ifdef CONFIG_PROC_FS
static int pfkey_seq_show(struct seq_file *f, void *v)
{
- struct sock *s;
+ struct sock *s = sk_entry(v);
- s = (struct sock *)v;
if (v == SEQ_START_TOKEN)
seq_printf(f ,"sk RefCnt Rmem Wmem User Inode\n");
else
@@ -3676,19 +3675,9 @@ static void *pfkey_seq_start(struct seq_file *f, loff_t *ppos)
{
struct net *net = seq_file_net(f);
struct netns_pfkey *net_pfkey = net_generic(net, pfkey_net_id);
- struct sock *s;
- struct hlist_node *node;
- loff_t pos = *ppos;
read_lock(&pfkey_table_lock);
- if (pos == 0)
- return SEQ_START_TOKEN;
-
- sk_for_each(s, node, &net_pfkey->table)
- if (pos-- == 1)
- return s;
-
- return NULL;
+ return seq_hlist_start_head(&net_pfkey->table, *ppos);
}
static void *pfkey_seq_next(struct seq_file *f, void *v, loff_t *ppos)
@@ -3696,10 +3685,7 @@ static void *pfkey_seq_next(struct seq_file *f, void *v, loff_t *ppos)
struct net *net = seq_file_net(f);
struct netns_pfkey *net_pfkey = net_generic(net, pfkey_net_id);
- ++*ppos;
- return (v == SEQ_START_TOKEN) ?
- sk_head(&net_pfkey->table) :
- sk_next((struct sock *)v);
+ return seq_hlist_next(v, &net_pfkey->table, ppos);
}
static void pfkey_seq_stop(struct seq_file *f, void *v)
--
1.6.3
^ permalink raw reply related
* [PATCH 10/13] net: x25: use seq_list_foo() helpers
From: Li Zefan @ 2010-02-05 1:47 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/x25/x25_proc.c | 84 +++++-----------------------------------------------
1 files changed, 8 insertions(+), 76 deletions(-)
diff --git a/net/x25/x25_proc.c b/net/x25/x25_proc.c
index c4cd322..7ff3737 100644
--- a/net/x25/x25_proc.c
+++ b/net/x25/x25_proc.c
@@ -25,49 +25,17 @@
#include <net/x25.h>
#ifdef CONFIG_PROC_FS
-static __inline__ struct x25_route *x25_get_route_idx(loff_t pos)
-{
- struct list_head *route_entry;
- struct x25_route *rt = NULL;
-
- list_for_each(route_entry, &x25_route_list) {
- rt = list_entry(route_entry, struct x25_route, node);
- if (!pos--)
- goto found;
- }
- rt = NULL;
-found:
- return rt;
-}
static void *x25_seq_route_start(struct seq_file *seq, loff_t *pos)
__acquires(x25_route_list_lock)
{
- loff_t l = *pos;
-
read_lock_bh(&x25_route_list_lock);
- return l ? x25_get_route_idx(--l) : SEQ_START_TOKEN;
+ return seq_list_start_head(&x25_route_list, *pos);
}
static void *x25_seq_route_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct x25_route *rt;
-
- ++*pos;
- if (v == SEQ_START_TOKEN) {
- rt = NULL;
- if (!list_empty(&x25_route_list))
- rt = list_entry(x25_route_list.next,
- struct x25_route, node);
- goto out;
- }
- rt = v;
- if (rt->node.next != &x25_route_list)
- rt = list_entry(rt->node.next, struct x25_route, node);
- else
- rt = NULL;
-out:
- return rt;
+ return seq_list_next(v, &x25_route_list, pos);
}
static void x25_seq_route_stop(struct seq_file *seq, void *v)
@@ -78,9 +46,9 @@ static void x25_seq_route_stop(struct seq_file *seq, void *v)
static int x25_seq_route_show(struct seq_file *seq, void *v)
{
- struct x25_route *rt;
+ struct x25_route *rt = list_entry(v, struct x25_route, node);
- if (v == SEQ_START_TOKEN) {
+ if (v == &x25_route_list) {
seq_puts(seq, "Address Digits Device\n");
goto out;
}
@@ -146,51 +114,16 @@ out:
return 0;
}
-static __inline__ struct x25_forward *x25_get_forward_idx(loff_t pos)
-{
- struct x25_forward *f;
- struct list_head *entry;
-
- list_for_each(entry, &x25_forward_list) {
- f = list_entry(entry, struct x25_forward, node);
- if (!pos--)
- goto found;
- }
-
- f = NULL;
-found:
- return f;
-}
-
static void *x25_seq_forward_start(struct seq_file *seq, loff_t *pos)
__acquires(x25_forward_list_lock)
{
- loff_t l = *pos;
-
read_lock_bh(&x25_forward_list_lock);
- return l ? x25_get_forward_idx(--l) : SEQ_START_TOKEN;
+ return seq_list_start_head(&x25_forward_list, *pos);
}
static void *x25_seq_forward_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct x25_forward *f;
-
- ++*pos;
- if (v == SEQ_START_TOKEN) {
- f = NULL;
- if (!list_empty(&x25_forward_list))
- f = list_entry(x25_forward_list.next,
- struct x25_forward, node);
- goto out;
- }
- f = v;
- if (f->node.next != &x25_forward_list)
- f = list_entry(f->node.next, struct x25_forward, node);
- else
- f = NULL;
-out:
- return f;
-
+ return seq_list_next(v, &x25_forward_list, pos);
}
static void x25_seq_forward_stop(struct seq_file *seq, void *v)
@@ -201,9 +134,9 @@ static void x25_seq_forward_stop(struct seq_file *seq, void *v)
static int x25_seq_forward_show(struct seq_file *seq, void *v)
{
- struct x25_forward *f;
+ struct x25_forward *f = list_entry(v, struct x25_forward, node);
- if (v == SEQ_START_TOKEN) {
+ if (v == &x25_forward_list) {
seq_printf(seq, "lci dev1 dev2\n");
goto out;
}
@@ -212,7 +145,6 @@ static int x25_seq_forward_show(struct seq_file *seq, void *v)
seq_printf(seq, "%d %-10s %-10s\n",
f->lci, f->dev1->name, f->dev2->name);
-
out:
return 0;
}
--
1.6.3
^ permalink raw reply related
* [PATCH 08/13] net: appletalk: use seq_hlist_foo() helpers
From: Li Zefan @ 2010-02-05 1:46 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/appletalk/atalk_proc.c | 30 +++---------------------------
1 files changed, 3 insertions(+), 27 deletions(-)
diff --git a/net/appletalk/atalk_proc.c b/net/appletalk/atalk_proc.c
index 80caad1..6ef0e76 100644
--- a/net/appletalk/atalk_proc.c
+++ b/net/appletalk/atalk_proc.c
@@ -144,40 +144,16 @@ out:
return 0;
}
-static __inline__ struct sock *atalk_get_socket_idx(loff_t pos)
-{
- struct sock *s;
- struct hlist_node *node;
-
- sk_for_each(s, node, &atalk_sockets)
- if (!pos--)
- goto found;
- s = NULL;
-found:
- return s;
-}
-
static void *atalk_seq_socket_start(struct seq_file *seq, loff_t *pos)
__acquires(atalk_sockets_lock)
{
- loff_t l = *pos;
-
read_lock_bh(&atalk_sockets_lock);
- return l ? atalk_get_socket_idx(--l) : SEQ_START_TOKEN;
+ return seq_hlist_start_head(&atalk_sockets, *pos);
}
static void *atalk_seq_socket_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct sock *i;
-
- ++*pos;
- if (v == SEQ_START_TOKEN) {
- i = sk_head(&atalk_sockets);
- goto out;
- }
- i = sk_next(v);
-out:
- return i;
+ return seq_hlist_next(v, &atalk_sockets, pos);
}
static void atalk_seq_socket_stop(struct seq_file *seq, void *v)
@@ -197,7 +173,7 @@ static int atalk_seq_socket_show(struct seq_file *seq, void *v)
goto out;
}
- s = v;
+ s = sk_entry(v);
at = at_sk(s);
seq_printf(seq, "%02X %04X:%02X:%02X %04X:%02X:%02X %08X:%08X "
--
1.6.3
^ permalink raw reply related
* [PATCH 07/13] net: ax25: use seq_hlist_foo() helpers
From: Li Zefan @ 2010-02-05 1:46 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/ax25/af_ax25.c | 18 +++---------------
net/ax25/ax25_uid.c | 25 ++++---------------------
2 files changed, 7 insertions(+), 36 deletions(-)
diff --git a/net/ax25/af_ax25.c b/net/ax25/af_ax25.c
index 5588ba6..a5beedf 100644
--- a/net/ax25/af_ax25.c
+++ b/net/ax25/af_ax25.c
@@ -1863,25 +1863,13 @@ static int ax25_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
static void *ax25_info_start(struct seq_file *seq, loff_t *pos)
__acquires(ax25_list_lock)
{
- struct ax25_cb *ax25;
- struct hlist_node *node;
- int i = 0;
-
spin_lock_bh(&ax25_list_lock);
- ax25_for_each(ax25, node, &ax25_list) {
- if (i == *pos)
- return ax25;
- ++i;
- }
- return NULL;
+ return seq_hlist_start(&ax25_list, *pos);
}
static void *ax25_info_next(struct seq_file *seq, void *v, loff_t *pos)
{
- ++*pos;
-
- return hlist_entry( ((struct ax25_cb *)v)->ax25_node.next,
- struct ax25_cb, ax25_node);
+ return seq_hlist_next(v, &ax25_list, pos);
}
static void ax25_info_stop(struct seq_file *seq, void *v)
@@ -1892,7 +1880,7 @@ static void ax25_info_stop(struct seq_file *seq, void *v)
static int ax25_info_show(struct seq_file *seq, void *v)
{
- ax25_cb *ax25 = v;
+ ax25_cb *ax25 = hlist_entry(v, struct ax25_cb, ax25_node);
char buf[11];
int k;
diff --git a/net/ax25/ax25_uid.c b/net/ax25/ax25_uid.c
index 832bcf0..9f13f6e 100644
--- a/net/ax25/ax25_uid.c
+++ b/net/ax25/ax25_uid.c
@@ -146,31 +146,13 @@ int ax25_uid_ioctl(int cmd, struct sockaddr_ax25 *sax)
static void *ax25_uid_seq_start(struct seq_file *seq, loff_t *pos)
__acquires(ax25_uid_lock)
{
- struct ax25_uid_assoc *pt;
- struct hlist_node *node;
- int i = 1;
-
read_lock(&ax25_uid_lock);
-
- if (*pos == 0)
- return SEQ_START_TOKEN;
-
- ax25_uid_for_each(pt, node, &ax25_uid_list) {
- if (i == *pos)
- return pt;
- ++i;
- }
- return NULL;
+ return seq_hlist_start_head(&ax25_uid_list, *pos);
}
static void *ax25_uid_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
- ++*pos;
- if (v == SEQ_START_TOKEN)
- return ax25_uid_list.first;
- else
- return hlist_entry(((ax25_uid_assoc *)v)->uid_node.next,
- ax25_uid_assoc, uid_node);
+ return seq_hlist_next(v, &ax25_uid_list, pos);
}
static void ax25_uid_seq_stop(struct seq_file *seq, void *v)
@@ -186,8 +168,9 @@ static int ax25_uid_seq_show(struct seq_file *seq, void *v)
if (v == SEQ_START_TOKEN)
seq_printf(seq, "Policy: %d\n", ax25_uid_policy);
else {
- struct ax25_uid_assoc *pt = v;
+ struct ax25_uid_assoc *pt;
+ pt = hlist_entry(v, struct ax25_uid_assoc, uid_node);
seq_printf(seq, "%6d %s\n", pt->uid, ax2asc(buf, &pt->call));
}
return 0;
--
1.6.3
^ permalink raw reply related
* [PATCH 06/13] net: netrom: use seq_hlist_foo() helpers
From: Li Zefan @ 2010-02-05 1:46 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/netrom/af_netrom.c | 21 ++----------------
net/netrom/nr_route.c | 53 ++++++++---------------------------------------
2 files changed, 12 insertions(+), 62 deletions(-)
diff --git a/net/netrom/af_netrom.c b/net/netrom/af_netrom.c
index 71604c6..a249127 100644
--- a/net/netrom/af_netrom.c
+++ b/net/netrom/af_netrom.c
@@ -1267,28 +1267,13 @@ static int nr_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
static void *nr_info_start(struct seq_file *seq, loff_t *pos)
{
- struct sock *s;
- struct hlist_node *node;
- int i = 1;
-
spin_lock_bh(&nr_list_lock);
- if (*pos == 0)
- return SEQ_START_TOKEN;
-
- sk_for_each(s, node, &nr_list) {
- if (i == *pos)
- return s;
- ++i;
- }
- return NULL;
+ return seq_hlist_start_head(&nr_list, *pos);
}
static void *nr_info_next(struct seq_file *seq, void *v, loff_t *pos)
{
- ++*pos;
-
- return (v == SEQ_START_TOKEN) ? sk_head(&nr_list)
- : sk_next((struct sock *)v);
+ return seq_hlist_next(v, &nr_list, pos);
}
static void nr_info_stop(struct seq_file *seq, void *v)
@@ -1298,7 +1283,7 @@ static void nr_info_stop(struct seq_file *seq, void *v)
static int nr_info_show(struct seq_file *seq, void *v)
{
- struct sock *s = v;
+ struct sock *s = sk_entry(v);
struct net_device *dev;
struct nr_sock *nr;
const char *devname;
diff --git a/net/netrom/nr_route.c b/net/netrom/nr_route.c
index e2e2d33..5cc6480 100644
--- a/net/netrom/nr_route.c
+++ b/net/netrom/nr_route.c
@@ -863,33 +863,13 @@ int nr_route_frame(struct sk_buff *skb, ax25_cb *ax25)
static void *nr_node_start(struct seq_file *seq, loff_t *pos)
{
- struct nr_node *nr_node;
- struct hlist_node *node;
- int i = 1;
-
spin_lock_bh(&nr_node_list_lock);
- if (*pos == 0)
- return SEQ_START_TOKEN;
-
- nr_node_for_each(nr_node, node, &nr_node_list) {
- if (i == *pos)
- return nr_node;
- ++i;
- }
-
- return NULL;
+ return seq_hlist_start_head(&nr_node_list, *pos);
}
static void *nr_node_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct hlist_node *node;
- ++*pos;
-
- node = (v == SEQ_START_TOKEN)
- ? nr_node_list.first
- : ((struct nr_node *)v)->node_node.next;
-
- return hlist_entry(node, struct nr_node, node_node);
+ return seq_hlist_next(v, &nr_node_list, pos);
}
static void nr_node_stop(struct seq_file *seq, void *v)
@@ -906,7 +886,9 @@ static int nr_node_show(struct seq_file *seq, void *v)
seq_puts(seq,
"callsign mnemonic w n qual obs neigh qual obs neigh qual obs neigh\n");
else {
- struct nr_node *nr_node = v;
+ struct nr_node *nr_node = hlist_entry(v, struct nr_node,
+ node_node);
+
nr_node_lock(nr_node);
seq_printf(seq, "%-9s %-7s %d %d",
ax2asc(buf, &nr_node->callsign),
@@ -949,31 +931,13 @@ const struct file_operations nr_nodes_fops = {
static void *nr_neigh_start(struct seq_file *seq, loff_t *pos)
{
- struct nr_neigh *nr_neigh;
- struct hlist_node *node;
- int i = 1;
-
spin_lock_bh(&nr_neigh_list_lock);
- if (*pos == 0)
- return SEQ_START_TOKEN;
-
- nr_neigh_for_each(nr_neigh, node, &nr_neigh_list) {
- if (i == *pos)
- return nr_neigh;
- }
- return NULL;
+ return seq_hlist_start_head(&nr_neigh_list, *pos);
}
static void *nr_neigh_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct hlist_node *node;
- ++*pos;
-
- node = (v == SEQ_START_TOKEN)
- ? nr_neigh_list.first
- : ((struct nr_neigh *)v)->neigh_node.next;
-
- return hlist_entry(node, struct nr_neigh, neigh_node);
+ return seq_hlist_next(v, &nr_neigh_list, pos);
}
static void nr_neigh_stop(struct seq_file *seq, void *v)
@@ -989,8 +953,9 @@ static int nr_neigh_show(struct seq_file *seq, void *v)
if (v == SEQ_START_TOKEN)
seq_puts(seq, "addr callsign dev qual lock count failed digipeaters\n");
else {
- struct nr_neigh *nr_neigh = v;
+ struct nr_neigh *nr_neigh;
+ nr_neigh = hlist_entry(v, struct nr_neigh, neigh_node);
seq_printf(seq, "%05d %-9s %-4s %3d %d %3d %3d",
nr_neigh->number,
ax2asc(buf, &nr_neigh->callsign),
--
1.6.3
^ permalink raw reply related
* [PATCH 05/13] net: packet: use seq_hlist_foo() helpers
From: Li Zefan @ 2010-02-05 1:45 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/packet/af_packet.c | 20 +++-----------------
1 files changed, 3 insertions(+), 17 deletions(-)
diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index 0f6bf59..256de26 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -2374,33 +2374,19 @@ static struct notifier_block packet_netdev_notifier = {
};
#ifdef CONFIG_PROC_FS
-static inline struct sock *packet_seq_idx(struct net *net, loff_t off)
-{
- struct sock *s;
- struct hlist_node *node;
-
- sk_for_each(s, node, &net->packet.sklist) {
- if (!off--)
- return s;
- }
- return NULL;
-}
static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
__acquires(seq_file_net(seq)->packet.sklist_lock)
{
struct net *net = seq_file_net(seq);
read_lock(&net->packet.sklist_lock);
- return *pos ? packet_seq_idx(net, *pos - 1) : SEQ_START_TOKEN;
+ return seq_hlist_start_head(&net->packet.sklist, *pos);
}
static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
struct net *net = seq_file_net(seq);
- ++*pos;
- return (v == SEQ_START_TOKEN)
- ? sk_head(&net->packet.sklist)
- : sk_next((struct sock *)v) ;
+ return seq_hlist_next(v, &net->packet.sklist, pos);
}
static void packet_seq_stop(struct seq_file *seq, void *v)
@@ -2415,7 +2401,7 @@ static int packet_seq_show(struct seq_file *seq, void *v)
if (v == SEQ_START_TOKEN)
seq_puts(seq, "sk RefCnt Type Proto Iface R Rmem User Inode\n");
else {
- struct sock *s = v;
+ struct sock *s = sk_entry(v);
const struct packet_sock *po = pkt_sk(s);
seq_printf(seq,
--
1.6.3
^ permalink raw reply related
* [PATCH 04/13] net: rose: use seq_hlist_foo() helpers
From: Li Zefan @ 2010-02-05 1:45 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
net/rose/af_rose.c | 22 +++-------------------
1 files changed, 3 insertions(+), 19 deletions(-)
diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c
index 8feb9e5..e90b9b6 100644
--- a/net/rose/af_rose.c
+++ b/net/rose/af_rose.c
@@ -1404,29 +1404,13 @@ static int rose_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
static void *rose_info_start(struct seq_file *seq, loff_t *pos)
__acquires(rose_list_lock)
{
- int i;
- struct sock *s;
- struct hlist_node *node;
-
spin_lock_bh(&rose_list_lock);
- if (*pos == 0)
- return SEQ_START_TOKEN;
-
- i = 1;
- sk_for_each(s, node, &rose_list) {
- if (i == *pos)
- return s;
- ++i;
- }
- return NULL;
+ return seq_hlist_start_head(&rose_list, *pos);
}
static void *rose_info_next(struct seq_file *seq, void *v, loff_t *pos)
{
- ++*pos;
-
- return (v == SEQ_START_TOKEN) ? sk_head(&rose_list)
- : sk_next((struct sock *)v);
+ return seq_hlist_next(v, &rose_list, pos);
}
static void rose_info_stop(struct seq_file *seq, void *v)
@@ -1444,7 +1428,7 @@ static int rose_info_show(struct seq_file *seq, void *v)
"dest_addr dest_call src_addr src_call dev lci neigh st vs vr va t t1 t2 t3 hb idle Snd-Q Rcv-Q inode\n");
else {
- struct sock *s = v;
+ struct sock *s = sk_entry(v);
struct rose_sock *rose = rose_sk(s);
const char *devname, *callsign;
const struct net_device *dev = rose->device;
--
1.6.3
^ permalink raw reply related
* [PATCH 03/13] net: x25: use seq_hlist_foo() helpers
From: Li Zefan @ 2010-02-05 1:45 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Simplify seq_file code.
Signed-off-by: Li Zefan <lizf@cn.fujtisu.com>
---
net/x25/x25_proc.c | 30 +++---------------------------
1 files changed, 3 insertions(+), 27 deletions(-)
diff --git a/net/x25/x25_proc.c b/net/x25/x25_proc.c
index 0a04e62..c4cd322 100644
--- a/net/x25/x25_proc.c
+++ b/net/x25/x25_proc.c
@@ -93,40 +93,16 @@ out:
return 0;
}
-static __inline__ struct sock *x25_get_socket_idx(loff_t pos)
-{
- struct sock *s;
- struct hlist_node *node;
-
- sk_for_each(s, node, &x25_list)
- if (!pos--)
- goto found;
- s = NULL;
-found:
- return s;
-}
-
static void *x25_seq_socket_start(struct seq_file *seq, loff_t *pos)
__acquires(x25_list_lock)
{
- loff_t l = *pos;
-
read_lock_bh(&x25_list_lock);
- return l ? x25_get_socket_idx(--l) : SEQ_START_TOKEN;
+ return seq_hlist_start_head(&x25_list, *pos);
}
static void *x25_seq_socket_next(struct seq_file *seq, void *v, loff_t *pos)
{
- struct sock *s;
-
- ++*pos;
- if (v == SEQ_START_TOKEN) {
- s = sk_head(&x25_list);
- goto out;
- }
- s = sk_next(v);
-out:
- return s;
+ return seq_hlist_next(v, &x25_list, pos);
}
static void x25_seq_socket_stop(struct seq_file *seq, void *v)
@@ -148,7 +124,7 @@ static int x25_seq_socket_show(struct seq_file *seq, void *v)
goto out;
}
- s = v;
+ s = sk_entry(v);
x25 = x25_sk(s);
if (!x25->neighbour || (dev = x25->neighbour->dev) == NULL)
--
1.6.3
^ permalink raw reply related
* [PATCH 02/13] net: add a wrapper sk_entry()
From: Li Zefan @ 2010-02-05 1:45 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
include/net/sock.h | 5 +++++
1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/include/net/sock.h b/include/net/sock.h
index 3f1a480..c8d4000 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -317,6 +317,11 @@ struct sock {
/*
* Hashed lists helper routines
*/
+static inline struct sock *sk_entry(const struct hlist_node *node)
+{
+ return hlist_entry(node, struct sock, sk_node);
+}
+
static inline struct sock *__sk_head(const struct hlist_head *head)
{
return hlist_entry(head->first, struct sock, sk_node);
--
1.6.3
^ permalink raw reply related
* Re: [PATCH 00/13] net: simplify seq_file code
From: Li Zefan @ 2010-02-05 1:44 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
In-Reply-To: <4B6B77FE.5000104@cn.fujitsu.com>
[PATCH 01/13] seq_file: Add helpers for iteration over a hlist
Some places in kernel need to iterate over a hlist in seq_file,
so provide some common helpers.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
fs/seq_file.c | 37 ++++++++++++++++++++++++++++++++++---
include/linux/seq_file.h | 11 +++++++++++
2 files changed, 45 insertions(+), 3 deletions(-)
diff --git a/fs/seq_file.c b/fs/seq_file.c
index eae7d9d..e833076 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -674,7 +674,6 @@ struct list_head *seq_list_start(struct list_head *head, loff_t pos)
return NULL;
}
-
EXPORT_SYMBOL(seq_list_start);
struct list_head *seq_list_start_head(struct list_head *head, loff_t pos)
@@ -684,7 +683,6 @@ struct list_head *seq_list_start_head(struct list_head *head, loff_t pos)
return seq_list_start(head, pos - 1);
}
-
EXPORT_SYMBOL(seq_list_start_head);
struct list_head *seq_list_next(void *v, struct list_head *head, loff_t *ppos)
@@ -695,5 +693,38 @@ struct list_head *seq_list_next(void *v, struct list_head *head, loff_t *ppos)
++*ppos;
return lh == head ? NULL : lh;
}
-
EXPORT_SYMBOL(seq_list_next);
+
+struct hlist_node *seq_hlist_start(struct hlist_head *head, loff_t pos)
+{
+ struct hlist_node *node;
+
+ hlist_for_each(node, head)
+ if (pos-- == 0)
+ return node;
+ return NULL;
+}
+EXPORT_SYMBOL(seq_hlist_start);
+
+struct hlist_node *seq_hlist_start_head(struct hlist_head *head, loff_t pos)
+{
+ if (!pos)
+ return SEQ_START_TOKEN;
+
+ return seq_hlist_start(head, pos - 1);
+}
+EXPORT_SYMBOL(seq_hlist_start_head);
+
+struct hlist_node *seq_hlist_next(void *v, struct hlist_head *head,
+ loff_t *ppos)
+{
+ struct hlist_node *node = v;
+
+ ++*ppos;
+ if (v == SEQ_START_TOKEN)
+ return head->first;
+ else
+ return node->next;
+}
+EXPORT_SYMBOL(seq_hlist_next);
+
diff --git a/include/linux/seq_file.h b/include/linux/seq_file.h
index 8366d8f..c95bcdc 100644
--- a/include/linux/seq_file.h
+++ b/include/linux/seq_file.h
@@ -135,4 +135,15 @@ extern struct list_head *seq_list_start_head(struct list_head *head,
extern struct list_head *seq_list_next(void *v, struct list_head *head,
loff_t *ppos);
+/*
+ * Helpers for iteration over hlist_head-s in seq_files
+ */
+
+extern struct hlist_node *seq_hlist_start(struct hlist_head *head,
+ loff_t pos);
+extern struct hlist_node *seq_hlist_start_head(struct hlist_head *head,
+ loff_t pos);
+extern struct hlist_node *seq_hlist_next(void *v, struct hlist_head *head,
+ loff_t *ppos);
+
#endif
--
1.6.3
^ permalink raw reply related
* [PATCH 00/13] net: simplify seq_file code
From: Li Zefan @ 2010-02-05 1:44 UTC (permalink / raw)
To: David Miller; +Cc: Andrew Morton, LKML, netdev@vger.kernel.org
This patchset introduces seq_hlist_foo() helpers, and convert
net/* to seq_hlist_foo() and seq_list_foo().
---
fs/seq_file.c | 37 ++++++++++++-
include/linux/seq_file.h | 11 ++++
include/net/sock.h | 5 ++
net/appletalk/atalk_proc.c | 30 +----------
net/atm/proc.c | 2 +-
net/atm/resources.c | 28 +++--------
net/ax25/af_ax25.c | 18 +-----
net/ax25/ax25_uid.c | 25 ++--------
net/ipx/ipx_proc.c | 89 ++++---------------------------
net/irda/irlan/irlan_common.c | 26 ++--------
net/key/af_key.c | 20 +------
net/netrom/af_netrom.c | 21 +-------
net/netrom/nr_route.c | 53 +++----------------
net/packet/af_packet.c | 20 +------
net/rose/af_rose.c | 22 +-------
net/x25/x25_proc.c | 114 ++++-------------------------------------
16 files changed, 116 insertions(+), 405 deletions(-)
^ permalink raw reply
* Re: [RFC Patch] net: reserve ports for applications using fixed port numbers
From: Octavian Purdila @ 2010-02-05 1:05 UTC (permalink / raw)
To: Tetsuo Handa
Cc: davem, amwang, linux-kernel, eric.dumazet, linux-rdma, netdev,
nhorman, linux-sctp
In-Reply-To: <201002050041.o150fCO8081208@www262.sakura.ne.jp>
On Friday 05 February 2010 02:41:12 you wrote:
> David Miller wrote:
> > > Octavian Purdila wrote:
> > >> int inet_is_reserved_local_port(int port)
> > >> {
> > >> if (test_bit(port, reserved_ports))
> > >> return 1;
> > >> return 0;
> > >> }
> > >
> > > Above check is exactly what I'm doing in the LSM hook.
> >
> > But his version can be done inline in 2 or 3 instructions.
> >
> > An LSM hook will result in an indirect function call,
> > all live registers spilled to the stack, then all of
> > those reloaded when the function returns.
> >
> > It will be much more expensive.
>
> If you can accept his version, I want to use his version (with an interface
> for updating above "reserved_ports" by not only root user's sysctl() but
> also MAC's policy configuration).
>
I think that simply using an interface to update the reserved_ports from MAC
policy configuration module wouldn't work, as root will be able to modify the
policy via sysctl.
I think that we might need to:
a) have a reserved_port updater
b) put a LSM hook into that
c) use the reserved_port updater from sysctl
^ permalink raw reply
* Re: [PATCH] seastar - SeaStar Ethernet driver
From: David Miller @ 2010-02-05 1:04 UTC (permalink / raw)
To: ktpedre; +Cc: netdev, linux-kernel
In-Reply-To: <1265330214.23523.78.camel@hawkeye.sandia.gov>
From: "Kevin Pedretti" <ktpedre@sandia.gov>
Date: Thu, 4 Feb 2010 17:36:54 -0700
> Sure, I'm open to doing this and would appreciate any example drivers
> you could point me at.
I can't think of any sorry.
> We do need to keep the capability though of talking to Cray's
> existing proprietary "ethernet" over seastar driver, which the
> existing driver does. If we go this route of being a non-Ethernet
> device, do you think it would still be possible to support this?
I don't think it will break communication with Cray's stuff.
^ permalink raw reply
* Re: [PATCH v2 net-next-2.6] packet: Add GSO/csum offload support.
From: Sridhar Samudrala @ 2010-02-05 1:02 UTC (permalink / raw)
To: Herbert Xu; +Cc: Michael S. Tsirkin, netdev
In-Reply-To: <20100205005256.GA9969@gondor.apana.org.au>
On Fri, 2010-02-05 at 11:52 +1100, Herbert Xu wrote:
> On Wed, Feb 03, 2010 at 03:22:20PM -0800, Sridhar Samudrala wrote:
> >
> > - skb_reserve(skb, LL_RESERVED_SPACE(dev));
> > - skb_reset_network_header(skb);
> > + skb_set_network_header(skb, reserve);
>
> Is this going to produce the right network header for the SOCK_DGRAM
> case?
In the case of SOCK_DGRAM reserve is not set and will be 0 and hence
it is same as calling skb_reset_network_header(skb)
Thanks
Sridhar
^ permalink raw reply
* Re: [PATCH] seastar - SeaStar Ethernet driver
From: Kevin Pedretti @ 2010-02-05 0:59 UTC (permalink / raw)
To: Alan Cox; +Cc: David Miller, netdev@vger.kernel.org,
linux-kernel@vger.kernel.org
In-Reply-To: <20100204231621.7a13695a@lxorguk.ukuu.org.uk>
On Thu, 2010-02-04 at 16:16 -0700, Alan Cox wrote:
> - tcpdump etc once coaxed would display seastar frames not fake ethernet
> - the config tools would actually report what it really was
> - non IP layers and userspace won't keep trying to do things you don't
> want (what does it do right now with vlans I wonder 8))
> - there will be no ARP confusion
>
> If it wants to stay compatible and pretend to be ethernet you probably
> need a message type for "encapsulated ethernet", you can then encapsulate
> anything not IP and stay compatible by keeping IP sent the way it is
> now ? Thats if it wants to in the first place.
I agree the current situation is rather confusing and a lot or most of
the standard tools will break. I have used tcpdump with some success,
but I'm sure there's lots of stuff broken.
We do need to stay compatible with the existing proprietary driver. The
usage model we're after is leaving all of the service nodes (nodes users
login to, serve I/O, etc.) booted with the existing proprietary software
and rebooting the compute nodes (99% of the nodes) with a modern Linux
kernel running this open-source seastar driver. The compute nodes need
to be able to communicate with each other and with the service nodes
using IP (point-to-point). Users login to the service nodes and then,
for example, can ssh to compute nodes, scp files to compute nodes,
etc.
Probably critical background: each service node and each compute node
has a seastar NIC that directly connects to nearest neighbors, forming a
3-D torus. The only way for the nodes to communicate is via this
network.
I like the idea of encapsulating whole ethernet frames in a new message
type. But won't the lack of broadcast (and multi-cast) still be an
issue for most protocols? I also don't have much control over the
seastar header, but can probably find an ignored bit somewhere to mark
the new message type.
Kevin
^ permalink raw reply
* Re: [PATCH v2 net-next-2.6] packet: Add GSO/csum offload support.
From: Herbert Xu @ 2010-02-05 0:53 UTC (permalink / raw)
To: Sridhar Samudrala; +Cc: Michael S. Tsirkin, netdev
In-Reply-To: <1265329020.31059.18.camel@w-sridhar.beaverton.ibm.com>
On Thu, Feb 04, 2010 at 04:17:00PM -0800, Sridhar Samudrala wrote:
>
> > So how is the caller going to pass the protocol value through?
>
> via sockaddr_ll->sll_protocol in msg->msg_name.
OK that works.
Thanks!
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* Re: [PATCH v2 net-next-2.6] packet: Add GSO/csum offload support.
From: Herbert Xu @ 2010-02-05 0:52 UTC (permalink / raw)
To: Sridhar Samudrala; +Cc: Michael S. Tsirkin, netdev
In-Reply-To: <1265239340.5600.21.camel@w-sridhar.beaverton.ibm.com>
On Wed, Feb 03, 2010 at 03:22:20PM -0800, Sridhar Samudrala wrote:
>
> - skb_reserve(skb, LL_RESERVED_SPACE(dev));
> - skb_reset_network_header(skb);
> + skb_set_network_header(skb, reserve);
Is this going to produce the right network header for the SOCK_DGRAM
case?
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
* Re: [RFC Patch] net: reserve ports for applications using fixed port numbers
From: Tetsuo Handa @ 2010-02-05 0:41 UTC (permalink / raw)
To: davem
Cc: opurdila, amwang, linux-kernel, eric.dumazet, linux-rdma, netdev,
nhorman, linux-sctp
In-Reply-To: <20100204.135639.69720709.davem@davemloft.net>
David Miller wrote:
> > Octavian Purdila wrote:
> >>
> >> int inet_is_reserved_local_port(int port)
> >> {
> >> if (test_bit(port, reserved_ports))
> >> return 1;
> >> return 0;
> >> }
> >>
> > Above check is exactly what I'm doing in the LSM hook.
>
> But his version can be done inline in 2 or 3 instructions.
>
> An LSM hook will result in an indirect function call,
> all live registers spilled to the stack, then all of
> those reloaded when the function returns.
>
> It will be much more expensive.
If you can accept his version, I want to use his version (with an interface for
updating above "reserved_ports" by not only root user's sysctl() but also MAC's
policy configuration).
^ permalink raw reply
* Re: [PATCH] seastar - SeaStar Ethernet driver
From: Kevin Pedretti @ 2010-02-05 0:36 UTC (permalink / raw)
To: David Miller; +Cc: netdev@vger.kernel.org, linux-kernel@vger.kernel.org
In-Reply-To: <20100204.151357.18104114.davem@davemloft.net>
On Thu, 2010-02-04 at 16:13 -0700, David Miller wrote:
> From: "Kevin Pedretti" <ktpedre@sandia.gov>
> Date: Thu, 4 Feb 2010 16:01:04 -0700
>
> > Please let me know if these issues are show-stoppers as far as inclusion
> > goes. We would like to get this open-source driver into the kernel so
> > us and others with Cray XT systems can start to benefit from it, and
> > continue to automatically track kernel API changes.
>
> I may let it in as-is, thanks for the explanations.
>
> Could you at least consider perhaps that since it isn't
> really ethernet, not pretending that it is may be something
> to look at?
Sure, I'm open to doing this and would appreciate any example drivers
you could point me at. We do need to keep the capability though of
talking to Cray's existing proprietary "ethernet" over seastar driver,
which the existing driver does. If we go this route of being a
non-Ethernet device, do you think it would still be possible to support
this?
Kevin
^ permalink raw reply
* Re: [PATCH v2 net-next-2.6] packet: Add GSO/csum offload support.
From: Sridhar Samudrala @ 2010-02-05 0:17 UTC (permalink / raw)
To: Herbert Xu; +Cc: Michael S. Tsirkin, netdev
In-Reply-To: <20100205000816.GA9575@gondor.apana.org.au>
On Fri, 2010-02-05 at 11:08 +1100, Herbert Xu wrote:
> On Wed, Feb 03, 2010 at 03:22:20PM -0800, Sridhar Samudrala wrote:
> >
> > - protocol value is assumed to be passed in by the caller.
>
> So how is the caller going to pass the protocol value through?
via sockaddr_ll->sll_protocol in msg->msg_name.
Thanks
Sridhar
^ permalink raw reply
* Re: [PATCH v2 net-next-2.6] packet: Add GSO/csum offload support.
From: Herbert Xu @ 2010-02-05 0:08 UTC (permalink / raw)
To: Sridhar Samudrala; +Cc: Michael S. Tsirkin, netdev
In-Reply-To: <1265239340.5600.21.camel@w-sridhar.beaverton.ibm.com>
On Wed, Feb 03, 2010 at 03:22:20PM -0800, Sridhar Samudrala wrote:
>
> - protocol value is assumed to be passed in by the caller.
So how is the caller going to pass the protocol value through?
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox