In case the binfmt_misc binary handler is registered *before* the e.g.
script one (when for example being compiled as a module) the following
situation may occur:
1. user launches a script, whose interpreter is a misc binary;
2. the load_misc_binary sets the misc_bang and returns -ENOEVEC,
since the binary is a script;
3. the load_script_binary loads one and calls for search_binary_hander
to run the interpreter;
4. the load_misc_binary is called again, but refuses to load the
binary due to misc_bang bit set.
The fix is to move the misc_bang setting lower - prior to the actual
call to the search_binary_handler.
Caused by the commit 3a2e7f47 (binfmt_misc.c: avoid potential kernel
stack overflow)
Signed-off-by: Pavel Emelyanov <[email protected]>
Reported-by: Kirill A. Shutemov <[email protected]>
Tested-by: Kirill A. Shutemov <[email protected]>
Cc: <[email protected]> [2.6.26.x]
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Deleting a timer with del_timer doesn't guarantee, that the
timer function is not running at the moment of deletion. Thus
in the xt_hashlimit case we can get into a ticklish situation
when the htable_gc rearms the timer back and we'll actually
delete an entry with a pending timer.
Fix it with using del_timer_sync().
AFAIK del_timer_sync checks for the timer to be pending by
itself, so I remove the check.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: Patrick McHardy <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The thing is that recent_mt_destroy first flushes the entries
from table with the recent_table_flush and only *after* this
removes the proc file, corresponding to that table.
Thus, if we manage to write to this file the '+XXX' command we
will leak some entries. If we manage to write there a 'clean'
command we'll race in two recent_table_flush flows, since the
recent_mt_destroy calls this outside the recent_lock.
The proper solution as I see it is to remove the proc file first
and then go on with flushing the table. This flushing becomes
safe w/o the lock, since the table is already inaccessible from
the outside.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: Patrick McHardy <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Fix the one describing what this function is and add one more - about
locking absence around pid namespaces loop.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Randy Dunlap <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This just makes the acct_proces walk the pid namespaces from current up to
the top and account a task in each with the accounting turned on.
ns->parent access if safe lockless, since current it still alive and holds
its namespace, which in turn holds its parent.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
All the bsd_acct_strcts with opened accounting are linked into a global
list. So, the acct_auto_close(_mnt) walks one and drops the accounting
for each.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Allocate the structure on the first call to sys_acct(). After this each
namespace, that ordered the accounting, will live with this structure till
its own death.
Two notes
- routines, that close the accounting on fs umount time use
the init_pid_ns's acct by now;
- accounting routine accounts to dying task's namespace
(also by now).
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This adds the appropriate pointer to all the internal (i.e. static)
functions that work with global acct instance. API calls pass a global
instance to them (while we still have such).
Mostly this is a s/acct_globals./acct->/ over the file.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Don't use per-bsd-acct-struct lock, but work with a global one.
This lock is taken for short periods, so it doesn't seem it'll become a
bottleneck, but it will allow us to easily avoid many locking difficulties
in the future.
So this is a mostly s/acct_globals.lock/acct_lock/ over the file.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We're going to have many bsd_acct_struct instances, not just one, so the
timer (currently working with a global one) has to know which one to work
with.
Use a handy setup_timer macro for it (thanks to Oleg for one).
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
All the bsdacct-related info will be stored in the area, pointer by this
one.
It will be NULL automatically for all new namespaces.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
It makes many fields initialization implicit helping in auto-setting
#ifdef-ed fields (bsd-acct related pointer will be such).
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
After I fixed access to task->tgid in kernel/acct.c, Oleg pointed out some
bad side effects with this accounting vs pid namespaces interaction. I.e.
when some task in pid namespace sets this accounting up, this blocks all
the others from doing the same. Restricting this to init namespace only
could help, but didn't look a graceful solution.
So here is the approach to make this accounting work with pid namespaces
properly.
The idea is simple - when a task dies it accounts itself in each namespace
it is visible from and which set the accounting up.
For example here are the commands run and the output of lastcomm from init
and sub namespaces:
init_ns# accton pacct
sub_ns# accton pacct (this is a different file - sub ns is run in
a chroot-ed environment)
init_ns# cat /dev/null
sub_ns# ls /dev/null
init_ns# accton
sub_ns# accton
sub_ns# lastcomm -f pacct
ls 0 [136,0] 0.00 secs Thu May 15 10:30
accton 0 [136,0] 0.00 secs Thu May 15 10:30
init_ns# lastcomm -f pacct
accton root pts/0 0.00 secs Thu May 15 14:30 << got from sub
cat root pts/1 0.00 secs Thu May 15 14:30
ls root pts/0 0.00 secs Thu May 15 14:30 << got from sub
accton root pts/1 0.00 secs Thu May 15 14:30
That was the summary, the details are in patches.
This patch:
It will be visible in pid_namespace.h file, so fix its name to look better
outside the acct.c file.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
It seems to me that it was a mistake marking this function as deprecated
and scheduling it for removal, rather than resolutely removing it after
the last caller's death.
Anyway - better late, then never.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This one had the only users so far - the kill_proc, which is removed, so
drop this (invalid in namespaced world) call too.
And of course - erase all references on it from comments.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This function operated on a pid_t to kill a task, which is no longer valid
in a containerized system.
It has finally lost all its users and we can safely remove it from the
tree.
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Oleg Nesterov <[email protected]>
Cc: "Eric W. Biederman" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently this list is protected with a simple spinlock, even for reading
from one. This is OK, but can be better.
Actually I want it to be better very much, since after replacing the
OpenVZ device permissions engine with the cgroup-based one I noticed, that
we set 12 default device permissions for each newly created container (for
/dev/null, full, terminals, ect devices), and people sometimes have up to
20 perms more, so traversing the ~30-40 elements list under a spinlock
doesn't seem very good.
Here's the RCU protection for white-list - dev_whitelist_item-s are added
and removed under the devcg->lock, but are looked up in permissions
checking under the rcu_read_lock.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: Paul Menage <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Ignoring their return values may result in counter underflow in the future -
when the value charged will be uncharged (or in "leaks" - when the value is
not uncharged).
This also prevents from using charging routines to decrement the
counter value (i.e. uncharge it) ;)
(Current code works OK with res_counter, however :) )
Signed-off-by: Pavel Emelyanov <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: Paul Menage <[email protected]>
Cc: KAMEZAWA Hiroyuki <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
There are already 7 of them - time to kill some duplicate code.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
After all this stuff is moved outside, this function can look better.
Besides, I tuned the error path in ip_proc_init_net to make it have
only 2 exit points, not 3.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This one has become per-net long ago, but the appropriate file
is per-net only now.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
All the statistics shown in this file have been made per-net already.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Now all the shown in it statistics is netnsizated, time to
show it in appropriate net.
The appropriate net init/exit ops already exist - they make
the sockstat file per net - so just extend them.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
After moving all the stuff outside this function it looks
a bit ugly - make it look better.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Proc temporary uses stats from init_net.
BTW, TCP_XXX_STATS are beautiful (w/o do { } while (0) facing) again :)
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
These ones are currently empty, but stuff from init_ipv4_mibs will
sequentially migrate there.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The only structure declared within is the netns_mib, which will
carry all our mibs within. I didn't put the mibs in the existing
netns_xxx structures to make it possible to mark this one as
properly aligned and get in a separate "read-mostly" cache-line.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Done with NET_XXX_STATS macros :)
To be continued...
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This one is tricky.
The thing is that this macro is only used when killing tw buckets,
but since this killer is promiscuous wrt to which net each particular
tw belongs to, I have to use it only when NET_NS is off. When the net
namespaces are on, I use the INET_INC_STATS_BH for each bucket.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
These places have a tcp_sock, but we'd prefer the sock itself to
get net from it. Fortunately, tcp_sk macro is just a type cast, so
this replace is really cheap.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The tcp_enter_memory_pressure calls NET_INC_STATS, but doesn't
have where to get the net from.
I decided to add a sk argument, not the net itself, only to factor
all the required sock_net(sk) calls inside the enter_memory_pressure
callback itself.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Same as before - the sock is always there to get the net from,
but there are also some places with the net already saved on
the stack.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Fortunately (almost) all the TCP code has a sock to get the net from :)
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This one sets TCP MIBs after zeroing them, and thus requires
the net.
The existing single caller can use init_net (temporarily).
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
TCP_INC_STATS_USER and TCP_ADD_STATS_BH are currently unused.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This is the same as the first patch in the set, but preparing
the net for TCP_XXX_STATS - save the struct net on the stack
where required and possible.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Very simple - only ip_evictor (fragments) requires such.
This patch ends up the IP_XXX_STATS patching.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
All the callers already have either the net itself, or the place
where to get it from.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Some places, that deal with IP statistics already have where to
get a struct net from, but use it directly, without declaring
a separate variable on the stack.
So, save this net on the stack for future IP_XXX_STATS macros.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
There are ICMP_XXX_STATS that are not used in the kernel, so I remove
them, not to "just patch" them later. But if there's some sense in
keeping them, kick me - I will remake this set keeping them.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Some places, that deal with ICMP statistics already have where
to get a struct net from, but use it directly, without declaring
a separate variable on the stack.
Since I will need this net soon, I declare a struct net on the
stack and use it in the existing places in a separate patch not
to spoil the future ones.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This routine deals with ICMP statistics, but doesn't have a
struct net at hands, so add one.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Two special cases here - one is rxrpc - I put init_net there
explicitly, since we haven't touched this part yet. The second
place is in __udp4_lib_rcv - we already have a struct net there,
but I have to move its initialization above to make it ready
at the "drop" label.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Denis V. Lunev <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Nothing special - all the places already have a struct sock
at hands, so use the sock_net() net.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Denis V. Lunev <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
There are some places in TCP that select one MIB index to
bump snmp statistics like this:
if (<something>)
NET_INC_STATS_BH(<some_id>);
else if (<something_else>)
NET_INC_STATS_BH(<some_other_id>);
...
else
NET_INC_STATS_BH(<default_id>);
or in a more tricky but still similar way.
On the other hand, this NET_INC_STATS_BH is a camouflaged
increment of percpu variable, which is not that small.
Factoring those cases out de-bloats 235 bytes on non-preemptible
i386 config and drives parts of the code into 80 columns.
add/remove: 0/0 grow/shrink: 0/7 up/down: 0/-235 (-235)
function old new delta
tcp_fastretrans_alert 1437 1424 -13
tcp_dsack_set 137 124 -13
tcp_xmit_retransmit_queue 690 676 -14
tcp_try_undo_recovery 283 265 -18
tcp_sacktag_write_queue 1550 1515 -35
tcp_update_reordering 162 106 -56
tcp_retransmit_timer 990 904 -86
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The problem is that while we work w/o the inet_frags.lock even
read-locked the secret rebuild timer may occur (on another CPU, since
BHs are still disabled in the inet_frag_find) and change the rnd seed
for ipv4/6 fragments.
It was caused by my patch fd9e63544c
([INET]: Omit double hash calculations in xxx_frag_intern) late
in the 2.6.24 kernel, so this should probably be queued to -stable.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
After the sctp_remaddr_proc_init failed, the proper rollback is
not the sctp_remaddr_proc_exit, but the sctp_assocs_proc_exit.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Vlad Yasevich <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
There are many possible ways to add this "salt", thus I made this
patch to be the last in the series to change it if required.
Currently I propose to use the struct net pointer itself as this
salt, but since this pointer is most often cache-line aligned, shift
this right to eliminate the bits, that are most often zeroed.
After this, simply add this mix to prepared hashfn-s.
For CONFIG_NET_NS=n case this salt is 0 and no changes in hashfn
appear.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Same as for inet_hashfn, prepare its ipv6 incarnation.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Although this hash takes addresses into account, the ehash chains
can also be too long when, for instance, communications via lo occur.
So, prepare the inet_hashfn to take struct net into account.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Listening-on-one-port sockets in many namespaces produce long
chains in the listening_hash-es, so prepare the inet_lhashfn to
take struct net into account.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Binding to some port in many namespaces may create too long
chains in bhash-es, so prepare the hashfn to take struct net
into account.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Every caller already has this one. The new argument is currently
unused, but this will be fixed shortly.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
They both calculate the hash chain, but currently do not have
a struct net pointer, so pass one there via additional argument,
all the more so their callers already have such.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Currently the chain to store a UDP socket is calculated with
simple (x & (UDP_HTABLE_SIZE - 1)). But taking net into account
would make this calculation a bit more complex, so moving it into
a function would help.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The bindv6only is tuned via sysctl. It is already on a struct net
and per-net sysctls allow for its modification (ipv6_sysctl_net_init).
Despite this the value configured in the init net is used for the
rest of them.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Consider you added a 'c foo:bar r' permission to some cgroup and then (a
bit later) 'c'foo:bar w' for it. After this you'll see the
c foo:bar r
c foo:bar w
lines in a devices.list file.
Another example - consider you added 10 'c foo:bar r' permissions to some
cgroup (e.g. by mistake). After this you'll see 10 c foo:bar r lines in
a list file.
This is weird. This situation also has one more annoying consequence.
Having many items in a white list makes permissions checking slower, sine
it has to walk a longer list.
The proposal is to merge permissions for items, that correspond to the
same device.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently even if a task sits in an all-denied cgroup it can still mount
any block device in any mode it wants.
Put a proper check in do_open for block device to prevent this.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Tested-by: Serge Hallyn <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Two functions, that need to get a device_cgroup from a task (they are
devcgroup_inode_permission and devcgroup_inode_mknod) make it in a strange
way:
They get a css_set from task, then a subsys_state from css_set, then a
cgroup from the state and then a subsys_state again from the cgroup.
Besides, the devices_subsys_id is read from memory, whilst there's a
enum-ed constant for it.
Optimize this part a bit:
1. Get the subsys_stats form the task and be done - no 2 extra
dereferences,
2. Use the device_subsys_id constant, not the value from memory
(i.e. one less dereference).
Found while preparing 2.6.26 OpenVZ port.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Acked-by: Paul Menage <[email protected]>
Cc: Balbir Singh <[email protected]>
Cc: James Morris <[email protected]>
Cc: Chris Wright <[email protected]>
Cc: Stephen Smalley <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Bad type/protocol specified result in sk leak.
Fix is simple - release the sk if bad values are given,
but to make it possible just to call sk_free(), I move
some sk initialization a bit lower.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Some places iterate over the checked list right after the check
itself, so even if the list is empty, the list_for_each_xxx
iterator will make everything right by himself.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Jay Vosburgh <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
Many places either do not modify the list under the list_for_each_xxx,
or break out of the loop as soon as the first element is removed.
Thus, this _safe iteration just occupies some unneeded .text space
and requires an additional variable.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Jay Vosburgh <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
While we're fixing the bond_create, I hope it's OK to polish it
a bit after the fixes.
The third argument is NULL at the first caller and is ignored by
the second one, so remove it.
Signed-off-by: Pavel Emelyanov <[email protected]>
Acked-by: Jay Vosburgh <[email protected]>
Signed-off-by: Jeff Garzik <[email protected]>
There are already two places, that kfree the mesh_table and
its buckets.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
This is the first (of two) clean ups after the fixes above.
The err variable is not even required after this cleaning.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
In case the hash grow failed, it is not fair to return error -
the new node _was_ _actually_ added in this case.
Besides, after my previous patch, this grow is more likely
to fail on large hashes.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
The mesh_path_node_copy() can be called like this:
mesh_path_add
`- write_lock(&pathtbl_resize_lock); /* ! */
`- mesh_table_grow
`- ->copy_node
`- mesh_path_node_copy
thus, the GFP_KERNEL is not suitable here.
The acceptable fix, I suppose, is make this allocation GPF_ATOMIC -
the mpath_node being allocated is 4 pointers, i.e. this allocation
is small enough to survive even under a moderate memory pressure.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
The mesh_path_node_copy() performs kmalloc() and thus - may fail
(well, it does not now, but I'm fixing this right now). Its caller -
the mesh_table_grow() - isn't prepared for such a trick yet.
This preparation is just flush the new hash and make copy_node()
return an int value.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
(This set applies OK without the previous one of 4 patches,
but with some fuzz in the 7th one)
The mesh_path_node_free() does so under hashwlock.
But, this one is called
1. from mesh_path_add() after an old hash is hidden and
synchronize_rcu() is calld
2. mesh_pathtbl_unregister(), when the module is being
unloaded and no devices exist to mess with this hash.
So, it seems to me, that simply removing the call is OK.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
There are already tree paths, that do incremental rollbacks, so
merge them together, rename labels and format the code to look a
bit nicer.
(I do not mind dropping/delaying this patch however).
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: John W. Linville <[email protected]>
These devices use the private area of appropriate size for
statistics. Turning them to use on-device ones make them
"privless" and thus - really small wrt kmalloc cache, they
are allocated from.
Besides, code looks nicer, because of absence of multi-braced
type casts and dereferences.
[ Fix build failures -DaveM ]
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This tunnel uses its own private structure and requires separate
patch to switch from private stats to on-device ones.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
All users already use on-device statistics, so this field can be
safely removed.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The ipmr uses ipip tunnels for its purposes and updates the
tunnels' stats, but the ipip driver is already switched to
use on-device ones.
Actually, this is a part of the patch #4 from this set.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Pavel Emelyanov