- 16 Apr, 2008 34 commits
-
-
Pavel Emelyanov authored
As for the IPIP tunnel, there are some ip_route_output_key() calls in there that require a proper net so give one to them. And a proper net for the __get_dev_by_index hanging around. Signed-off-by:
Pavel Emelyanov <xemul@openvz.org> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Pavel Emelyanov authored
Very similar to what was done for the IPIP code. Signed-off-by:
-
Pavel Emelyanov authored
Everything is prepared for this change now. Create on in init callback, use it over the code and destroy on net exit. Signed-off-by:
-
Pavel Emelyanov authored
This is the part#2 of the patch #2 - get the proper net for these functions. This change in a separate patch in order not to get lost in a large previous patch. Signed-off-by:
-
Pavel Emelyanov authored
The fallback device and hashes are to become per-net, but many code doesn't have anything to get the struct net pointer from. So pass the proper net there with an extra argument. Signed-off-by:
-
Pavel Emelyanov authored
Signed-off-by:
-
Pavel Emelyanov authored
This one was disabled by default for sanity. Signed-off-by:
-
Pavel Emelyanov authored
Set the proper net before calling register_netdev and disable the tunnel device netns changing. Signed-off-by:
-
Pavel Emelyanov authored
There are some ip_route_output_key() calls in there that require a proper net so give one to them. Besides - give a proper net to a single __get_dev_by_index call in ipip_tunnel_bind_dev(). Signed-off-by:
-
Pavel Emelyanov authored
Either net or ipip_net already exists in all the required places, so just use one. Besides, tune net_init and net_exit calls to respectively initialize the hashes and destroy devices. Signed-off-by:
-
Pavel Emelyanov authored
This is the part#2 of the previous patch - get the proper net for these functions. I make it in a separate patch, so that this change does not get lost in a large previous patch. Signed-off-by:
-
Pavel Emelyanov authored
The hashes of tunnels will be per-net too, so prepare all the functions that uses them for this change by adding an argument. Use init_net temporarily in places, where the net does not exist explicitly yet. Signed-off-by:
-
Pavel Emelyanov authored
Create on in ipip_init_net(), use it all over the code (the proper place to get the net from already exists) and destroy in ipip_net_exit(). Signed-off-by:
-
Pavel Emelyanov authored
Signed-off-by:
-
Pavel Emelyanov authored
When van device is moved to another namespace proc files, related to this device, should also change one. Use the netdev REGISTER and UNREGISTER event handlers for this. Signed-off-by:
Patrick McHardy <kaber@trash.net> Signed-off-by:
-
Pavel Emelyanov authored
This one is similar to what I've done for TUN - set the proper net after device allocation and clean VLANs on net exit (use the rtnl_kill_links helper finally). Plus, drop explicit init_net usage and net != &init_net checks. Signed-off-by:
-
Pavel Emelyanov authored
This includes moving one on the struct vlan_net and s/vlan_name_type/vn->name_type/ over the code. Signed-off-by:
-
Pavel Emelyanov authored
It is created in a proper net, so make is show info, related to this particular net. Signed-off-by:
-
Pavel Emelyanov authored
The proc_vlan_dir and proc_vlan_conf migrate on the struct vlan_net and their creation uses the struct net. The devices' entries use the corresponding device's net. Signed-off-by:
-
Pavel Emelyanov authored
All proc files will be created in each net, so prepare them for this change now, not to mess it with real creation patch. The net != &init_net checks in them are for git-bisect sanity, but I will drop them soon. Signed-off-by:
-
Pavel Emelyanov authored
Unlike TUN, it is empty from the very beginning, and will be eventually populated later. Signed-off-by:
-
Pavel Emelyanov authored
Currently vlan group is searched using one key - the ifindex. We'll have to lookup the vlan_group by two keys - ifindex and net. Turning the vlan_group lookup key to struct net_device pointer will make this process easier. Besides, this will eliminate one more place in the networking, that assumes that indexes are unique in the kernel. Signed-off-by:
-
Pavel Emelyanov authored
This one is responsible for calling ->dellink on each net device found in net to help with vlan net_exit hook in the nearest future. Signed-off-by:
-
Pavel Emelyanov authored
Each potential list_del (happening from inside a ->dellink call) is followed by goto restart, so there's no need in _safe iteration. Signed-off-by:
-
Pavel Emelyanov authored
This is basically means that a net is set for a new device, but actually also involves two more steps: 1. mark the tun device as "local", i.e. do not allow for it to move across namespaces. This is done so, since tun device is most often associated to some file (and thus to some process) and moving the device alone is not valid while keeping the file and the process outside. The need in ability to move a detached persistent device is to be investigated later. 2. get the tun device's net when tun becomes attached and put one when it becomes detached. This is needed to handle the case when a task owning the tun dies, but a files lives for some more time - in this case we must not allow for net to be freed, since its exit hook will spoil that file's private data by unregistering the tun from under tun_chr_close. Signed-off-by:
-
Pavel Emelyanov authored
Remove the static tun_dev_list and replace its occurrences in driver with per-net one. It is used in two places - in tun_set_iff and tun_cleanup. In the first case it's legal to use current net_ns. In the cleanup call - move the loop, that unregisters all devices in net exit hook. Signed-off-by:
-
Pavel Emelyanov authored
This is the first step in making tuntap devices work in net namespaces. The structure mentioned is pointed by generic net pointer with tun_net_id id, and tun driver fills one on its load. It will contain only the tun devices list. So declare this structure and introduce net init and exit hooks. Signed-off-by:
-
Ilpo Järvinen authored
Needed can only be more strict than what was checked by the earlier common case check for non-tail skbs, thus cwnd_len <= needed will never match in that case anyway. Signed-off-by:
Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by:
-
Mandeep Singh Baines authored
Currently, it is not possible to read/write to an eeprom larger than 128k in size because the buffer used for temporarily storing the eeprom contents is allocated using kmalloc. kmalloc can only allocate a maximum of 128k depending on architecture. Modified ethtool_get/set_eeprom to only allocate a page of memory and then copy the eeprom a page at a time. Updated original patch as per suggestions from Joe Perches. Signed-off-by:
Mandeep Singh Baines <msb@google.com> Signed-off-by:
-
Oliver Hartkopp authored
Make use of hrtimers to support high resolution capabilities, when provided by the system clocksource. The conversion to hrtimers additionally discovered and solved an unlikely race condition that has been reproduced under (unrealistic) massive receive load, which can only be produced on vcan software devices. [ Fix printf format warnings on 64-bit -DaveM ] Signed-off-by:
Oliver Hartkopp <oliver@hartkopp.net> Signed-off-by:
-
Allan Stephens authored
This patch ensures that TIPC properly handles incoming messages that have incorrect or unexpected formats. Most significantly, it now ensures that each sl_buff has at least as much data as the message header indicates it should, and that the entire message header is stored contiguously; this prevents TIPC from accidentally accessing memory that is not part of the sk_buff. Signed-off-by:
Allan Stephens <allan.stephens@windriver.com> Signed-off-by:
-
Allan Stephens authored
This patch allows TIPC to process incoming messages that are stored in a fragmented sk_buff, by forcing the linearization of any such messages it receives. Note: This is an interim solution to allow TIPC to operate with Ethernet devices that generate non-linear buffers (such as the gianfar driver), until such time as the rest of TIPC is enhanced to handle sk_buffs with multiple data areas. Signed-off-by:
-
Allan Stephens authored
This patch causes TIPC to allocate fast clonable sk_buffs, rather than standard ones. This speeds up the cloning operation done by the link code each time a message is sent off-node. Signed-off-by:
-
Allan Stephens authored
This patch eliminates a null pointer check when discarding a TIPC message buffer, since kfree_skb() already handles this situation. Acknowledgements to Florian Westphal (fw@strlen.de> for suggesting this enhancement. Signed-off-by:
-
- 15 Apr, 2008 6 commits
-
-
Pavel Emelyanov authored
Add the elastic array of void * pointer to the struct net. The access rules are simple: 1. register the ops with register_pernet_gen_device to get the id of your private pointer 2. call net_assign_generic() to put the private data on the struct net (most preferably this should be done in the ->init callback of the ops registered) 3. do not store any private reference on the net_generic array; 4. do not change this pointer while the net is alive; 5. use the net_generic() to get the pointer. When adding a new pointer, I copy the old array, replace it with a new one and schedule the old for kfree after an RCU grace period. Since the net_generic explores the net->gen array inside rcu read section and once set the net->gen->ptr[x] pointer never changes, this grants us a safe access to generic pointers. Quoting Paul: "... RCU is protecting -only- the net_generic structure that net_generic() is traversing, and the [pointer] returned by net_generic() is protected by a reference counter in the upper-level struct net." Signed-off-by:Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by:
-
Pavel Emelyanov authored
To make some per-net generic pointers, we need some way to address them, i.e. - IDs. This is simple IDA-based IDs generator for pernet subsystems. Addressing questions about potential checkpoint/restart problems: these IDs are "lite-offsets" within the net structure and are by no means supposed to be exported to the userspace. Since it will be used in the nearest future by devices only (tun, vlan, tunnels, bridge, etc), I make it resemble the functionality of register_pernet_device(). The new ids is stored in the *id pointer _before_ calling the init callback to make this id available in this callback. Signed-off-by:
-
Adrian Bunk authored
This patch remove the usage of a nonexisting kconfig variable. Reported-by:
Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by:
Adrian Bunk <bunk@kernel.org> Signed-off-by:
-
Adrian Bunk authored
Even kernel 2.2.26 (sic) already contains the #undef CONFIG_IRLAN_SEND_GRATUITOUS_ARP with the comment "but for some reason the machine crashes if you use DHCP". Either someone finally looks into this or it's simply time to remove this dead code. Reported-by:
-
Adrian Bunk authored
Robert P. J. Day spotted that my removal of the Sangoma drivers missed a few bits. Signed-off-by:
-
Allan Stephens authored
This patch modifies TIPC's socket code to follow the same approach used by other protocols. This change eliminates the need for a mutex in the TIPC-specific portion of the socket protocol data structure -- in its place, the standard Linux socket backlog queue and associated locking routines are utilized. These changes fix a long-standing receive queue bug on SMP systems, and also enable individual read and write threads to utilize a socket without unnecessarily interfering with each other. Signed-off-by:
-
