- 12 May, 2015 19 commits
-
-
Scott Feldman authored
STP update is just a settable port attribute, so convert switchdev_port_stp_update to an attr set. For DSA, the prepare phase is skipped and STP updates are only done in the commit phase. This is because currently the DSA drivers don't need to allocate any memory for STP updates and the STP update will not fail to HW (unless something horrible goes wrong on the MDIO bus, in which case the prepare phase wouldn't have been able to predict anyway). Signed-off-by:
Scott Feldman <sfeldma@gmail.com> Acked-by:
Jiri Pirko <jiri@resnulli.us> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Scott Feldman authored
For rocker, support prepare-commit transaction model for setting attributes (and for adding objects). This requires rocker to preallocate memory needed for the commit up front in the prepare phase. Since rtnl_lock is held between prepare-commit, store the allocated memory on a queue hanging off of the rocker_port. Also, in prepare phase, do everything right up to calling into HW. The same code paths are tranversed in the driver for both prepare and commit phases. In some cases, any state modified in the prepare phase must be reverted before returning so the commit phase makes the same decisions. As a consequence of holding rtnl_lock in process context for all attr sets (and obj adds), all memory is GFP_KERNEL allocated and we don't need to busy spin waiting for the device to complete the command. So the bulk of this patch is simplifying the memory allocations to only use GFP_KERNEL and to remove the nowait flag and busy spin loop. Signed-off-by:
-
Scott Feldman authored
Switch ID is just a gettable port attribute. Convert switchdev op switchdev_parent_id_get to a switchdev attr. Note: for sysfs and netlink interfaces, SWITCHDEV_ATTR_PORT_PARENT_ID is called with SWITCHDEV_F_NO_RECUSE to limit switch ID user-visiblity to only port netdevs. So when a port is stacked under bond/bridge, the user can only query switch id via the switch ports, but not via the upper devices Signed-off-by:
-
Scott Feldman authored
Add two new swdev ops for get/set switch port attributes. Most swdev interactions on a port are gets or sets on port attributes, so rather than adding ops for each attribute, let's define clean get/set ops for all attributes, and then we can have clear, consistent rules on how attributes propagate on stacked devs. Add the basic algorithms for get/set attr ops. Use the same recusive algo to walk lower devs we've used for STP updates, for example. For get, compare attr value for each lower dev and only return success if attr values match across all lower devs. For sets, set the same attr value for all lower devs. We'll use a two-phase prepare-commit transaction model for sets. In the first phase, the driver(s) are asked if attr set is OK. If all OK, the commit attr set in second phase. A driver would NACK the prepare phase if it can't set the attr due to lack of resources or support, within it's control. RTNL lock must be held across both phases because we'll recurse all lower devs first in prepare phase, and then recurse all lower devs again in commit phase. If any lower dev fails the prepare phase, we need to abort the transaction for all lower devs. If lower dev recusion isn't desired, allow a flag SWITCHDEV_F_NO_RECURSE to indicate get/set only work on port (lowest) device. Signed-off-by:
-
Jiri Pirko authored
Turned out that "switchdev" sticks. So just unify all related terms to use this prefix. Signed-off-by:
Roopa Prabhu <roopa@cumulusnetworks.com> Acked-by:
Andy Gospodarek <gospo@cumulusnetworks.com> Signed-off-by:
-
Jiri Pirko authored
Turned out that "switchdev" sticks. So just unify all related terms to use this prefix. Signed-off-by:
-
David Ward authored
In a GRED qdisc, if the default "virtual queue" (VQ) does not have drop parameters configured, then packets for the default VQ are not subjected to RED and are only dropped if the queue is larger than the net_device's tx_queue_len. This behavior is useful for WRED mode, since these packets will still influence the calculated average queue length and (therefore) the drop probability for all of the other VQs. However, for some drivers tx_queue_len is zero. In other cases the user may wish to make the limit the same for all VQs (including the default VQ with no drop parameters). This change adds a TCA_GRED_LIMIT attribute to set the GRED queue limit, in bytes, during qdisc setup. (This limit is in bytes to be consistent with the drop parameters.) The default limit is the same as for a bfifo queue (tx_queue_len * psched_mtu). If the drop parameters of any VQ are configured with a smaller limit than the GRED queue limit, that VQ will still observe the smaller limit instead. Signed-off-by:
David Ward <david.ward@ll.mit.edu> Signed-off-by:
-
Nicolas Schichan authored
Signed-off-by:
Nicolas Schichan <nschichan@freebox.fr> Signed-off-by:
-
David S. Miller authored
Alexander Duyck says: ==================== Refactor netdev page frags and move them into mm/ This patch series addresses several things. First I found an issue in the performance of the pfmemalloc check from build_skb. To work around it I have provided a cached copy of pfmemalloc to be used in __netdev_alloc_skb and __napi_alloc_skb. Second I moved the page fragment allocation logic into the mm tree and added functionality for freeing page fragments. I had to fix igb before I could do this as it was using a reference to NETDEV_FRAG_PAGE_MAX_SIZE incorrectly. Finally I went through and replaced all of the duplicate code that was calling put_page and replaced it with calls to skb_free_frag. With these changes in place a simple receive and drop test increased from a packet rate of 8.9Mpps to 9.8Mpps. The gains breakdown as follows: 8.9Mpps Before 9.8Mpps After ------------------------ ------------------------ 7.8% put_compound_page 9.1% __free_page_frag 3.9% skb_free_head 1.1% put_page 4.9% build_skb 3.8% __napi_alloc_skb 2.5% __alloc_rx_skb 1.9% __napi_alloc_skb ==================== Signed-off-by: -
Alexander Duyck authored
Signed-off-by:
Alexander Duyck <alexander.h.duyck@redhat.com> Signed-off-by:
-
Alexander Duyck authored
Signed-off-by:
-
Alexander Duyck authored
Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
Alexander Duyck authored
Signed-off-by:
-
Alexander Duyck authored
Signed-off-by:
-
Alexander Duyck authored
This change adds a function called skb_free_frag which is meant to compliment the function netdev_alloc_frag. The general idea is to enable a more lightweight version of page freeing since we don't actually need all the overhead of a put_page, and we don't quite fit the model of __free_pages. Signed-off-by:
-
Alexander Duyck authored
This change moves the __alloc_page_frag functionality out of the networking stack and into the page allocation portion of mm. The idea it so help make this maintainable by placing it with other page allocation functions. Since we are moving it from skbuff.c to page_alloc.c I have also renamed the basic defines and structure from netdev_alloc_cache to page_frag_cache to reflect that this is now part of a different kernel subsystem. I have also added a simple __free_page_frag function which can handle freeing the frags based on the skb->head pointer. The model for this is based off of __free_pages since we don't actually need to deal with all of the cases that put_page handles. I incorporated the virt_to_head_page call and compound_order into the function as it actually allows for a signficant size reduction by reducing code duplication. Signed-off-by:
-
Alexander Duyck authored
This change makes it so that we store the virtual address of the page in the netdev_alloc_cache instead of the page pointer. The idea behind this is to avoid multiple calls to page_address since the virtual address is required for every access, but the page pointer is only needed at allocation or reset of the page. While I was at it I also reordered the netdev_alloc_cache structure a bit so that the size is always 16 bytes by dropping size in the case where PAGE_SIZE is greater than or equal to 32KB. Signed-off-by:
-
Alexander Duyck authored
This change updates igb so that it will correctly perform the descriptor count calculation. Previously it was taking NETDEV_FRAG_PAGE_MAX_SIZE into account with isn't really correct since a different value is used to determine the size of the pages used for TCP. That is actually determined by SKB_FRAG_PAGE_ORDER. Signed-off-by:
-
Alexander Duyck authored
While testing I found that the testing for pfmemalloc in build_skb was rather expensive. I found the issue to be two-fold. First we have to get from the virtual address to the head page and that comes at the cost of something like 11 cycles. Then there is the cost for reading pfmemalloc out of the head page which can be cache cold due to the fact that put_page_testzero is likely invalidating the cache-line on one or more CPUs as the fragments can be shared. To avoid this extra expense I have added a pfmemalloc member to the netdev_alloc_cache. I then pushed pieces of __alloc_rx_skb into __napi_alloc_skb and __netdev_alloc_skb so that I could rewrite them to make use of the cached pfmemalloc value. The result is that my perf traces show a reduction from 9.28% overhead to 3.7% for the code covered by build_skb, __alloc_rx_skb, and __napi_alloc_skb when performing a test with the packet being dropped instead of being handed to napi_gro_receive. Signed-off-by:
-
- 11 May, 2015 20 commits
-
-
Eric Dumazet authored
Only left enqueue_root() user is netem, and it looks not necessary : qdisc_skb_cb(skb)->pkt_len is preserved after one skb_clone() Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Michal Simek authored
Use one return statement instead of two to simplify the code. Both are returning the same value. Signed-off-by:
Michal Simek <michal.simek@xilinx.com> Signed-off-by:
-
Philippe Reynes authored
This enables the ethtool's "-d" and "--register-dump" options for fec devices. Signed-off-by:
Philippe Reynes <tremyfr@gmail.com> Signed-off-by:
-
Daniel Borkmann authored
This should have been #ifdef not #if. Reported-by:
Fengguang Wu <fengguang.wu@intel.com> Fixes: d2788d34 ("net: sched: further simplify handle_ing") Signed-off-by:
Daniel Borkmann <daniel@iogearbox.net> Signed-off-by:
-
David S. Miller authored
Daniel Borkmann says: ==================== handle_ing update These are a couple of cleanups to make ingress a bit more lightweight. ==================== Signed-off-by: -
Daniel Borkmann authored
Ingress qdisc has no other purpose than calling into tc_classify() that executes attached classifier(s) and action(s). It has a 1:1 relationship to dev->ingress_queue. After having commit 087c1a60 ("net: sched: run ingress qdisc without locks") removed the central ingress lock, one major contention point is gone. The extra indirection layers however, are not necessary for calling into ingress qdisc. pktgen calling locally into netif_receive_skb() with a dummy u32, single CPU result on a Supermicro X10SLM-F, Xeon E3-1240: before ~21,1 Mpps, after patch ~22,9 Mpps. We can redirect the private classifier list to the netdev directly, without changing any classifier API bits (!) and execute on that from handle_ing() side. The __QDISC_STATE_DEACTIVATE test can be removed, ingress qdisc doesn't have a queue and thus dev_deactivate_queue() is also not applicable, ingress_cl_list provides similar behaviour. In other words, ingress qdisc acts like TCQ_F_BUILTIN qdisc. One next possible step is the removal of the dev's ingress (dummy) netdev_queue, and to only have the list member in the netdevice itself. Note, the filter chain is RCU protected and individual filter elements are being kfree'd by sched subsystem after RCU grace period. RCU read lock is being held by __netif_receive_skb_core(). Joint work with Alexei Starovoitov. Signed-off-by:
Alexei Starovoitov <ast@plumgrid.com> Signed-off-by:
-
Daniel Borkmann authored
Given quite some code has been removed from ing_filter(), we can just consolidate that function into handle_ing() and get rid of a few instructions at the same time. Signed-off-by:
-
Xi Wang authored
Extend the testcase to catch a signedness bug in the arm64 JIT: test_bpf: #58 load 64-bit immediate jited:1 ret -1 != 1 FAIL (1 times) This is useful to ensure other JITs won't have a similar bug. Link: https://lkml.org/lkml/2015/5/8/458 Cc: Alexei Starovoitov <ast@plumgrid.com> Cc: Will Deacon <will.deacon@arm.com> Signed-off-by:
Xi Wang <xi.wang@gmail.com> Acked-by:
-
David S. Miller authored
Jonathan Toppins says: ==================== add netlink support for new lacp bonding parameters This is a resubmit of Mahesh's last 3 bonding patches from this series (http://marc.info/?l=linux-netdev&m=142432864626179&w=2) with one additional kernel patch which adds the netlink bits. I have noted any modifications I did to the original patches just above my signoff line. Patch 5 is the iproute2 support for these bonding options. All patches were coded against the net-next branch of their respective projects. v2: * rebased * only send these new parameters via netlink when bond is in mode 4 * fixed ad_actor_sys_prio to be 0xFFFF by default even when the bond is initially created in mode 0 and switched to mode 4 v3: * reverted changes to bond_option_ad_actor_system_set() from v1 in Mahesh's patch "bonding: Allow userspace to set actors' macaddr in an AD-system." Instead implementing all setting in the option specific set function as Nik suggested. ==================== Signed-off-by:
-
Andy Gospodarek authored
Adds netlink support for the following bonding options: * BOND_OPT_AD_ACTOR_SYS_PRIO * BOND_OPT_AD_ACTOR_SYSTEM * BOND_OPT_AD_USER_PORT_KEY When setting the actor system mac address we assume the netlink message contains a binary mac and not a string representation of a mac. Signed-off-by:
Jonathan Toppins <jtoppins@cumulusnetworks.com> Signed-off-by:
Nikolay Aleksandrov <razor@blackwall.org> Signed-off-by:
-
Mahesh Bandewar authored
The port key has three components - user-key, speed-part, and duplex-part. The LSBit is for the duplex-part, next 5 bits are for the speed while the remaining 10 bits are the user defined key bits. Get these 10 bits from the user-space (through the SysFs interface) and use it to form the admin port-key. Allowed range for the user-key is 0 - 1023 (10 bits). If it is not provided then use zero for the user-key-bits (default). It can set using following example code - # modprobe bonding mode=4 # usr_port_key=$(( RANDOM & 0x3FF )) # echo $usr_port_key > /sys/class/net/bond0/bonding/ad_user_port_key # echo +eth1 > /sys/class/net/bond0/bonding/slaves ... # ip link set bond0 up Signed-off-by:
Mahesh Bandewar <maheshb@google.com> Reviewed-by:
Nikolay Aleksandrov <nikolay@redhat.com> [jt: * fixed up style issues reported by checkpatch * fixed up context from change in ad_actor_sys_prio patch] Signed-off-by:
-
Mahesh Bandewar authored
In an AD system, the communication between actor and partner is the business between these two entities. In the current setup anyone on the same L2 can "guess" the LACPDU contents and then possibly send the spoofed LACPDUs and trick the partner causing connectivity issues for the AD system. This patch allows to use a random mac-address obscuring it's identity making it harder for someone in the L2 is do the same thing. This patch allows user-space to choose the mac-address for the AD-system. This mac-address can not be NULL or a Multicast. If the mac-address is set from user-space; kernel will honor it and will not overwrite it. In the absence (value from user space); the logic will default to using the masters' mac as the mac-address for the AD-system. It can be set using example code below - # modprobe bonding mode=4 # sys_mac_addr=$(printf '%02x:%02x:%02x:%02x:%02x:%02x' \ $(( (RANDOM & 0xFE) | 0x02 )) \ $(( RANDOM & 0xFF )) \ $(( RANDOM & 0xFF )) \ $(( RANDOM & 0xFF )) \ $(( RANDOM & 0xFF )) \ $(( RANDOM & 0xFF ))) # echo $sys_mac_addr > /sys/class/net/bond0/bonding/ad_actor_system # echo +eth1 > /sys/class/net/bond0/bonding/slaves ... # ip link set bond0 up Signed-off-by: -
Mahesh Bandewar authored
This patch allows user to randomize the system-priority in an ad-system. The allowed range is 1 - 0xFFFF while default value is 0xFFFF. If user does not specify this value, the system defaults to 0xFFFF, which is what it was before this patch. Following example code could set the value - # modprobe bonding mode=4 # sys_prio=$(( 1 + RANDOM + RANDOM )) # echo $sys_prio > /sys/class/net/bond0/bonding/ad_actor_sys_prio # echo +eth1 > /sys/class/net/bond0/bonding/slaves ... # ip link set bond0 up Signed-off-by: -
David S. Miller authored
Eric W. Biederman says: ==================== Cleanup the kernel sockets. Right now the situtation for allocating kernel sockets is a mess. - sock_create_kern does not take a namespace parameter. - kernel sockets must not reference count a network namespace and keep it alive or else we will have a reference counting loop. - The way we avoid the reference counting loop with sk_change_net and sk_release_kernel are major hacks. This patchset addresses this mess by fixing sock_create_kern to do everything necessary to create a kernel socket. None of the current users of kernel sockets need the network namespace reference counted. Either kernel sockets are network namespace aware (and using the current hacks) or kernel sockets are limited to the initial network namespace in which case it does not matter. This patchset starts by addressing tun which should be using normal userspace sockets like macvtap. Then sock_create_kern is fixed to take a network namespace. Then the in kernel status of sockets are passed through to sk_alloc. Then sk_alloc is fixed to not reference count the network namespace of kernel sockets. Then the callers of sock_create_kern are fixed up to stop using hacks. Then netlink which uses it's own flavor of sock_create_kern is fixed. Finally the hacks that are sk_change_net and sk_release_kernel are removed. When it is all done the code is easier to follow, easier to use, easier to maintain and shorter by about 70 lines. ==================== Reported-by:Ying Xue <ying.xue@windriver.com> Signed-off-by:
-
Eric W. Biederman authored
These functions are no longer needed and no longer used kill them. Signed-off-by:
"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
-
Eric W. Biederman authored
Utilize the new functionality of sk_alloc so that nothing needs to be done to suprress the reference counting on kernel sockets. Signed-off-by:
-
Eric W. Biederman authored
Now that sk_alloc knows when a kernel socket is being allocated modify it to not reference count the network namespace of kernel sockets. Keep track of if a socket needs reference counting by adding a flag to struct sock called sk_net_refcnt. Update all of the callers of sock_create_kern to stop using sk_change_net and sk_release_kernel as those hacks are no longer needed, to avoid reference counting a kernel socket. Signed-off-by:
-
Eric W. Biederman authored
In preparation for changing how struct net is refcounted on kernel sockets pass the knowledge that we are creating a kernel socket from sock_create_kern through to sk_alloc. Signed-off-by:
-
Eric W. Biederman authored
This is long overdue, and is part of cleaning up how we allocate kernel sockets that don't reference count struct net. Signed-off-by:
-
Eric W. Biederman authored
There is no need for tun to do the weird network namespace refcounting. The existing network namespace refcounting in tfile has almost exactly the same lifetime. So rewrite the code to use the struct sock network namespace refcounting and remove the unnecessary hand rolled network namespace refcounting and the unncesary tfile->net. This change allows the tun code to directly call sock_put bypassing sock_release and making SOCK_EXTERNALLY_ALLOCATED unnecessary. Remove the now unncessary tun_release so that if anything tries to use the sock_release code path the kernel will oops, and let us know about the bug. The macvtap code already uses it's internal socket this way. Signed-off-by:
-
- 10 May, 2015 1 commit
-
-
Eric Dumazet authored
For DCTCP or similar ECN based deployments on fabrics with shallow buffers, hosts are responsible for a good part of the buffering. This patch adds an optional ce_threshold to codel & fq_codel qdiscs, so that DCTCP can have feedback from queuing in the host. A DCTCP enabled egress port simply have a queue occupancy threshold above which ECT packets get CE mark. In codel language this translates to a sojourn time, so that one doesn't have to worry about bytes or bandwidth but delays. This makes the host an active participant in the health of the whole network. This also helps experimenting DCTCP in a setup without DCTCP compliant fabric. On following example, ce_threshold is set to 1ms, and we can see from 'ldelay xxx us' that TCP is not trying to go around the 5ms codel target. Queue has more capacity to absorb inelastic bursts (say from UDP traffic), as queues are maintained to an optimal level. lpaa23:~# ./tc -s -d qd sh dev eth1 qdisc mq 1: dev eth1 root Sent 87910654696 bytes 58065331 pkt (dropped 0, overlimits 0 requeues 42961) backlog 3108242b 364p requeues 42961 qdisc codel 8063: dev eth1 parent 1:1 limit 1000p target 5.0ms ce_threshold 1.0ms interval 100.0ms Sent 7363778701 bytes 4863809 pkt (dropped 0, overlimits 0 requeues 5503) rate 2348Mbit 193919pps backlog 255866b 46p requeues 5503 count 0 lastcount 0 ldelay 1.0ms drop_next 0us maxpacket 68130 ecn_mark 0 drop_overlimit 0 ce_mark 72384 qdisc codel 8064: dev eth1 parent 1:2 limit 1000p target 5.0ms ce_threshold 1.0ms interval 100.0ms Sent 7636486190 bytes 5043942 pkt (dropped 0, overlimits 0 requeues 5186) rate 2319Mbit 191538pps backlog 207418b 64p requeues 5186 count 0 lastcount 0 ldelay 694us drop_next 0us maxpacket 68130 ecn_mark 0 drop_overlimit 0 ce_mark 69873 qdisc codel 8065: dev eth1 parent 1:3 limit 1000p target 5.0ms ce_threshold 1.0ms interval 100.0ms Sent 11569360142 bytes 7641602 pkt (dropped 0, overlimits 0 requeues 5554) rate 3041Mbit 251096pps backlog 210446b 59p requeues 5554 count 0 lastcount 0 ldelay 889us drop_next 0us maxpacket 68130 ecn_mark 0 drop_overlimit 0 ce_mark 37780 ... Signed-off-by:
Neal Cardwell <ncardwell@google.com> Signed-off-by:
-
