- 08 Apr, 2016 40 commits
-
-
Vivien Didelot authored
The switchdev design implies that a software error should not happen in the commit phase since it must have been previously reported in the prepare phase. If an hardware error occurs during the commit phase, there is nothing switchdev can do about it. The DSA layer separates port_vlan_prepare and port_vlan_add for simplicity and convenience. If an hardware error occurs during the commit phase, there is no need to report it outside the driver itself. Make the DSA port_vlan_add routine return void for explicitness. Signed-off-by:
Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Vivien Didelot authored
The switchdev design implies that a software error should not happen in the commit phase since it must have been previously reported in the prepare phase. If an hardware error occurs during the commit phase, there is nothing switchdev can do about it. The DSA layer separates port_fdb_prepare and port_fdb_add for simplicity and convenience. If an hardware error occurs during the commit phase, there is no need to report it outside the DSA driver itself. Make the DSA port_fdb_add routine return void for explicitness. Signed-off-by:
-
Vivien Didelot authored
The DSA layer doesn't care about the return code of the port_stp_update routine, so make it void in the layer and the DSA drivers. Replace the useless dsa_slave_stp_update function with a dsa_slave_stp_state function used to reply to the switchdev SWITCHDEV_ATTR_ID_PORT_STP_STATE attribute. In the meantime, rename port_stp_update to port_stp_state_set to explicit the state change. Signed-off-by:
-
Vivien Didelot authored
Add description for the missing port_vlan_prepare, port_fdb_prepare, port_fdb_dump functions in the DSA documentation. Signed-off-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
David S. Miller authored
Merge tag 'mac80211-next-for-davem-2016-04-06' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next Johannes Berg says: ==================== For the 4.7 cycle, we have a number of changes: * Bob's mesh mode rhashtable conversion, this includes the rhashtable API change for allocation flags * BSSID scan, connect() command reassoc support (Jouni) * fast (optimised data only) and support for RSS in mac80211 (myself) * various smaller changes ==================== Signed-off-by:
-
Daniel Borkmann authored
The verifier needs to go through every path of the program in order to check that it terminates safely, which can be quite a lot of instructions that need to be processed f.e. in cases with more branchy programs. With search pruning from f1bca824 ("bpf: add search pruning optimization to verifier") the search space can already be reduced significantly when the verifier detects that a previously walked path with same register and stack contents terminated already (see verifier's states_equal()), so the search can skip walking those states. When working with larger programs of > ~2000 (out of max 4096) insns, we found that the current limit of 32k instructions is easily hit. For example, a case we ran into is that the search space cannot be pruned due to branches at the beginning of the program that make use of certain stack space slots (STACK_MISC), which are never used in the remaining program (STACK_INVALID). Therefore, the verifier needs to walk paths for the slots in STACK_INVALID state, but also all remaining paths with a stack structure, where the slots are in STACK_MISC, which can nearly double the search space needed. After various experiments, we find that a limit of 64k processed insns is a more reasonable choice when dealing with larger programs in practice. This still allows to reject extreme crafted cases that can have a much higher complexity (f.e. > ~300k) within the 4096 insns limit due to search pruning not being able to take effect. Furthermore, we found that a lot of states can be pruned after a call instruction, f.e. we were able to reduce the search state by ~35% in some cases with this heuristic, trade-off is to keep a bit more states in env->explored_states. Usually, call instructions have a number of preceding register assignments and/or stack stores, where search pruning has a better chance to suceed in states_equal() test. The current code marks the branch targets with STATE_LIST_MARK in case of conditional jumps, and the next (t + 1) instruction in case of unconditional jump so that f.e. a backjump will walk it. We also did experiments with using t + insns[t].off + 1 as a marker in the unconditionally jump case instead of t + 1 with the rationale that these two branches of execution that converge after the label might have more potential of pruning. We found that it was a bit better, but not necessarily significantly better than the current state, perhaps also due to clang not generating back jumps often. Hence, we left that as is for now. Signed-off-by:
Daniel Borkmann <daniel@iogearbox.net> Acked-by:
Alexei Starovoitov <ast@kernel.org> Signed-off-by:
-
Jiri Pirko authored
Ptr to devlink structure can be easily obtained from devlink_port->devlink. So share user_ptr[0] pointer for both and leave user_ptr[1] free for other users. Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Reviewed-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
-
David S. Miller authored
Jiri Pirko says: ==================== mlxsw: small driver update + one tiny devlink dependency Cosmetics, in preparation to sharedbuffer patchset. First patch is here to allow patch number two. ==================== Signed-off-by: -
Jiri Pirko authored
Fix copy&paste error and state the name of SBPM register correctly. Signed-off-by:
-
Jiri Pirko authored
Same field, same values, so share the same enum. Signed-off-by:
-
Jiri Pirko authored
Instead of that, pass mlxsw_core and use a helper to get driver priv from driver code. Looks much cleaner that way. Signed-off-by:
-
Jiri Pirko authored
Instead of passing around driver priv, pass struct mlxsw_core * directly. Signed-off-by:
-
Jiri Pirko authored
Remove devlink port reg/unreg from spectrum and switchx2 code and rather do the common work in core. That also ensures code separation where devlink is only used in core.c. Signed-off-by:
-
Jiri Pirko authored
As we rely on caller zeroing or correctly set the struct before the call, this implicit type set is either no-op (DEVLINK_PORT_TYPE_NOTSET is 0) or it rewrites wanted value. So remove this. Signed-off-by:
-
David S. Miller authored
Jakub Kicinski says: ==================== MTU/buffer reconfig changes I re-discussed MPLS/MTU internally, dropped it from the patch 1, re-tested everything, found out I forgot about debugfs pointers, fixed that as well. v5: - don't reserve space in RX buffers for MPLS label stack (patch 1); - fix debugfs pointers to ring structures (patch 5). v4: - cut down on unrelated patches; - don't "close" the device on error path. --- v4 cover letter Previous series included some not entirely related patches, this one is cut down. Main issue I'm trying to solve here is that .ndo_change_mtu() in nfpvf driver is doing full close/open to reallocate buffers - which if open fails can result in device being basically closed even though the interface is started. As suggested by you I try to move towards a paradigm where the resources are allocated first and the MTU change is only done once I'm certain (almost) nothing can fail. Almost because I need to communicate with FW and that can always time out. Patch 1 fixes small issue. Next 10 patches reorganize things so that I can easily allocate new rings and sets of buffers while the device is running. Patches 13 and 15 reshape the .ndo_change_mtu() and ethtool's ring-resize operation into desired form. ==================== Signed-off-by: -
Jakub Kicinski authored
Since much of the required changes have already been made for changing MTU at runtime let's use it for ring size changes as well. Signed-off-by:
Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by:
-
Jakub Kicinski authored
Soon ring resize will call this functions with values different than the current configuration we need to explicitly pass the ring count as parameter. Signed-off-by:
-
Jakub Kicinski authored
When changing MTU on running device first allocate new rings and buffers and once it succeeds proceed with changing MTU. Allocation of new rings is not really necessary for this operation - it's done to keep the code simple and because size of the extra ring memory is quite small compared to the size of buffers. Operation can still fail midway through if FW communication times out. In that case we retry with old MTU (rings). Signed-off-by:
-
Jakub Kicinski authored
Free list buffer size needs to be propagated to few functions as a parameter and added to struct nfp_net_rx_ring since soon some of the functions will be reused to manage rings with buffers of size different than nn->fl_bufsz. Signed-off-by:
-
Jakub Kicinski authored
FW reconfiguration in .ndo_open()/.ndo_stop() should reset/ restore queue state. Since we need IRQs to be disabled when filling rings on RX path we have to move disable_irq() from .ndo_open() all the way up to IRQ allocation. nfp_net_start_vec() becomes trivial now so it's inlined. Signed-off-by:
-
Jakub Kicinski authored
Divide .ndo_open() and .ndo_stop() into logical, callable chunks. No functional changes. Signed-off-by:
-
Jakub Kicinski authored
nfp_net_[rt]x_ring_{alloc,free} should only allocate or free ring resources without touching the device. Move setting parameters in the BAR to separate functions. This will make it possible to reuse alloc/free functions to allocate new rings while the device is running. Signed-off-by: -
Jakub Kicinski authored
We want the .ndo_open() to have following structure: - allocate resources; - configure HW/FW; - enable the device from stack perspective. Therefore filling RX rings needs to be moved to the beginning of .ndo_open(). Signed-off-by:
-
Jakub Kicinski authored
Separate allocation of buffers from giving them to FW, thanks to this it will be possible to move allocation earlier on .ndo_open() path and reuse buffers during runtime reconfiguration. Similar to TX side clean up the spill of functionality from flush to freeing the ring. Unlike on TX side, RX ring reset does not free buffers from the ring. Ring reset means only that FW pointers are zeroed and buffers on the ring must be placed in [0, cnt - 1) positions. Signed-off-by:
-
Jakub Kicinski authored
Since we never used flush without freeing the ring later the functionality of the two operations is mixed. Rename flush to ring reset and move there all the things which have to be done after FW ring state is cleared. While at it do some clean-ups. Signed-off-by:
-
Jakub Kicinski authored
To be able to switch rings more easily on config changes allocate them dynamically, separately from nfp_net structure. Signed-off-by:
-
Jakub Kicinski authored
nfp_net_[rt]x_ring_init functions used to be called from probe path only and some of their functionality was spilled to the call site. In order to reuse them for ring reconfiguration we need them to do all the init. Signed-off-by:
-
Jakub Kicinski authored
nfp_net_{alloc|free}_rings contained strange mix of allocations and vector initialization. Remove it, declare vector init as a separate function and handle allocations explicitly. Signed-off-by: -
Jakub Kicinski authored
We need to be able to disable the link state interrupt when the device is brought down. We used to just free the IRQ at the beginning of .ndo_stop(). As we now move towards more ordered .ndo_open()/.ndo_stop() paths LSC allocation should be placed in the "allocate resource" section. Since the IRQ can't be freed early in .ndo_stop(), it is disabled instead. Signed-off-by:
-
Jakub Kicinski authored
When calculating the RX buffer length we need to account for up to 2 VLAN tags. Rounding up to 1k is an relic of a distant past and can be removed. While at it also remove trivial print statement. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 10GbE Intel Wired LAN Driver Updates 2016-04-07 This series contains updates to ixgbe and ixgbevf. This entire series (except for one patch from Alex) comes from Mark and is mainly to add support for our new MAC (x550em_a). So let's get Alex's patch out of the way first before we cover Mark's many changes. Alex does his enable bulk free in transmit cleanup for ixgbe and ixgbevf, like his has done for all of our other drivers. First Mark cleans up registers that were not being used, so do some house cleaning. Then to avoid casting lan_id and func fields, just make them u8 since they only hold small values anyways. Found and fixed an issue where on read operations it could be possible to modify locations beyond the length passed in, so change the check to round up in the same way. Cleaned up the interface for issuing firmware commands to use a void * instead of a u32 * which eliminates a number of casts. Added support for the new MAC and provided method pointers and use them to access IOSF-attached devices, since the new MAC will also need a new access method. Added support for SFPs with an external retimer and for an SGMII backplane interface. ==================== Signed-off-by: -
David S. Miller authored
Alexei Starovoitov says: ==================== allow bpf attach to tracepoints Hi Steven, Peter, v1->v2: addressed Peter's comments: - fixed wording in patch 1, added ack - refactored 2nd patch into 3: 2/10 remove unused __perf_addr macro which frees up an argument in perf_trace_buf_submit 3/10 split perf_trace_buf_prepare into alloc and update parts, so that bpf programs don't have to pay performance penalty for update of struct trace_entry which is not going to be accessed by bpf 4/10 actual addition of bpf filter to perf tracepoint handler is now trivial and bpf prog can be used as proper filter of tracepoints v1 cover: last time we discussed bpf+tracepoints it was a year ago [1] and the reason we didn't proceed with that approach was that bpf would make arguments arg1, arg2 to trace_xx(arg1, arg2) call to be exposed to bpf program and that was considered unnecessary extension of abi. Back then I wanted to avoid the cost of buffer alloc and field assign part in all of the tracepoints, but looks like when optimized the cost is acceptable. So this new apporach doesn't expose any new abi to bpf program. The program is looking at tracepoint fields after they were copied by perf_trace_xx() and described in /sys/kernel/debug/tracing/events/xxx/format We made a tool [2] that takes arguments from /sys/.../format and works as: $ tplist.py -v random:urandom_read int got_bits; int pool_left; int input_left; Then these fields can be copy-pasted into bpf program like: struct urandom_read { __u64 hidden_pad; int got_bits; int pool_left; int input_left; }; and the program can use it: SEC("tracepoint/random/urandom_read") int bpf_prog(struct urandom_read *ctx) { return ctx->pool_left > 0 ? 1 : 0; } This way the program can access tracepoint fields faster than equivalent bpf+kprobe program, which is the main goal of these patches. Patch 1-4 are simple changes in perf core side, please review. I'd like to take the whole set via net-next tree, since the rest of the patches might conflict with other bpf work going on in net-next and we want to avoid cross-tree merge conflicts. Alternatively we can put patches 1-4 into both tip and net-next. Patch 9 is an example of access to tracepoint fields from bpf prog. Patch 10 is a micro benchmark for bpf+kprobe vs bpf+tracepoint. Note that for actual tracing tools the user doesn't need to run tplist.py and copy-paste fields manually. The tools do it automatically. Like argdist tool [3] can be used as: $ argdist -H 't:block:block_rq_complete():u32:nr_sector' where 'nr_sector' is name of tracepoint field taken from /sys/kernel/debug/tracing/events/block/block_rq_complete/format and appropriate bpf program is generated on the fly. [1] http://thread.gmane.org/gmane.linux.kernel.api/8127/focus=8165 [2] https://github.com/iovisor/bcc/blob/master/tools/tplist.py [3] https://github.com/iovisor/bcc/blob/master/tools/argdist.py ==================== Signed-off-by: -
Alexei Starovoitov authored
the first microbenchmark does fd=open("/proc/self/comm"); for() { write(fd, "test"); } and on 4 cpus in parallel: writes per sec base (no tracepoints, no kprobes) 930k with kprobe at __set_task_comm() 420k with tracepoint at task:task_rename 730k For kprobe + full bpf program manully fetches oldcomm, newcomm via bpf_probe_read. For tracepint bpf program does nothing, since arguments are copied by tracepoint. 2nd microbenchmark does: fd=open("/dev/urandom"); for() { read(fd, buf); } and on 4 cpus in parallel: reads per sec base (no tracepoints, no kprobes) 300k with kprobe at urandom_read() 279k with tracepoint at random:urandom_read 290k bpf progs attached to kprobe and tracepoint are noop. Signed-off-by: -
Alexei Starovoitov authored
modify offwaketime to work with sched/sched_switch tracepoint instead of kprobe into finish_task_switch Signed-off-by:
-
Alexei Starovoitov authored
Recognize "tracepoint/" section name prefix and attach the program to that tracepoint. Signed-off-by:
-
Alexei Starovoitov authored
during bpf program loading remember the last byte of ctx access and at the time of attaching the program to tracepoint check that the program doesn't access bytes beyond defined in tracepoint fields This also disallows access to __dynamic_array fields, but can be relaxed in the future. Signed-off-by:
-
Alexei Starovoitov authored
needs two wrapper functions to fetch 'struct pt_regs *' to convert tracepoint bpf context into kprobe bpf context to reuse existing helper functions Signed-off-by:
-
Alexei Starovoitov authored
register tracepoint bpf program type and let it call the same set of helper functions as BPF_PROG_TYPE_KPROBE Signed-off-by:
-
Alexei Starovoitov authored
introduce BPF_PROG_TYPE_TRACEPOINT program type and allow it to be attached to the perf tracepoint handler, which will copy the arguments into the per-cpu buffer and pass it to the bpf program as its first argument. The layout of the fields can be discovered by doing 'cat /sys/kernel/debug/tracing/events/sched/sched_switch/format' prior to the compilation of the program with exception that first 8 bytes are reserved and not accessible to the program. This area is used to store the pointer to 'struct pt_regs' which some of the bpf helpers will use: +---------+ | 8 bytes | hidden 'struct pt_regs *' (inaccessible to bpf program) +---------+ | N bytes | static tracepoint fields defined in tracepoint/format (bpf readonly) +---------+ | dynamic | __dynamic_array bytes of tracepoint (inaccessible to bpf yet) +---------+ Not that all of the fields are already dumped to user space via perf ring buffer and broken application access it directly without consulting tracepoint/format. Same rule applies here: static tracepoint fields should only be accessed in a format defined in tracepoint/format. The order of fields and field sizes are not an ABI. Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by:
-
Alexei Starovoitov authored
split allows to move expensive update of 'struct trace_entry' to later phase. Repurpose unused 1st argument of perf_tp_event() to indicate event type. While splitting use temp variable 'rctx' instead of '*rctx' to avoid unnecessary loads done by the compiler due to -fno-strict-aliasing Signed-off-by:
-
