aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/.gitignore (unfollow)
Commit message (Collapse)AuthorFilesLines
2018-06-18mpmc_ptr_ring: move atomic_read outside loopThomas Gschwantner1-2/+2
2018-06-18selftest/mpmc_ring: use rwsem instead of multiple completionsJason A. Donenfeld1-10/+10
2018-06-18selftest/mpmc_ring: switch back to our own ringJason A. Donenfeld1-9/+8
2018-06-18selftest/mpmc_ring: use completion and switch to ptr_ringJason A. Donenfeld1-17/+23
2018-06-18mpmc_ptr_ring: Fix a wordJonathan Neuschäfer1-1/+1
Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
2018-06-18selftest/mpmc_ring: use kthreadsJason A. Donenfeld1-40/+26
2018-06-18selftest/mpmc_ring: don't starve other queuesJason A. Donenfeld1-21/+28
2018-06-18selftest/mpmc_ring: round-robin CPUsJason A. Donenfeld1-2/+5
2018-06-18selftest/mpmc_ring: small restructuringJason A. Donenfeld1-18/+17
2018-06-18Test commitThomas Gschwantner4-17/+56
2018-06-18selftest/mpmc_ring: use int64_t for countersThomas Gschwantner1-6/+6
Previously used long would overflow on 32-bit machines.
2018-06-18selftest/mpmc_ring: Add more __init/__initdata annotationsJonathan Neuschäfer1-3/+3
The main function, mpmc_ring_selftest, is already marked with __init, but the callback functions and the ring pointer were not.
2018-06-18selftest/mpmc_ring: Remove const qualifier from mpmc_ptr_ring_produce argumentJonathan Neuschäfer1-1/+1
mpmc_ptr_ring_produce takes a void*. This fixes the following warning: net/wireguard/selftest/mpmc_ring.h: In function ‘producer_function’: net/wireguard/selftest/mpmc_ring.h:43:38: warning: passing argument 2 of ‘mpmc_ptr_ring_produce’ disc] while (mpmc_ptr_ring_produce(ring, (const void *) count)) ^
2018-06-18mpmc_ptr_ring: Include all necessary headersJonathan Neuschäfer1-0/+11
mpmc_ptr_ring.h should include the headers that it needs explicitly. This commit adds the following: - <asm/barrier.h>: smp_rmb, smp_wmb, smp_mb__before_atomic - <linux/atomic.h>: atomic_t, atomic_read, atomic_set, atomic_cmpxchg, atomic_inc - <linux/cache.h>: ____cacheline_aligned_in_smp - <linux/compiler.h>: READ_ONCE, WRITE_ONCE - <linux/errno.h>: ENOMEM - <linux/log2.h>: is_power_of_2 - <linux/processor.h>: cpu_relax - <linux/slab.h>: kcalloc, kfree - <linux/stddef.h>: NULL I'm not sure if all of them are really needed because I wasn't about to provoke a compile error due to the missing includes. Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
2018-06-18mpmc_ptr_ring: Fix some style detailsJonathan Neuschäfer2-1/+2
2018-06-18mpmc_ptr_ring: add selftestThomas Gschwantner4-1/+125
2018-06-18mpmc_ptr_ring: calculate mask once and store itThomas Gschwantner1-3/+5
2018-06-18mpmc_ptr_ring: use unsigned int instead of size_tThomas Gschwantner1-10/+10
For producers/consumers we use atomic_t which is really int, so size_t could either be too large, wasting memory, or too small (unlikely). For size, we also want to be using unsigned int, since the mask that we derive from it is ANDed with producer/consumer.
2018-06-18mpmc_ptr_ring: add include guardThomas Gschwantner1-0/+6
2018-06-18mpmc_ptr_ring: Switch to smp_[rw]mb()Jonathan Neuschäfer1-8/+8
From https://www.kernel.org/doc/Documentation/memory-barriers.txt: > SMP memory barriers are reduced to compiler barriers on uniprocessor > compiled systems because it is assumed that a CPU will appear to be > self-consistent, and will order overlapping accesses correctly with > respect to itself. Since we only order CPU memory accesses with the memory barriers in mpmc_ptr_ring.h, smp_[rw]mb() should be sufficient. Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
2018-06-18messages: Ensure that there are more queue slots than CPUsJonathan Neuschäfer2-1/+9
I'm not completely sure about this. It also doesn't fix all the errors: sometimes the test suite reports that it fails to send packets.
2018-06-18mpmc_ptr_ring: Place producer_head and producer_tail in the same cachelineJonathan Neuschäfer1-5/+3
They are updated together, so it doesn't make much sense to keep them separate in the cache.
2018-06-18mpmc_ptr_ring: Use atomic_t instead of atomic_long_tJonathan Neuschäfer1-20/+20
The index range doesn't get large enough to warrant 64-bit indices.
2018-06-18mpmc_ptr_ring: Eliminate false sharing in struct mpmc_ptr_ringJonathan Neuschäfer1-3/+3
2018-06-18mpmc_ptr_ring: Reduce the memory barrier usageJonathan Neuschäfer1-26/+22
2018-06-18mpmc_ptr_ring: Fix the remaining crashJonathan Neuschäfer1-2/+2
2018-06-18[WIP] Implement a lock-free MPMC ring bufferJonathan Neuschäfer6-13/+225
TODO: actually use the right memory barriers in the right places TODO: eliminate false sharing between mpmc_ptr_ring members TODO: reconsider atomic_long_t vs. atomic_t, vs. the type of size in _init() TODO: sprinkle likely/unlikely on some branches FIXME: it still crashes
2018-06-17qemu: bump default kernelJason A. Donenfeld1-1/+1
2018-06-17wg-quick: android: don't forget to free compiled regexesJason A. Donenfeld1-5/+6
2018-06-17wg-quick: android: disable roaming to v6 networks when v4 is specifiedJason A. Donenfeld1-1/+54
This works around an unfortunate bug in 464XLAT transitions.
2018-06-17dns-hatchet: apply resolv.conf's selinux context to new resolv.confJason A. Donenfeld1-0/+2
2018-06-17simd: no need to restore fpu state when no preemptionJason A. Donenfeld1-0/+2
2018-06-17simd: encapsulate fpu amortization into nice functionsJason A. Donenfeld7-71/+83
2018-06-16queueing: re-enable preemption periodically to lower latencyJason A. Donenfeld2-0/+18
2018-06-16queueing: remove useless spinlocks on scJason A. Donenfeld3-5/+1
Since these are the only consumers, there's no need for locking.
2018-06-14tools: getentropy requires 10.12Jason A. Donenfeld1-1/+7
2018-06-14chacha20poly1305: use slow crypto on -rt kernels on arm tooJason A. Donenfeld1-1/+1
2018-06-13version: bump snapshot0.0.20180613Jason A. Donenfeld2-2/+2
2018-06-13chacha20poly1305: use slow crypto on -rt kernelsJason A. Donenfeld1-1/+1
In rt kernels, spinlocks call schedule(), which means preemption can't be disabled. The FPU disables preemption. Hence, we can either restructure things to move the calls to kernel_fpu_begin/end to be really close to the actual crypto routines, or we can do the slower lazier solution of just not using the FPU at all on -rt kernels. This patch goes with the latter lazy solution. The reason why we don't place the calls to kernel_fpu_begin/end close to the crypto routines in the first place is that they're very expensive, as it usually involves a call to XSAVE. So on sane kernels, we benefit from only having to call it once.
2018-06-08tools: support getentropy(3)Jason A. Donenfeld1-0/+11
2018-06-06tools: encoding: add missing static array constraintsJason A. Donenfeld3-5/+5
2018-06-04wg-quick: android: change name of intentJason A. Donenfeld1-1/+1
2018-06-02chacha20: add missing include to headerJason A. Donenfeld1-0/+1
2018-05-31wg-quick: android: delay setting users until endJason A. Donenfeld1-1/+6
`ndc users add` eventually invokes SOCK_DESTROY on user sockets, causing them to reconnect. By delaying this until after routes are set, we ensure that the sockets reconnect using the tunnel, rather than the old route.
2018-05-31version: bump snapshot0.0.20180531Jason A. Donenfeld2-2/+2
2018-05-31qemu: bump default versionJason A. Donenfeld1-1/+1
2018-05-31tools: constanter time encodingJason A. Donenfeld2-22/+28
2018-05-31device: do not assume dst is always validJason A. Donenfeld1-1/+1
The new flow offloading feature at the moment does not set the dst. We have a patch pending to fix this upstream, but in the meantime, work around it here.
2018-05-31poly1305: mips: compute S on flyRené van Dorst1-31/+22
This reduces memory access and the total opaque size. Signed-off-by: René van Dorst <opensource@vdorst.com>
2018-05-31chacha20poly1305: test for authtag failureJason A. Donenfeld1-3/+21
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.