| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | remove warning about amdgpu userptr ioctl being unimplemented |   jsg | 2021-02-20 | 1 | -1/+0 |
| | | | | | | matches radeon and i915 reported by Benjamin Baier | ||||
| * | move the rearming of the cq after the refill of the rq. |   dlg | 2021-02-15 | 1 | -4/+5 |
| | | | | | | | | | this is the only real diff we have left outstanding on a box that experienced rx lockups. since adding this change it's been happy for the last 4 weeks and counting so far. ok jmatthew@ | ||||
| * | regen |   kettenis | 2021-02-14 | 2 | -2/+12 |
| | | |||||
| * | Add a few more devices that show up on the Apple M1 mini. | kettenis | 2021-02-14 | 1 | -1/+3 |
| | | | | | ok patrick@, deraadt@ | ||||
| * | correct drm work flush behaviour | jsg | 2021-02-14 | 2 | -6/+10 |
| | | | | | | | | | | Don't set taskq to system_wq in INIT_WORK(). Test if taskq pointer is non-NULL before calling taskq_barrier() in flush functions. fixes a black screen on boot problem with 5.10.y drm using nano x1 bisected by jcs@ to 'drm/i915: Always flush the active worker before returning from the wait' | ||||
| * | Add missing break in switch statement of rge_activate(). |   stsp | 2021-02-11 | 1 | -1/+2 |
| | | | | | | | | CID 1501716 ok kevlo@ and mestre@ had the same diff | ||||
| * | Simplify sleep_setup API to two operations in preparation for splitting |   mpi | 2021-02-08 | 2 | -9/+7 |
| | | | | | | | | | | | | | the SCHED_LOCK(). Putting a thread on a sleep queue is reduce to the following: sleep_setup(); /* check condition or release lock */ sleep_finish(); Previous version ok cheloha@, jmatthew@, ok claudio@ | ||||
| * | Add basic support for BCM4378 as found on the Apple M1 SoCs. There's a |   patrick | 2021-01-31 | 1 | -1/+4 |
| | | | | | little bit more to do though before it can be enabled. | ||||
| * | regen | patrick | 2021-01-31 | 2 | -2/+7 |
| | | |||||
| * | Add Broadcom BCM4378. | patrick | 2021-01-31 | 1 | -1/+2 |
| | | |||||
| * | regen | jsg | 2021-01-27 | 2 | -2/+177 |
| | | |||||
| * | add some more amdgpu devices | jsg | 2021-01-27 | 1 | -1/+36 |
| | | |||||
| * | do better accounting of how many msix interrupts we want to use. | dlg | 2021-01-27 | 1 | -2/+8 |
| | | | | | ok jmatthew@ | ||||
| * | intrmap attribute missing from mcx; ok dlg |   deraadt | 2021-01-26 | 1 | -3/+3 |
| | | |||||
| * | raise the max number of queues/interrupts to 16, up from 1. | dlg | 2021-01-25 | 1 | -3/+12 |
| | | | | | | | | | | jmatthew@ has tried this before, but hrvoje popovski experienced breakage so it wasn't enabled. we've tightened the code up since then so it's time to try again. this diff has been tested by hrvoje popovski and myself ok jmatthew@ | ||||
| * | if the rx descriptor reports the rss hash, use it for the mbuf flowid. | dlg | 2021-01-25 | 1 | -2/+10 |
| | | | | | ok jmatthew@ | ||||
| * | don't lose the M_FLOWID flag if the ipv4 cksum is ok. | dlg | 2021-01-25 | 1 | -2/+2 |
| | | | | | | found while poking around with hrvoje popovski yes jmatthew@ | ||||
| * | use an intrmap when establishing interrupts for queues. | dlg | 2021-01-25 | 1 | -21/+44 |
| | | | | | | | | mcx is still hardcoded/limited to 1 queue for now, but this lets different mcx devices use different cpus for handling packets. looks good jmatthew@ | ||||
| * | match on Intel Alder Lake and Meteor Lake I219 Ethernet ids | jsg | 2021-01-24 | 3 | -3/+30 |
| | | |||||
| * | regen | jsg | 2021-01-24 | 2 | -2/+47 |
| | | |||||
| * | add Intel Alder Lake and Meteor Lake I219 Ethernet ids | jsg | 2021-01-24 | 1 | -1/+10 |
| | | |||||
| * | match on Realtek RTL8168H ids for Killer E2500V2 and E2600 | jsg | 2021-01-24 | 1 | -1/+3 |
| | | | | | checked against linux and windows drivers | ||||
| * | regen | jsg | 2021-01-24 | 2 | -2/+12 |
| | | |||||
| * | add Realtek RTL8168H ids for Killer E2500V2 and E2600 | jsg | 2021-01-24 | 1 | -1/+3 |
| | | | | | checked against linux and windows drivers | ||||
| * | match on another Realtek RTL8168 id | jsg | 2021-01-23 | 1 | -1/+2 |
| | | | | | reported and tested by John Batteen on a TP-Link TG-3468 | ||||
| * | regen | jsg | 2021-01-23 | 2 | -2/+7 |
| | | |||||
| * | add another Realtek RTL8168 id | jsg | 2021-01-23 | 1 | -1/+2 |
| | | | | | shows up on a TP-Link TG-3468 John Batteen has | ||||
| * | Check management capabilities before trying to attach temperature sensors, |   jmatthew | 2021-01-20 | 1 | -1/+43 |
| | | | | | | | | | avoiding an unhelpful error message if the card's firmware doesn't expose the sensor registers. tested by chris@, who saw the unhelpful error message ok dlg@ | ||||
| * | add another iwx device |   jcs | 2021-01-17 | 2 | -11/+21 |
| | | | | | ok stsp jsg | ||||
| * | regen | jcs | 2021-01-17 | 2 | -2/+7 |
| | | |||||
| * | add another ax201 | jcs | 2021-01-17 | 1 | -1/+2 |
| | | |||||
| * | this hardware is fine with BUS_DMA_64BIT mappings. | dlg | 2021-01-17 | 1 | -4/+7 |
| | | | | | | | | | this raises performance of tcpbench on an m3000 from ~3kpps and ~8MB/s to ~70kpps and ~191MB/s when transmitting, and ~10kpps and ~15MB/s to ~120kpps and 174MB/s when receiving. i also tested this on a v245 and an m4000 a while back. | ||||
| * | recognise Realtek ALC287 | jsg | 2021-01-16 | 1 | -1/+4 |
| | | |||||
| * | regen | jsg | 2021-01-16 | 2 | -4/+109 |
| | | |||||
| * | add devices mostly from x1 nano jcs@ has | jsg | 2021-01-16 | 1 | -2/+23 |
| | | |||||
| * | revert drm vmalloc changes | jsg | 2021-01-13 | 3 | -126/+39 |
| | | | | | | | | It is suspected they were to blame for a machine with inteldrm running X (xterms and chromium) running out of resources after a few days. ok kettenis@ | ||||
| * | Bring the emulated Linux memory allocation interfaces more in line with | kettenis | 2021-01-08 | 3 | -39/+126 |
| | | | | | | | | | | what Linux does. Let vmalloc() use km_alloc(9) instead of malloc(9) and let kvmalloc() only use malloc(9) for small (less than a page) allocations and atomic allocations. This should reduce the pressure on the "interrupt-safe" map. ok jsg@ | ||||
| * | the tx doorbell is next to the rx doorbell, not on top of it. | dlg | 2021-01-04 | 1 | -2/+2 |
| | | |||||
| * | use bus_dmamap_sync around updates to the doorbells. | dlg | 2021-01-04 | 1 | -33/+69 |
| | | | | | ok jmatthew@ | ||||
| * | Remove useless redundant call of ttyopen(). |   jan | 2021-01-01 | 1 | -6/+1 |
| | | | | | | | Also remove dead DIALOUT macro. ok mpi@ | ||||
| * | don't oversleep when waiting on fences | jsg | 2020-12-31 | 1 | -15/+17 |
| | | | | | original diff from and ok cheloha@ | ||||
| * | regen | kettenis | 2020-12-28 | 2 | -2/+14 |
| | | |||||
| * | Add Synopsys vendor and their DesignWare PCIe bridge. | kettenis | 2020-12-28 | 1 | -1/+5 |
| | | |||||
| * | have mcx_process_txeof return the number of slots it processed. | dlg | 2020-12-27 | 1 | -7/+8 |
| | | | | | | | | it used a pointer in an argument to communicate that back to the caller, while being a void functon. this seems more natural and brings it in line with how the rx completion function returns free slots to its caller too. | ||||
| * | do a bus space barrier after arming the eq. | dlg | 2020-12-27 | 1 | -2/+2 |
| | | | | | ok jmatthew@ | ||||
| * | disable timestamping a little bit harder to avoid divide by 0. | dlg | 2020-12-27 | 1 | -1/+5 |
| | | | | | | hrvoje popovski reports the current code faults on some boxes. i'm working on it, but the code isn't being used right now. | ||||
| * | shuffle filling the rx ring so the sw prod is updated before the hw. | dlg | 2020-12-27 | 1 | -24/+19 |
| | | | | | ok jmatthew@ | ||||
| * | reuse the calculated vector as the argument to pci_intr_map_msix. | dlg | 2020-12-26 | 1 | -2/+2 |
| | | | | | doing the maths again feels error prone. | ||||
| * | add bus_dmamap_sync ops around the eq. | dlg | 2020-12-26 | 1 | -6/+27 |
| | | | | | ok jmatthew@ | ||||
| * | add some bus_dmamap_syncs around the rq. | dlg | 2020-12-26 | 1 | -2/+16 |
| | | | | | ok jmatthew@ | ||||
 |
index : wireguard-openbsd | |
| WireGuard implementation for the OpenBSD kernel | Matt Dunwoodie |
| summaryrefslogtreecommitdiffstats |