summaryrefslogtreecommitdiffstats
path: root/sys (follow)
Commit message (Collapse)AuthorAgeFilesLines
* Somes fixes for the VIA PadLock drivers.fcambus2017-12-242-5/+7
| | | | | | | | - Do not ignore the return value of m_copyback() on i386 - Always free sc->op_buf before returning on both amd64 and i386, pointed out by mikeb@, thanks! OK mikeb@
* There was a corner case where linking the inp to the state key didbluhm2017-12-241-3/+15
| | | | | | | | work in pf. The function pf_inp_link() takes the state key from the mbuf and not the one pf_find_state() has just found. Introduce a new function pf_state_key_link_inpcb() that links the given state key and inpcb together with some sanity checks. OK sashan@
* Allow SD/MMC controller drivers to allocate their own DMA mapping structurekettenis2017-12-242-3/+5
| | | | | | to account for any relevant hardware constraints. ok stsp@
* Generate SIGBUS/BUS_ADRALN when we see a SP Alignment Exception from userland.kettenis2017-12-241-1/+6
| | | | ok jsg@
* regenmlarkin2017-12-232-2/+12
|
* Add missing Sunrise Point-H I2C controller PCI IDsmlarkin2017-12-231-1/+3
|
* Remove stray space.kettenis2017-12-231-2/+2
|
* Sync Allwinner H3/H5 pin definitions with Linux.kettenis2017-12-231-3/+2
|
* Avoid null-pointer dereference wen a device tree refers to a pin-functionkettenis2017-12-231-1/+3
| | | | that doesn't exist.
* scrolling: take care of noise.bru2017-12-231-3/+5
|
* Remove dead assignment.fcambus2017-12-231-4/+1
| | | | | | Now that we call rasops_putchar_rotated(), we don't need ri anymore. OK kettenis@
* Use regmap interface to tweak "SYSREG" PHY configuration registers andkettenis2017-12-236-138/+18
| | | | | | | | remove exsysreg(4). The exsysreg(4) driver would no longer attach as syscon(4) would match its node in the device tree. As a result the driver corss-call would panic. Makes te Odroid-XU4 work again.
* enable snooping on Intel 200 Series HDAjsg2017-12-231-1/+2
|
* match on intel 200 seriesjsg2017-12-231-2/+3
|
* regenjsg2017-12-232-2/+247
|
* add ids for intel coffee lake-s and series 200/300 pchjsg2017-12-231-1/+50
|
* If we use the cookie-based interrupt API on sun4v, whe shouldn't put thekettenis2017-12-223-22/+22
| | | | | | | | | interrupts in the lookup table. On machines like the Oracle T7-2 this would make the code believe that there is interrupt sharing, but since we bypass intr_list_handler() when using the cookie-based interrupt API, we wouldn't properly acknowledge the interrupt. This fixes the last remaining problem with interrupts on the Oracle T7-2.
* 1. Use unfiltered deltas for scrolling. 2. (Re-)Simplify the edge areabru2017-12-221-74/+63
| | | | setup.
* Expand u_short and u_int to unsigned short and unsigned intmillert2017-12-211-4/+4
| | | | | | respectively to avoid compilation errors when one of the POSIX or X/OPEN version macros is defined. Also sync the field descriptions with kqueue.2. OK deraadt@
* Call if_start() directly to send the EAPOL key.mpi2017-12-211-3/+8
| | | | | | | | | | | | | | | | | | | | Using if_enqueue() here, from interrupt context, might result in the packet beeing enqueued, incorrectly encrypted, on the TX ring. This race has been recently exposed by the re-introduction of the TX mitigation. It exists because the net80211 stack sets IEEE80211_NODE_TXPROT on the node while processing the 3rd message, assuming the answer has already been transmitted. However a CPU returns from if_enqueue() it cannot assume that the send queue is empty. So call if_start() to flush this queue. Encrypting the 4th message of the 4way handshake with the new key breaks WPA handshake as found the hardway by anton@. Race analysed by dlg@, a lot of net80211 inputs and suggetions from stsp@. ok stsp@, dlg@
* Two width specifiers changed from %02 to %0 in previous. Change them back.tb2017-12-211-2/+2
| | | | ok deraadt
* Do not use %hx for chars, ok kettenisderaadt2017-12-211-3/+3
|
* drivers might want to call ifiq_barrier.dlg2017-12-211-1/+2
|
* add some ifq and ifiq barriers in ixl_down.dlg2017-12-211-4/+5
| | | | move rxfill in ixl_up so the qtail is only written once.
* now that we have multiple input queues in ifnet structs, use them.dlg2017-12-211-26/+24
| | | | | for now we still only have one set of tx and rx rings, but sets us up to bump that number.
* prototype if_attach_iqueues so drivers can configure multiple iqs.dlg2017-12-211-1/+2
|
* Two background scan fixes for iwn(4) (4965 devices only):stsp2017-12-201-6/+19
| | | | | | | | | | | | | | | | | | | 1) Use only one antenna while scanning on 5GHz. Fixes very low RSSI values. 2) During a background scan while associated to a 5GHz AP, send probe requests on 2GHz channels with an OFDM rate (6Mbps) because the buggy firmware does not like sending with a CCK rate (1Mbps) in this state. CCK rates are not valid for 5GHz, which could explain this firmware bug. Taken from FreeBSD r222679. jca@ tested and confirmed the first problem but lacks a 5GHz access point to associate to and test the second one with. I am going to *blindly trust* FreeBSD on the second one! It is not a big deal if it doesn't actually apply to our firmware version as the change itself is rather inconsequential in practice. And 4965 hardware is quite rare nowadays. ok jca@
* Sync C mutex implementations.mpi2017-12-204-32/+62
| | | | ok kettenis@
* Switch x86 architectures to the common C mutex implementation.mpi2017-12-205-6/+307
| | | | | | This is a step towards MI mutexes. ok kettenis@
* Add a sanity check on the number of chunks beingkrw2017-12-201-1/+8
| | | | | assembled into a volume, to enure the on-disk metadata and the in-memory metadata agree.
* Pass calibrated TSC frequency to vmm(4) guests. On machines where TSCpatrick2017-12-191-3/+21
| | | | | | | | | | | cannot be calculated from the CPUID register, because the CPU does not support it, but can be calibrated from another timer, the vmm(4) guest doesn't have a chance to read or calibrate its own TSC frequency since it has no access to a precise enough time counter. In that case, fake the existence of the register and supply our calibrated TSC frequency. mlarkin@ tells me this also greatly helps Linux guests. ok mlarkin@
* curproc access isn't necessarily as cheap as a local variable access,deraadt2017-12-191-3/+4
| | | | | so only get it once ok guenther
* Remove unused ps_stackgap from process structstefan2017-12-192-6/+2
| | | | | Nothing uses this field since Linux compat was removed. ok mpi@ deraadt@ guenther@
* Remove a 27 year old #ifdef notdef chunk involving SB_LOCK.mpi2017-12-191-17/+1
| | | | ok bluhm@
* Inline socket buffer related defines, no functional change.mpi2017-12-191-11/+8
| | | | ok bluhm@
* Remove unnecessary unlock/lock dance when following a goto.mpi2017-12-191-5/+5
| | | | ok bluhm@
* Properly implement the rsb attribute and actually use it.kettenis2017-12-183-10/+11
|
* Add sensors for the AXP809.kettenis2017-12-181-3/+10
|
* Generate SIGBUS/BUS_ADRALN when we see a PC Alignment Exception from userland.kettenis2017-12-181-1/+6
| | | | ok patrick@
* Use axppmic as a dependency instead of axppmic_rsb.kettenis2017-12-181-2/+2
|
* Read RAM size for chips with SOCRAM like the BCM4334 on one of mypatrick2017-12-181-2/+18
| | | | Intel machines.
* Read RAM size for chips with SOCRAM like the BCM4334 on one of mypatrick2017-12-181-5/+55
| | | | Intel machines.
* Pass the BSSID to the join command so that we connect to the AP thatpatrick2017-12-181-4/+7
| | | | we chose instead of any AP the firmware chooses based on the SSID.
* Pass scan type directly to the scan command instead of using the globalpatrick2017-12-182-8/+6
| | | | variable.
* Add sizes for free() in libsa for amd64 and i386.fcambus2017-12-184-12/+12
| | | | OK tom@, deraadt@
* Enable RSB attachment for axppmic(4).kettenis2017-12-181-1/+2
|
* Add i2c interface code to the new axppmic(4) code and add support for thekettenis2017-12-183-164/+305
| | | | | | AXP152 and AXP209 here as well. The AXP209 is extended to support regulators and sensors. The ACIN and VBUS indicators are now exported as sensors. Retire the old AXP152/AXP209 support code.
* Map framebuffer into userland as non-cachable (but normal) memory. We stillkettenis2017-12-181-2/+4
| | | | | | | | | map the framebuffer as device memory in the kernel. Using mismatched memory attributes like this is discouraged (and we should probably fix this at some point) but this particular case is well-defined and the potential side-effects should not matter for thus use-case. ok patrick@
* Revert support for multiple threads to enter kqueue_scan() in parallel.mpi2017-12-182-64/+3
| | | | | | It is not clear if this change is responsible for the lockups experienced by dhill@ and jcs@ but since we're no longer grabbing the socket lock in kqueue(2) filters there's no need for this change.
* Revert grabbing the socket lock in kqueue(2) filters.mpi2017-12-183-28/+10
| | | | | | | | | | This change exposed or created a situation where a CPU started to be irresponsive while holding the KERNEL_LOCK(). These led to lockups and even with MP_LOCKDEBUG it was not clear what happened to this CPU. These situations have been experience by dhill@ with dcrwallet and jcs@ with syncthing. Both applications are written in Go and do kevent(2) & networking across multiple threads.
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.