summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2016-01-13Don't confuse the HID descriptor address (which really is an office into theMark Kettenis
register space of the i2c device) with the i2c address. For i2c busses enumerated by ACPI we get the address from the I2C Serial Bus Connection Resource Descriptor returned by the _CRS methide of the i2c device. Pass the HID descriptor address in the ia_size member. ok jcs@
2016-01-13Add data structures and defines for Generic and I2C Serial Bus ConnectionMark Kettenis
descriptors. ok jcs@
2016-01-13Implement VFS read clustering for MSDOSFS.Martin Pieuchot
The logic used in msdosfs_bmap() to loop calling pcbmap() comes from FreeBSD and is not really efficient but it is good enough since it is only called when generating I/O. With this diff I get a 100% improvement when reading big files from a crappy USB stick. With this and bread_cluster(9) modified to not re-fetch B_CACHED buffers, reading large contiguous files with chunk sizes of MAXPHYS is almost as fast as physio(9) on the same device. For a 'real world' example, when copying music files from a USB stick I see a speed jump from 15MB/s on -current to 24Mb/s with this diff. While here rename some 'lbn' variables into 'cn' to better reflect what we're dealing with. Tested by Mathieu, with support from deraadt@
2016-01-13Prevent a double if_put().Martin Pieuchot
ok mikeb@, bluhm@
2016-01-13Tell the iwn(4) firmware to retry failed Tx at 1Mbit/s instead of MCS 0.Stefan Sperling
Seems to make Tx rate scaling go up faster and helps Rx performance. ok deraadt@
2016-01-13Fix iwn(4) firmware SYSASSERT during hardware initialization.Stefan Sperling
For some reason we would end up passing channel number zero to the RXON firmware command, which the firmware doesn't like. ok sthen@ reyk@ and also tested by stefan@, deraadt@, mikeb@
2016-01-13dont propagate any of the parent interfaces offload features on svlanDavid Gwynne
this avoids confusion about checksum offload if you configure ips on svlan interfaces on top of interfaces that do offload, like bge or em. the problem was reported by and this fix was tested by denis fondras analyzed by naddy@ thanks guys
2016-01-12naddy observed the TSC fallback code could encounter (high word) %edxTheo de Raadt
being 0; after multiply there is no perturbance to the from-disk entropy buffer. Further investigation showed perturbance was biased towards the lower bytes of a word. Compensate for this with a hocus pocus bit-spreading operation which applies a result byte by byte. discussion with kettenis, tb, mlarkin, naddy ok naddy
2016-01-12properly initialize quirks before sending them to hid layer, brokenJoshua Stein
during hid abstraction problem noticed and fix tested by abieber
2016-01-12rename new i2c_attach_args fields to avoid conflict with cpp magic inTheo de Raadt
isavar.h (a few parts of the tree include both) ok jcs
2016-01-12dev/isa/isavar.h does not appear to be needed by any of consumers.Theo de Raadt
(noticed because of another issue: one consumer is a high level busdriver, but also contains i2c support, so the isa #define for ia_irq clashes against i2c)
2016-01-12include files.hid before files.i2cTheo de Raadt
2016-01-12Use uiomove() instead of uiomovei(); from Martin NatanoStefan Kempf
ok deraadt@
2016-01-12A couple of unused variables have sneaked in from the debug codeMike Belopuhov
2016-01-12Convert XenStore code to free running producer/consumer indicesMike Belopuhov
After some hair pulling while implementing xnf(4) I've realised that Xen uses free running producer/consumer indices wrapping with their type (unsigned 32 bit integer). Later I've confirmed it with free chapters of the "The Definitive Guide to the Xen Hypervisor" by David Chisnall available online.
2016-01-12Kill the "doclusterread" and "doclusterwrite" defines.Martin Pieuchot
They are defined to 0 since the import of ext2fs 19 years ago.
2016-01-12Remove the "doclusterread" and "doclusterwrite" buttons.Martin Pieuchot
Cluster read is unconditionnally enabled for FFS since some years and toggling "doclusterread" has no effect. OpenBSD no longer support write clustering since the Dynamic Buffer Cache went in, so there is reason to delay the call to bawrite(9). Tested by Mathieu, ok dlg@
2016-01-12fix a misplaced curly braceStefan Sperling
2016-01-12Use task to execute command except polling context.Masao Uebayashi
2016-01-12Prevent a possible NULL-derefenced introduced by the move of RA sleepingMartin Pieuchot
code in a task. Such dereferenced can be triggered by receiving a RA with the 'on-link' bit set to 0 apparently generated by dnsmasq and reported by matthieu@. ok matthieu@, sthen@, bluhm@
2016-01-12Remove wireless turbo mode support. It is a non-standard extensionStefan Sperling
which only worked with ath(4) devices from a decade ago. Diff tested on ath(4) hardware by me to verify that 11a/b modes still work. ok reyk deraadt chris sthen kettenis
2016-01-12Compare if the prefix lengths are equal rather than checking if theMartin Pieuchot
netmasks are non-NULL for RTM_CHANGE and RTM_LOCK. Fix route(8) regressions tests 14 and 16 with ART kernels. These tests were failing because with ART, RTF_HOST routes always have a mask. ok claudio@, bluhm@
2016-01-12Prevent a NULL-reference if SIOCGVH is issued without carpdev argument.Martin Pieuchot
Problem reported by and diff tested by Fabian Raetz, thanks! ok benno@, bluhm@
2016-01-12Use task to execute watchdog tickle.Masao Uebayashi
2016-01-12Refactor ipmi_watchdog(); split tickle and set into separate functions.Masao Uebayashi
2016-01-12Improve emulation of PM registers on Hardware-redcued ACPI by emulatingMark Kettenis
SPL_TYPx, SLP_EN and WAK_STS using the new SLEEP_STATUS_REG and SLEEP_CONTROL_REG registers. Makes the transition into S5 (powerdown) work on the ASUS X205TA. ok deraadt@, mlarkin@
2016-01-12Add dwiic, a driver for the Synopsys DesignWare i2c controller foundJoshua Stein
on the Samsung ATIV Book 9 laptop. This initial version only supports ACPI config/attachment. Add ihidev, a HID-over-i2c driver largely based on uhidev. dwiic handles attaching ihidev devices found in ACPI. Add ims, a HID-over-i2c mouse/trackpad driver to get basic cursor and button functionality from HID-compliant i2c trackpads. ok kettenis deraadt
2016-01-12post the packet on em_82547 chips after bpfDavid Gwynne
now that start and txeof can run on different cpus, txeof could have freed the mbuf before bpf got to it.
2016-01-11Enable sdhc* at acpi?.Mark Kettenis
2016-01-11Make sdhc(4) attach to hardware IDs "80860F14" and "PNP0FFF".Mark Kettenis
2016-01-11Add a "mux -1" locator to wsdisplay(4) instances representing secondaryMark Kettenis
displays. This makes sure the keyboard mux gets picked up by the primary (console) display and that USB keyboards get paired with the console even if they are not marked as the console keyboard. This is what most people expect to happen. If you really want to associate your USB keyboard with a secondary display, you can detach it from the mux using wsconscfg -f /dev/ttyCcfg -k -d and attach it to another display using wsconscfg -f /dev/ttyDcfg -k
2016-01-11Use uiomove() instead of uiomovei(); from Martin NatanoStefan Kempf
ok kettenis@
2016-01-11xs_cmd should always return errno(2) codesMike Belopuhov
2016-01-11Do not fail when receiving an empty reply (or directory node) inReyk Floeter
xs_intr() but put an empty message in the queue. This prevents xs_reply() from being stuck in an endless loop because it expectes a message in the queue to break out of it. Depends on mikeb@'s previous commit because it would otherwise panic on trying to cleanup the empty message. OK mikeb@
2016-01-11Handle zero lenght messages in the xs_parse by returning an empty stringMike Belopuhov
Problem was reported and analyzed by reyk@
2016-01-11Make sendmsg() and recvmsg() functions take only struct ipmi_cmd *.Masao Uebayashi
Handle interface specific command length more nicely. No functional changes.
2016-01-11Allocate command data buffer on softc and use it from both ipmi_sendcmd()Masao Uebayashi
and ipmi_recvcmd(). No functional changes.
2016-01-11Only include sdhc_acpi.c if shdc* at acpi? is included in the kernel config.Mark Kettenis
2016-01-11Refactor buildmsg() functions to take struct ipmi_cmd * instead of 6Masao Uebayashi
arguments. No functional changes.
2016-01-11Introduce struct ipmi_cmd and have ipmi_cmd() function. IPMI commandMasao Uebayashi
sequence is always a pair of send and receive; call both from ipmi_cmd(). By making ipmi_cmd() take only one argument struct ipmi_cmd *, this also helps to prepare to call ipmi_cmd() as a task. No functional changes intended.
2016-01-11Remove useless sc_poll flag because it is always 1. Remove unused globalMasao Uebayashi
ipmi_poll too.
2016-01-11Read values from disabled sensors if possible. From jmatthew@.Masao Uebayashi
2016-01-11Add glue to make sdhc(4) attach to acpi(4). Needed to support the SD hostMark Kettenis
controllers integrated on Intel's Bay Trail SoCs.
2016-01-11Improve CIS parsing to recognize CISTPL_NULL codes instead of bailing out onMark Kettenis
them. Add symbolic constant for CISTPL_END. ok jsg@
2016-01-11Add (limited) support for controllers that implement version 3.0 of theMark Kettenis
SD host controller standard. Support the larger base clock and larger clock divisors. ok jsg@
2016-01-11do further work on the em transmit path to simplify the code.David Gwynne
noone could understand how em_txeof worked, so i rewrote it. this also gets rid of the sc_tx_desc_free var that needed atomic ops. space to use in em_start and space to free in em_txeof is now calculated from the producer and consumer. testers have reported better responsiveness with this. somehow. if em issues persist after this, im rolling back to pre-mpsafe changes.
2016-01-10match up code and comments, add missing $OpenBSD$; ok jsing@ a while agoChristian Weisgerber
2016-01-10Page fault handling tweaks for vmm:Stefan Kempf
- compute fault reason for uvm_fault() (e.g. page not present, protection violation) instead of passing a protection code - a page does not need to be zero'd after faulting it in. uvm_fault() does that for fresh anon pages already, and we also do not want a page that is swapped back in have its contents wiped. ok mlarkin@
2016-01-10Introduce acpi_intr_establish(9), an API to establish interrupt handlers forMark Kettenis
"platform" devices that attach to acpi(4). ok mlarkin@, deraadt@
2016-01-10Add thinkpad tablet dock/undock ACPI events.Stefan Sperling
ok deraadt kettenis
generated by cgit v1.2.3 (git 2.25.1) at 2024-12-04 18:14:58 +0000