IABSD.fr/src/sys/dev
Branch :
Log
| Author | Commit | Date | CI | Message |
|---|---|---|---|---|
 |
0a3861fc | 2017-07-06 10:09:26 | sync the list of pci devices which don't require aperture ok kettenis@ | |
 |
e12511f5 | 2017-07-05 20:30:13 | Fix native/raw backlight support in inteldrm(4). | |
|
b98ddde8 | 2017-07-05 12:43:32 | regen | |
|
784b66df | 2017-07-05 12:43:10 | Add Intel Braswell Sensor Hub device. | |
|
e4dbbea0 | 2017-07-04 22:27:23 | Ignore entries with specific PCI subvendor/subdevice such that we don't inadvertedly match the Intel HD Graphics P4000 as a "Quanta transcode" device. Thanks to Joe Gidi for figuring out that I inadvertedly brought this back. Hopefully fixing it this way prevents it from happening again. | |
 |
76ead70e | 2017-07-03 23:27:53 | blacklist MSI on intel m.2 optane parts. it just doesnt work there. discovered by jmatthew@ tweaks and ok jsg@ kettenis@ | |
 |
50cc059d | 2017-07-03 13:26:04 | merge error, revealed by clang; ok kettenis@ | |
 |
8443256d | 2017-07-03 09:21:09 | Replace slot time durations with macros. ok stsp@ | |
|
c2fbad6e | 2017-07-03 06:33:24 | regen | |
|
7704e1c4 | 2017-07-03 06:32:54 | intel optanes are a thing | |
|
1a8d1b18 | 2017-07-02 20:58:55 | Enable MIPI/DSI support for inteldrm(4). Makes the display on my ASUS Transformer Book T100HA work (with some ugly error/warning messages). | |
|
e198220f | 2017-07-02 14:48:19 | Add the definition of IEEE80211_DUR_DS_SHSLOT. From IEEE Std. 802.11-2016, Table 18-5 "ERP characteristics", p. 2332: aSlotTime characteristic: If dont11OperatingClassesRequired is false: Long = 20 us Short = 9 us ok stsp@ | |
|
3253c27b | 2017-07-01 16:14:10 | Update inteldrm(4) to code based on Linux 4.4.70. This brings us support for Skylake and Cherryview and better support for Broadwell and Valleyview. Also adds MST support. Some tweaks to the TTM code and radeondrm(4) to keep it working with the updated generic DRM code needed for inteldrm(4). Tested by many. | |
|
4e5201e3 | 2017-07-01 16:12:03 | Update inteldrm(4) to code based on Linux 4.4.70. This brings us support for Skylake and Cherryview and better support for Broadwell and Valleyview. Also adds MST support. Some tweaks to the TTM code and radeondrm(4) to keep it working with the updated generic DRM code needed for inteldrm(4). Tested by many. | |
|
e0b53cee | 2017-07-01 16:00:25 | Update inteldrm(4) to code based on Linux 4.4.70. This brings us support for Skylake and Cherryview and better support for Broadwell and Valleyview. Also adds MST support. Some tweaks to the TTM code and radeondrm(4) to keep it working with the updated generic DRM code needed for inteldrm(4). Tested by many. | |
|
5fb81a6c | 2017-07-01 15:56:11 | - Avoid possible use of uninitialized variable; noticed by jsg@ - Only set value in R92C_FAST_EDCA_CTRL register if the device is combo WiFi/Bluetooth; confirmed by Yan-Hsuan Chuan from Realtek - There's no need to set the default value for the R92C_PIFS since default value is 0x1c; confirmed by Yan-Hsuan Chuan from Realtek - Disable usb Rx aggregation mode for rtl8188cu/rtl8192cu; vendor drivers use DMA Rx aggregation mode for rtl8188cu/rtl8192cu/rtl8188eu and use USB Rx aggregation mode for later chipsets. Tested by stsp and me. ok stsp@ | |
 |
a796f3c8 | 2017-06-30 06:25:29 | Fix a format string specifier error in a debug printf. Noticed and reported by Fasse: fasse dot f2 at gmail dot com Thanks. | |
 |
6b1c7616 | 2017-06-29 17:36:16 | kill RCSID macros; discussed with millert | |
 |
8ad1e6fc | 2017-06-29 03:48:44 | set c.c_maxrxlen to something. apparently lost a line somewhere. maybe it works now? :) from Colin Stolley via Paul B. Henson | |
 |
d16b5843 | 2017-06-27 13:56:15 | Provide a handler stub for the "channel rescind" message | |
|
f2643dce | 2017-06-27 12:04:26 | Avoid testing an uninitialised variable in the SCSI_BUSY/SCSI_QUEUE_FULL paths of hvs_scsi_cmd_done(). ok mikeb@ | |
|
f62d7836 | 2017-06-26 20:12:14 | Minor cleanup; when polling delay before calling the interrupt handler | |
|
398942c6 | 2017-06-26 18:42:23 | Rework the deferred interrupt loop By performing a task_add an interrupt handler can rely on the taskq_thread to invoke it again with an additional benefit of being able to sched_pause when required. In the long run more than 99.8% of calls do not require an additional iteration. | |
 |
c14d8484 | 2017-06-26 07:02:16 | Factor a variable assignment in audiopoll(). From Michael Bombardieri, thanks. | |
|
6ab8fabf | 2017-06-23 19:09:21 | Event interrupt handler should call hv_channel_schedule now | |
|
f1ed34a9 | 2017-06-23 19:07:13 | Request deferred interrupt processing | |
|
8f767439 | 2017-06-23 19:05:42 | Introduce deferred interrupt processing capability Hyper-V interrupts seem to be sometimes delivered before the message becomes available on the channel ring. This is reproducible on hvs(4) under load. This change is modelled on the workaround found in the Linux driver. | |
|
c2027333 | 2017-06-23 18:54:44 | s/membar_sync/virtio_membar_sync/ | |
|
c16b4e2e | 2017-06-23 18:51:28 | Finish up minor changes in the output and clear the BATCHED channel flag | |
|
08300c66 | 2017-06-23 14:41:54 | - Fix Tx queues to USB endpoints mapping and merge urtwn_r92c_dma_init() and urtwn_r88e_dma_init() into one; from FreeBSD r291902. - Change the way the number of pages for public queue is calculated; Based on the vendor driver's _InitQueueReservedPage(), the formula for calculation of the number of pages for public priority queue is as follows: if the device has 3 endpoints: PUBQ = total number of Tx pages - HQ - LQ - NQ if the device has 2 endpoints (i.e. has high and normal priority queues): PUBQ = total number of Tx pages - HQ - NQ - Don't write MCS8..MCS15 Tx power if the WNIC is a 1-stream Tx; - Set R88E_TX_PAGE_COUNT to a correct value; - Minimal enhancement in urtwn_open_pipes() for upcoming rtl8192eu chipset support; Tested with USB-N10 Nano (rtl8188cu), DWA-131 rev B (rtl8192cu), and TL-WN725N v2 (rtl8188eu) on amd64. ok stsp@ | |
 |
8d8f7df7 | 2017-06-22 06:21:12 | Add a hypervisor-specific function pointer in pvbus(4) that gets called during cpu_hatch() on each cpu, allowing initialization of per-cpu features on each cpu as it starts up. input from reyk@ and mikeb@ ok mikeb@ | |
 |
31de4d1e | 2017-06-22 05:32:47 | Fix another ;; in sys/dev | |
|
114e9c63 | 2017-06-22 02:44:37 | double ;;. xhci one found by geoffhill | |
|
4f93181d | 2017-06-21 18:18:08 | Don't confuse SCSI command status and transfer error code. While here, make SCSI command completion routine more robust. | |
|
f16a32e7 | 2017-06-21 16:37:48 | Return early from NVS command submission if no response is required | |
|
c099a505 | 2017-06-21 11:35:11 | Include the request id in the I/O debug output | |
 |
bd38a9bd | 2017-06-20 15:49:00 | Check the validity of the `qtd' pointer after each interrupt case and bail out if the pointer has become invalid. This prevents use-after-free memory accesses that corrupted the dwc2qtd pool. This fix improves stability on the Raspberry Pi 3. From Linux commit dc8730846948e517169f630826cd2c97615f5ee8 OK kettenis@ | |
 |
53de56e5 | 2017-06-20 13:52:40 | Remove the IWM_FLAG_HW_INITED flag from iwm(4). This flag was redundant. The IFF_RUNNING flag in struct ifnet serves the same purpose. ok mpi@ kettenis@ deraadt@ | |
 |
8c3115d8 | 2017-06-19 11:03:30 | Add a handler for the WSDISPLAYIO_GINFO ioctl in radeondrm, allowing to retrieve basic information about a framebuffer display. OK visa@, deraadt@ | |
|
08624e9e | 2017-06-19 11:00:18 | Add a handler for the WSDISPLAYIO_GINFO ioctl in inteldrm, allowing to retrieve basic information about a framebuffer display. OK visa@, tedu@ | |
 |
acc71172 | 2017-06-19 09:36:27 | Do not call txeof nor rxeof in the watchdog routine. They cannot be serialized with the interrupt routine and are useless because the driver is reinitialized right after. Pointed by and ok mikeb@ | |
 |
995fbf01 | 2017-06-18 13:34:03 | pms/alps: configure compat mode | |
|
ec55cf64 | 2017-06-18 13:21:48 | Remove remnants of old versions. | |
|
b172cb7a | 2017-06-17 19:20:30 | The fonts we are looping through in vga_selectfont() are the builtin font which is using the IBM encoding, and some potentially loaded fonts which are either IBM or ISO encoded. Therefore the condition checked by vga_valid_font() is always true, and we can remove it. Thanks to miod@ for pointing this out. OK tb@ | |
|
34a2cade | 2017-06-16 14:57:51 | - Fix incorrect values in the computation of transmit power for the rtl8188eu chipset. - Small code refactoring: - replace hardcoded rate indexes with their name, like athn(4). - replace fields offsets with the structure; from FreeBSD r294198. - be compliant with definitions of the efuse in vendor hal_pg.h and rename struct r92c_rom member names: s/channel_plan/reserved5/, s/xtal_calib/channel_plan. - no need to disable HWPDN twice in urtwn_r88e_power_on(). Tested by stsp@ and myself with RTL8188CUS, RTL8192CU, RTL8188EU, and RTL8188CE devices. ok stsp@ | |
|
5ae9823c | 2017-06-16 08:45:34 | In iwm_init() save the generation counter once before looping over tsleep(9), instead of saving it during every iteration. | |
|
175c85ce | 2017-06-15 20:12:12 | Don't defer attaching PV devices Now that both hvn(4) and hvs(4) can perform VMBus channel operations during autoconf, it's no longer necessary to defer their attachment. | |
|
5328d3b1 | 2017-06-15 20:08:56 | Make sure we can free GPA handles when cold | |
|
b6117676 | 2017-06-15 20:06:13 | Support executing NVSP and RNDIS commands during autoconf | |
|
2bedbdd7 | 2017-06-15 18:21:08 | Provide storage for the response message in each CCB | |
|
559d9ea0 | 2017-06-15 18:05:52 | Remove the structure definition for an old (pre Win8) I/O request type since the new one is backwards compatible and the correct length of the request can be specified w/o an actual definition. | |
|
332f60c2 | 2017-06-15 18:00:26 | Put back copying of the response message into the polling callback The response message is allocated on the stack of hvs_intr and its content is valid during a single message processing loop iteration. Due to the fact that polled transfers finish completion in a callee context, the cached pointer to the response message becomes stale very quickly. To avoid this situation the poll completion callback needs to make sure the content of the response is provided to the next completion routine. | |
|
d917c196 | 2017-06-15 15:01:28 | Correct the number of targets vs number of LUNs jsg@ has found out that the number of targets wasn't correct and it turns out that the device supports up to 2 targets and 1 or 64 LUNs depending on whether it's taking over an IDE device or is a virtual SCSI. While here make sure the command response is copied back only for synchronous commands that are issued during initialization phase. | |
|
96ea3eb5 | 2017-06-15 11:48:49 | We do not support iso7 nor pcvt encoding, so remove macro definitions and commented entries. OK mlarkin@ | |
|
ae828fe0 | 2017-06-14 22:24:11 | Borrow some of dlg's smarts to make this look like a real SCSI driver | |
|
8ce927c3 | 2017-06-14 16:58:28 | Currently, iwm(4) uses only one phy context, so stop sending commands to the firmware for unused ones. This should be refactored further, but for now this change is good enough to reduce the amount of commands we throw at the firmware. ok deraadt@ | |
|
31bde467 | 2017-06-14 16:57:47 | When a task in iwm(4) obtains an rwlock(9), bail out if the device was reset while the task was asleep waiting for the lock. ok deraadt@ | |
|
90e46a7b | 2017-06-14 16:56:50 | Make the iwm(4) ioctl handler wait until the transition to SCAN has completed when bringing the interface up. Otherwise we get racy behaviour when running ifconfig commands which issue several ioctls in quick succession. Should fix the occasional "could not initiate scan" errors. ok deraadt@ | |
|
626e1a05 | 2017-06-14 16:56:04 | Make iwm(4) wait for the PHY calibration result notification during HW init. Code inspection revealed that Linux does this, and so should we. ok deraadt@ | |
|
77269c59 | 2017-06-14 12:28:41 | Special handling for request 0 clashes with CCB 0 | |
|
01903976 | 2017-06-14 12:05:37 | Increase the maximum amount of requests and do some cleanup | |
|
7f9668fd | 2017-06-14 10:25:40 | Hookup hvs(4) | |
|
3ae168bd | 2017-06-14 10:24:51 | Hyper-V storage driver Based on the FreeBSD driver, attaches only to emulated SCSI controllers. | |
|
1f3992c0 | 2017-06-14 09:47:37 | Add another type of a scatter-gather list operation used by StorVSC Obtained from FreeBSD. | |
|
13e4c7ef | 2017-06-13 08:00:24 | Increase WSFONT_NAME_SIZE size, this will allow storing longer font names in wsdisplay_font structures. OK deraadt@ | |
 |
4fe5a5de | 2017-06-12 16:39:51 | Limit the maximum size of softraid crypto volumes that can be created to the size that is currently supported (a data area of 16TB, or 2^32 * 32 + 528 blocks including the softraid metadata/boot area), rather than successfully creating the volume and then failing during resource allocation (without properly reporting why it failed). Found the hard way by sharon s. <mymlact at gmx dot com> | |
|
67b01768 | 2017-06-12 15:43:25 | Upon failure, only return zero from the softraid ioctl handler when there is at least one bio status message. It is a much nicer user experience when we properly report failures, however in the case that we fail to do this, getting an errno back from the ioctl is somewhat better than simply printing "unknown error" from bioctl(8). | |
|
f6a8865d | 2017-06-12 15:15:08 | Fix another long standing softraid crypto bug where if all 32 keys are in use, when freeing crypto sessions we run straight off the end of the array and start blatting memory - clearly no one has a softraid crypto volume that exceeds 15.5TB in size... | |
|
2dc30802 | 2017-06-12 15:09:07 | Factor out and reuse the code that frees the crypto sessions, instead of duplicating it. | |
|
b436e181 | 2017-06-12 14:46:00 | Fix a long standing off-by-one bug in the softraid crypto number of keys calculation - we allow one key per 0.5TB, which should allow up to 16TB disks, however the disk size was treated like a block offset and shifted. This meant that the maximum size was actually 16TB minus one block. While here also calculate the number of keys as an absolute value, rather than as the upper inclusive bound - adjust the logic in the associated for and if statements to match. ok krw@ tb@ | |
|
56bd49a9 | 2017-06-12 13:45:39 | specify free() size from the old allocation, not new allocation. fix from C Turt | |
|
44ac566c | 2017-06-12 12:35:07 | ifq_start does OACTIVE and RUNNING checks for the driver Reminded by dlg@. | |
|
7f0b5f10 | 2017-06-12 04:57:42 | typo; from bryan vyhmeister | |
|
56bdc47d | 2017-06-12 03:00:26 | From FreeBSD (r227593, r307982): More and more RealTek controllers started to implement EEE feature. Vendor driver seems to load a kind of firmware for EEE with additional PHY fixups. It is known that the EEE feature may need ASPM support. Unfortunately there is no documentation for EEE of the controller so enabling ASPM may cause more problems. The Realtek vendor drivers for FreeBSD and Linux also disable ASPM and clock request. While here, add a define for the ECPM (Enable Clock Power Management) bit. Tested by stsp@ and myself. ok stsp@ | |
|
921381e4 | 2017-06-11 02:06:36 | integer overflow for two range checks fix from C Turt, ok miod | |
|
d5bd7054 | 2017-06-10 12:58:37 | Pass M_CANFAIL to malloc(9) calls which use M_WAITOK but are tested for failure. ok armani@ | |
|
088d0fa3 | 2017-06-09 20:38:48 | Fill RX ring during init and bail early on send if OACTIVE is set or IFF_RUNNING is not. | |
|
1d4f8416 | 2017-06-09 14:36:43 | Remove a WAITOK that has sneaked in | |
|
a24f5c54 | 2017-06-09 14:34:10 | Convert to ifq_dequeue and perform m_defrag if the mbuf doesn't fit With suggestions from and OK dlg | |
|
5a92a6a8 | 2017-06-09 13:47:26 | Introduce iwm_nic_assert_locked() to verify that the driver has correctly requested MAC access before accessing certain registers, as required by the hardware. Use it to assert that hardware is still in an accessible state before reading or writing such a register. For now, panic if that check fails. The long term goal is to make this a non-fatal error and handle it properly in all code paths that end up reading or writing such a register. Fix a missing NIC lock on 8000 hardware, found by this new assertion. Also, grab the NIC lock early during hardware init and keep it until init is done. The previous code relinquished and reacquired the NIC lock several times during the init sequence. It seems this is what was causing some random errors when the interface was brought up, such as "could not enable Tx queue", "could not add aux station", and "could not add phy context". For some reason, bsd.rd kernels were suffering particularly hard from such problems, to the point where some machines could not be upgraded over iwm(4). This change does not eliminate such problems entirely but is a step forward. Prodded by deraadt@ This change has already been in snaps for a while. | |
|
e334da63 | 2017-06-09 13:46:15 | Fix the resume code path in iwm(4) to no longer call iwm_stop() before initializing the hardware. Prevents "acquiring device failed" messages during resume. Also, start the hardware up in DVACT_RESUME already and verify that it has started by the time we reach DVACT_WAKEUP, before scheduling the iwm_init_task which loads firmware etc. With help from deraadt@ test & ok tb@ | |
|
bf196442 | 2017-06-08 01:34:00 | make the gem tx path MPSAFE. this mostly follows the pattern in ifq.h ok jmatthew@ | |
|
a338ceb4 | 2017-06-07 15:49:21 | Acquire submit queue mutex only once per xbf_complete_cmd invocation and remove some leftover assertions. | |
|
bc260131 | 2017-06-06 21:53:07 | Set the weak variant of hysteresis as default. | |
|
705639a4 | 2017-06-06 21:12:01 | Rewrite the driver to handle 64kb transfers Although several codepaths in the kernel such as coredump and buffercache read-ahead feature assume that underlying hardware is capable of handling 64kb transfers without any issues, xbf was setup to rely on a single descriptor per transfer which limited the maximum size of an individual transfer to 11 4k segments amounting to 44k bytes. To avoid overbooking, a metadata object is allocated for each transfer to keep track of associated descriptors limiting the maximum amount of outstanding transfers to half the ring size. The issue was reported by Dan Cross <crossd at gmail.com>, thanks! | |
|
9aecda2c | 2017-06-06 20:33:28 | Call xbf_intr for polled transfers that can't sleep | |
|
31a33b6e | 2017-06-06 19:47:22 | Add support for tap gestures. | |
 |
5396f3b8 | 2017-06-06 19:01:27 | Provide sufficient info in error messages to identify softraid volume and backing disk when i/o errors occur. Original issue reported by Paul de Weerd. Suggestions from jsing@. ok deraadt@ mikeb@ | |
|
42000b0b | 2017-06-06 01:02:18 | Regen. | |
|
0e67be84 | 2017-06-06 00:52:02 | Sort DLINK section and add USB device ID of D-Link DWA-131 rev E1. ok stsp@ | |
|
91637d79 | 2017-06-05 04:57:37 | use byte swapping loads and stores. shrinks the code a bit on sparc64. ok jmatthew@ | |
|
e0047aae | 2017-06-04 20:28:05 | fix a bungled logical/bitwise expression pointed out by clang; ok deraadt@ | |
|
a86c5d82 | 2017-06-04 19:23:20 | Remove unused function, as pointed out by clang. ok deraadt@ kettenis@ | |
|
37bc6d24 | 2017-06-04 15:06:22 | Switch the radeondrm(4) driver over to TTM using the new DRM VMA manager. ok jsg@ | |
|
beea9bb1 | 2017-06-04 14:02:24 | Switch the TTM code over to the generic DRM VMA manager. ok jsg@ | |
|
6aba050a | 2017-06-04 05:04:24 | update link to Ken Kato's vmware backdoor protocol documentation from Seth Jackson | |
|
35debf69 | 2017-06-04 04:29:23 | use htolem32 and lemtoh32 instead of htole32 and letoh32. saves about half a k on sparc64. | |
 |
a28b3ce1 | 2017-06-03 08:50:38 | vioblk/virtio: Re-arrange some struct members This reduces holes/padding and makes the structs smaller. | |
|
c929981a | 2017-06-02 20:25:50 | Perform grant table page allocation outside of the table mutex witness(4) has found that km_alloc will trigger an rw_enter via uvm_map and vm_map_lock. While rw_enter is called with RW_SLEEPFAIL, there's also an msleep in there, so it's easier to avoid getting in the middle of that. | |
|
04c31ef2 | 2017-06-02 16:04:36 | regen |