summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2016-02-09bring in dlg's fix from rev 1.10 of the equivalent hppa file:Jonathan Gray
make __cpu_simple_lock provide serialisation of the critical section. that in turn makes atomic sequences actually atomic, which in turn means the refcnt api asserts wont fire erronously when if_get and if_put are actually used correctly.
2016-02-09Log frames which fall outside the BlockAack window in dmesg if theStefan Sperling
interface debug flag is set (enabled with: ifconfig iwn0 debug). Shows the frame's sequence number and the current BA window. I'm adding this for diagnosis, just in case it will be needed to make future decisions about tuning the heuristic which works around network stalls caused by such frames.
2016-02-09make __cpu_simple_lock provide serialisation of the critical section.David Gwynne
that in turn makes atomic sequences actually atomic, which in turn means the refcnt api asserts wont fire erronously when if_get and if_put are actually used correctly. such embarrassment. reported by landry@ who also let me debug on the affected machines ok jmatthew@
2016-02-09sync a function's comment with its signaturemmcc
ok tedu@, with a slight tweak
2016-02-08"exceeds past the end" -> "extends past the end" in comment.Kenneth R Westerback
2016-02-08Set EPT bits of guest pages in pmap_enter instead of doing itStefan Kempf
after an uvm_fault: uvm_fault maps in neighboring pages of the faulting page. We want EPT bits set for those as soon as possible as well. This avoids additional EPT violations causing further uvm_faults when the guest accesses the neighboring pages. discussion with and ok mlarkin@
2016-02-08Convert to uiomove. From Martin Natano.Stefan Kempf
2016-02-08Stop requiring a BlockAck session timeout (again), and just use it if the APStefan Sperling
is asking for it. This timeout should not be required anymore now that krw@'s hangs are fixed by working around APs which make sequence numbers jump about.
2016-02-08Work around buggy APs which occasionally emit sequence numbers much higherStefan Sperling
than the current 11n BlockAck window. The previous code would be fooled into moving the window forward and then drop packets until their sequence numbers catch up with the new window, which can take several minutes. Fixes traffic stalls observed with Broadcom APs. ok krw@ tb@
2016-02-07Fix off-by-one in loop termination condition.tb
ok stsp@
2016-02-07Hide some excessive net80211 debug printfs behind ieee80211_debug >= 2.Stefan Sperling
2016-02-07Comment.Masao Uebayashi
2016-02-07Remove a too strict assertion.Masao Uebayashi
2016-02-07Make sure the maximum iwm(4) firmware command payload size fits intoStefan Sperling
the 12 bits available for it in struct iwm_tfd_tb.hi_n_len. Patch by Imre Vadasz via tech@
2016-02-07Convert to uiomove. While there, use size_t consistently forStefan Kempf
variables 'left' and 'tocopy' and adapt printf and DPRINTF format strings accordingly. From Martin Natano.
2016-02-06Convert to uiomove. From Martin Natano.Stefan Kempf
2016-02-06Add support for the 9 series PCH.Mark Kettenis
From Dmitri Alenichev.
2016-02-06Enable the code that reads the BIOS from the ACPI VFCT table on platformsMark Kettenis
with ACPI. ok jsg@
2016-02-05Fix iwn(4) CCMP replay detection so it does not drop out-of-order A-MPDUStefan Sperling
subframes. Helps 11n mode with WPA. tested by me, tb@, and krw@
2016-02-05Count A-MPDU subframes with sequence number below the current BA window asStefan Sperling
duplicates rather than input errors. These subframes have either already been received, or the window was moved by the gap timeout which should only happen with buggy APs. Neither condition indicates a severe problem. Perhaps we will introduce a separate counter for this later.
2016-02-05Guard against integer overflow when checking whetherStefan Kempf
writing to a file on msdosfs stays within the max. file size. ok kettenis@, krw@
2016-02-05Define EDCA tables for 11n mode. Per 802.11-2012 they are the same asStefan Sperling
the tables for 11a/11g modes. Add a function to append a Microsoft WME parameter element to a frame, using EDCA tables to construct the parameter records. Some 11n AP's I have observed provide this element. None of this code is used yet, so no functional change. I wrote this while debugging BlockAck and then realized it was only needed for hostap mode. Once we support 11n in hostap mode and send A-MPDUs, this code will be needed.
2016-02-05In iwn(4), set max A-MPDU length to 64k instead of 4k and tell the firmwareStefan Sperling
about A-MPDU spacing. Makes BlockAck happy. tested by me, tb@, krw@, sthen@, and Henrik Friedrichsen
2016-02-05In iwm(4), set max A-MPDU length to 64k instead of 4k and tell the firmwareStefan Sperling
about A-MPDU spacing. Makes BlockAck happy. tested by myself and abieber@
2016-02-05Store ADDBA request and response parameters in the block ack record ofStefan Sperling
ieee80211_node. This way, we can keep track of the ACK policy and echo it back to the AP as required by the standard. And use the correct bit flag for the policy -- this code was confused between BlockAck and ADDBA, both of which have a policy bit but in different places. Fixes apple airport APs. tested by tb@, krw@, sthen@, abieber@, and Henrik Friedrichsen
2016-02-05Implement acpi_get_table_with_size(). Will soon be used to read VFCTMark Kettenis
tables in radeondrm(4).
2016-02-05return if the bpf_if passed to bpf_tap and _bpf_mtap are NULL.David Gwynne
this works around a toctou bug in a very common idiom in our tree, in between the two lines below: if (ifp->if_bpf) bpf_mtap(ifp->if_bpf, m, BPF_DIRECTION_OUT); figured out by and diff from haesbart
2016-02-05Simple moderation of Tx completion notificationsMike Belopuhov
Follow FreeBSD and schedule the next Tx completion event to fire when about half of the packet segments scheduled for transmission are consumed.
2016-02-05Silence warnings from static analyzers; found by jsg@Mike Belopuhov
2016-02-05Improve Linux PCI compatibility code.Mark Kettenis
2016-02-05Fix a pci vendor/device conversion missed in rev 1.6.Mark Kettenis
2016-02-05Implement FreeBSD-compatible IOCTL to access BMC in ipmi(4)Masao Uebayashi
Initial help & testing by jmatthew@ Code review & input by mpi@ Final review & OK by jsg@
2016-02-04Make iwn(4) pass 802.11 control frames in monitor mode.Stefan Sperling
Now we can see rts/cts, ack, blockack etc. in tcpdump(8). ok kettenis@
2016-02-04In iwn(4), don't forget about restoring the CCMP key to firmware afterStefan Sperling
HT protection setting updates. Unbreaks WPA in 11n mode. ok sthen@
2016-02-04Restore the BlockAck session timer. It is still required to work aroundStefan Sperling
stalled BA sessions observed with iwn(4). We can revisit this later once the underlying problem in iwn(4) has been found. Prompted by report from krw@, I could reproduce the issue. ok krw@
2016-02-04Bail early if there's no token; found by jsg@Mike Belopuhov
2016-02-04Prevent signed overflow when calculating i_modrev initial valueMike Belopuhov
Reported and fixed by Martin Natano <natano at natano ! net>, thanks! ok stefan
2016-02-04Some (buggy) UEFI implementations don't report the correct framebufferMark Kettenis
parameters unless SetMode() has been called, even if we don't actually change the mode. Other (equally buggy) UEFI implementations report an error if we call SetMode() but don't actually change the mode. Work around these issues by always calling SetMode() (like we did before rev 1.10), but don't report an error if the current mode is the same as the desired mode. ok yasuoka@
2016-02-04If the ROM size is 0, we won't be able to read the video BIOS from it.Mark Kettenis
Bail out early to prevent a panic from calling bus_space_map(9) with size 0.
2016-02-03Do not access the scsi link of the disk at the beginning of sdopen()Alexander Bluhm
and sdminphys() if the scsi disk is dying. The memory of the link may have been freed already. OK krw@
2016-02-03When accessing the scsi link of a scsi disk, use a variable "sc_link"Alexander Bluhm
everywhere. This is the first step to fix a use after free of the sc_link when the disk detaches. If a function gets a scsi transfer, the scsi link is always valid. Call this variable "link" consistently. OK krw@
2016-02-03Increase the RPC buffer once more to 4096 bytes. The guestinfo.ovfEnvReyk Floeter
mechanism is used to configure VMs in the VMware vSphere world: instead of using individual key-value guestinfo.* properties, it uses the guestinfo.ovfEnv value to pass an enterprise-compliant XML file that includes key-value properties. This file can be rather large, especially with comments, but 4k ought to be enough for anybody. Also change a stack buffer to malloc'ed memory in the ioctl path. OK mikeb@
2016-02-03Prevent efifb(4) from attaching if we're the console.Mark Kettenis
Unfortunately, making this decision in radeondrm_attachhook() is too late because at that point efifb(4) would have already been attached. This means that if we decide to bail in radeondrm_attachhook() we may end up without a glass console. That may happen for example if the firmware package has not been installed. I'm still looking for a solution for that problem. ok jsg@
2016-02-03Test cpuid_level or ci->ci_pnfeatset before using a CPUID leaf; some BIOSesPhilip Guenther
can disable leaves that CPU feature flags would seem to imply. Corrects signal delivery on systems where the AVX leaf is disabled. report and debugging help from Marcus MERIGHI (mcmer-openbsd (at) tor.at) ok kettenis@
2016-02-02give sxidog it's own mapping as well so it wont't require sxitimerJonathan Gray
from Artturi Alm
2016-02-02enable xen devices here as wellJonathan Gray
tested by and ok mikeb@
2016-02-02A few reliability improvements in the power management interfaceMike Belopuhov
Nathanael Rensen <nathanael at list ! polymorpheus ! com> came up with a few improvements to the event watcher and power management interface, namely: o Make sure to put our watcher on a list before issuing an XS_WATCH command since Xen will raise the event right after it's been set up. o The first time xen_control is called the "control/shutdown" node may not exist, so skip printing the error message in this case. o Acknowledge requests by writing back an empty string. o log(9) reboot and halt requests like vmt(4) does. Huge thanks!
2016-02-02Convert to uiomove(). From Martin Natano.Stefan Kempf
2016-02-02remove uneeded break statementsJonathan Gray
2016-02-01Add A20 support to sxirtc.Jonathan Gray
From Artturi Alm.
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-08 07:08:21 +0000