summaryrefslogtreecommitdiff
path: root/sys
AgeCommit message (Collapse)Author
2019-09-05Fix previous commit.Mark Kettenis
ok millert@, sf@, deraadt@
2019-09-05Update Spleen kernel fonts to the latest released version, bringingFrederic Cambus
the following improvements: - Character at position CB is 'E' with diaeresis, not with tilde (12x24 version) - Character at position D6 is 'O' with diaeresis, not with tilde (12x24 version) - Character at position DC is 'U' with diaeresis, not with tilde (12x24 version) - Fix middle bar thickness in the upper case 'eth' character (12x24 version) - Fix ring alignment for the 'A' and 'a' characters (16x32 and 32x64 versions) - Fix tilde alignment in the 'o' with tilde character (32x64 version) - Remove strain pixel on the '3' and 'k' characters (32x64 version) OK deraadt@, mpi@
2019-09-05Nuke some 25+year old comments that the oracles at delphi would haveKenneth R Westerback
found obscure. Repair a broken WangTEK 5150ES quirks entry that apparently suffered some damage when imported in 1995. It never made NetBSD/FreeBSD quirk lists at all. Some whitespace tweaking to make everything line up.
2019-09-05Add mvdog(4), a driver to support the watchdog on the Armada 3700 SoC.Patrick Wildt
At the moment it only supports disabling the watchdog, which lets me continue to work on the Turris Mox. ok kettenis@
2019-09-05Supermicro X8DTH-i/6/iF/6F fails to attach uhci(4) via PCI and AHCI.Alexander Bluhm
Result is a crash at power down. Skip the activate function if attach has failed. Solution taken from ehci(4). OK deraadt@ kettenis@
2019-09-05Devices that have no quirks don't need to appear in the list ofKenneth R Westerback
devices with quirks.
2019-09-05Add missing RCS Id.Patrick Wildt
2019-09-05Revert iwm(4) to software WPA crypto because of apparent firmware issues.Stefan Sperling
For a small proportion of frames on 8265 we have observed firmware reporting "not encrypted" even though in Wireshark these frames appear as regular encrypted frames (rx_pkt_status=0xc0400007 type=0x8 subtype=0x80 m_len=1542). This causes frame drops which do not occur with software crypto. We might try again after a firmware upgrade. There seems to be no public record of bug fixes made by Intel between firmware versions which degrades our development process into pure trial-and-error. We need to waste our own time to find out information which Intel should be providing to the public. If anyone is aware of a source of such information, please let us know. All this is of course very disappointing. Problem reported by and debugged with help from solene@
2019-09-05Nuke unused struct block_desc_cipher. The Eighties called and wantKenneth R Westerback
their "CIPHER ST150S (old drive)" devices back.
2019-09-05Add Allwinner H6 support.Mark Kettenis
ok patrick@
2019-09-05em: Fix potential endless loopStefan Fritsch
If the NIC is in some error state (seen on a i219LM when em_read_phy_reg_ex() returns at "MDI Error"), it can happen that we loop endlessly because the loop variable is modified again somewhere down in the call stack. Use a separate variable to make the attach fail with "Hardware Initialization Failed" instead of hanging. yes deraadt@
2019-09-05Use the new set_blksz() interface to set the block size.Alexandre Ratchov
This simplifies the code and allows any block size multiple of 1ms to be used when play and recording number of channels are not the same. ok mpi@
2019-09-05Use the new set_{blksz,nblks}() interface to set the block size.Alexandre Ratchov
This simplifies the code and allows any block size supported by the hardware to be used. ok mpi@
2019-09-05Add the set_blksz() and set_nblks() audio driver functions.Alexandre Ratchov
The first sets the block size in frames, which is necessarily common to play and recording directions no matter the number of channels. The second sets the number of blocks per buffer for the given direction. Together, these two functions allow audio drivers to easily set the block size, matching both playback and recording constraints. The round_blocksize() didn't allow to do so because it returns the block size in *bytes*. Since the driver doesn't know if it's called for the play or for the record block size, it's impossible to calculate the block size in all cases if play and record number of channels are different. ok mpi@
2019-09-05Implement splassert() on mips64.Visa Hankala
2019-09-05Adjust interrupt priority levels on mips64 so that prioritiesVisa Hankala
of soft interrupts are lower than priorities of hard interrupts. This allows the delivery of hard interrupts while soft interrupts are masked.
2019-09-05Prepare the bat for kernels greater > 8MB of code, why because clang.Theo de Raadt
ok kettenis
2019-09-05Nuke unused st_softc field 'numblks'.Kenneth R Westerback
2019-09-05There can be only one -- mode.Kenneth R Westerback
So fold 'modeflags' field of st_softc into 'flags' field of same. Nuke a bunch of dubious/obvious comments.
2019-09-04Shuffle sdstart() and sdminphys() declarations closer to their friends.Kenneth R Westerback
Nuke pointless 'quirkdata' pointer in st_softc. We use the data once at attach time and don't need to remember where we got it from.
2019-09-04Hoist the quirks up another rung. drive_quirks is superfluous, justKenneth R Westerback
store the discovered quirks in st_softc's 'quirks' field.
2019-09-04'quirks' field of struct modes is superfluous. There is only 1 mode. Just useKenneth R Westerback
the 'quirks' field of struct quirksdata.
2019-09-04Various whitespace nits.Kenneth R Westerback
Reformat quirks array the way the autoindenter likes it to minimize the effort needed to keep diffs short.
2019-09-04The great density mode purge of 2006 (r1.62) was never carried throughKenneth R Westerback
to a logical conclusion. Start cleaning up the left overs. First, delete the unused quirk ST_Q_BLKSIZE and simplify the one condition referencing it in light of the fact that !ST_Q_BLKSIZE would always be true.
2019-09-04Add RCS Id.Alexander Bluhm
2019-09-04Fix a route use after free in IPv6 multicast route. Move theAlexander Bluhm
mrt6_mcast6_del() out of the rtable_walk(). This avoids recursion to prevent stack overflow. Also it allows freeing the route outside of the walk. Now mrt6_mcast_del() frees the route only when it is deleted from the routing table. If that fails, it must not be freed. After the route is returned by mf6c_find(), it is reference counted. Then we need a rtfree(), but not in the other case. Name mrt6_mcast_add() and mrt6_mcast_del() consistently. Move rt_timer_remove_all() into mrt6_mcast_del(). Reported-by: syzbot+af7d510593d74c825960@syzkaller.appspotmail.com OK mpi@
2019-09-04msdosfs: remove timezone supportcheloha
This support is undocumented, only works if you're using the kernel timezone, and breaks during a DST shift. It also preferences file systems managed by a Windows installation: many implementations, like ours, use UTC by default (think: phones, digital cameras). No complaints on tech@. "good riddance" tedu@, "Yep." deraadt@
2019-09-04Remove DST/TIMEZONE options(4) from kernel configs; ok jca@ deraadt@cheloha
2019-09-04gettimeofday, settimeofday(2): limit timezone supportcheloha
For gettimeofday(2), always copy out an empty timezone struct. For settimeofday(2), still copyin(9) the struct but ignore the contents. In gettimeofday(2)'s case we have not changed the original BSD semantics: the kernel only tracks UTC time without an offset for DST, so a zeroed timezone struct is the correct thing to return to the caller. Future work could move these out into libc as stubs for clock_gettime and clock_settime(2). But, definitely a "later" thing, given that we are in beta. Update the manpage to de-emphasize the timezone parameters for these syscalls. Discussed with tedu@, deraadt@, millert@, kettenis@, yasuoka@, jca@, and guenther@. Tested by job@. Ports input from jca@ and sthen@. Manpage input from jca@. ok jca@ deraadt@
2019-09-04Add comment.Mark Kettenis
2019-09-03Add support for getting the frequency of the CPU clocks.Mark Kettenis
2019-09-03Shorten normal dmesg attach verbiage and expand SCSIDEBUG verbiage.Kenneth R Westerback
ok deraadt@
2019-09-03Enable amlmmc(4).Mark Kettenis
ok patrick@
2019-09-03Add clang support.Mark Kettenis
2019-09-03a sprinkling of function ansificationTheo de Raadt
2019-09-03drop support for building with gcc; ok deraadt@Christian Weisgerber
2019-09-03only need one start: label; ok kettenisTheo de Raadt
2019-09-03delete incorrect casts, and a bit if ansi functions; ok jcsTheo de Raadt
2019-09-03Use cmplw mnemonic instead of cmpl with only 3 operands. Make clang happy.Mark Kettenis
ok deraadt@
2019-09-03some cleanup for clang; ok kettenisTheo de Raadt
2019-09-03Replace empty body of while loop with continue statement.Mark Kettenis
ok deraadt@
2019-09-03correct INET6 path in mpls_getttl() and avoid uninitialised variableJonathan Gray
ok dlg@
2019-09-03remove duplicate test introduced in rev 1.3Jonathan Gray
2019-09-03Avoid uninitialized variable; using a for loop like in siif_recvmsg() wasMark Kettenis
what I had intended to write anyway. spotted-by & ok jsg@
2019-09-03bnxt: Support MSI-XStefan Fritsch
Tested on a BCM57412 ok jmatthew@
2019-09-03remove unreachable returnJonathan Gray
ok kettenis@
2019-09-03ansify functionsTheo de Raadt
ok jsg
2019-09-03correct indent that triggered clang to freak out; with jsg.Theo de Raadt
2019-09-02Various fixes to make clang happy.Mark Kettenis
ok deraadt@
2019-09-02Implement setting the CPU clock for Allwinner A64 SoCs.Mark Kettenis
From Krystian Lewandowski.
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-18 17:29:04 +0000