| created | 2020-02-22T19:44:05Z |
|---|---|
| begin | 2020-02-16T00:00:00Z |
| end | 2020-02-17T00:00:00Z |
| path | src/sys |
| commits | 7 |
| date | 2020-02-16T03:23:05Z | |||
|---|---|---|---|---|
| author | jmatthew | |||
| files | src/sys/dev/pv/vmt.c | log | diff | annotate |
| message |
Force a guest info update when we get a Capabilities_Register command. ESXi doesn't carry the guest info over when a guest gets moved to a different host, and this appears to be the only hint the guest gets that this has happened. ok dlg@ |
|||
| date | 2020-02-16T07:55:30Z | |||
|---|---|---|---|---|
| author | anton | |||
| files | src/sys/sys/sysctl.h | log | diff | annotate |
| message |
Cope with latest changes to the ps_flags field of struct process which is unsigned by now. ok millert@ visa@ |
|||
| date | 2020-02-16T07:59:08Z | |||
|---|---|---|---|---|
| author | anton | |||
| files | src/sys/kern/sys_pipe.c | log | diff | annotate |
| message |
Unconditionally acquiring a write lock in pipe_ioctl() is quite excessive as only one command actually modifies the pipe. The sigio subsystem is already internally protected using its own lock. This is similar to what soo_ioctl() already does. ok mpi@ visa@ |
|||
| date | 2020-02-16T09:33:47Z | |||
|---|---|---|---|---|
| author | kettenis | |||
| files | src/sys/arch/arm64/arm64/locore.S | log | diff | annotate |
| src/sys/arch/arm64/arm64/syscall.c | log | diff | annotate | |
| message |
Change arm64 system call ABI to skip two instructions. This allows us to insert a barrier after each system call to address a speculative execution issue discovered by Anthony Steinhauser. A change to insert two nop instructions after each system call instruction in crt0.o/libc/ld.so has been committed a few weeks ago, so the transition should go smoothly. However, old static binaries and binaries that roll their own syscalls (like go) will no longer work. Snapshot packages that work with the new syscall ABI are available (including a fixed go). Together with deraadt@ and guenther@ ok deraadt@ |
|||
| date | 2020-02-16T17:39:47Z | |||
|---|---|---|---|---|
| author | krw | |||
| files | src/sys/dev/isa/wds.c | log | diff | annotate |
| message |
16 << PGSHIFT (which is >= 12) is always >= MAXPHYS (64 * 1024) so wds_minphys() is just duplicating what minphys() will do. Nuke it. |
|||
| date | 2020-02-16T19:34:59Z | |||
|---|---|---|---|---|
| author | anton | |||
| files | src/sys/kern/vfs_vnops.c | log | diff | annotate |
| message |
Rework vn_ioctl() to only have a single point of return. This will make it easier to grab the kernel lock once ioctl() is unlocked. Thanks to semarie@ who came up with an improved diff. ok mpi@ semarie@ visa@ |
|||
| date | 2020-02-16T23:37:23Z | |||
|---|---|---|---|---|
| author | jca | |||
| files | src/sys/arch/loongson/dev/apm.c | log | diff | annotate |
| message |
Send a resume event to apmd(8), like done by apm(4) and acpi(4) Test & ok visa@. Sadly his Lemote doesn't come back from sleep, with or without this change. |
|||