Age | Commit message (Collapse) | Author |
|
OK lteo@ sashan@
|
|
Check if n_blocks is less than UINT32_MAX and not the image size
in various places (READ_CAPACITY_* and GET_CONFIGURATION).
ok mlarkin@ and dlg@
|
|
of doing it manually.
OK sashan@ visa@
|
|
won't flood the terminal, since the first redirect changes the host
discussed with sthen
|
|
The error was:
--8<--
ffs.S:57:2: error: unsupported relocation on symbol
adrne r2, .L_ffs_table
^
-->8--
Minimal workaround suggested by deraadt@, ok guenther@ kettenis@
|
|
data. diff provided by alf <a.schlichting at lemarit.com>
ok millert
|
|
|
|
came also up with this diff. Simple, no bias and benchmarks show the extra
random calls disappear in te measurement noise.
|
|
|
|
devices from attaching (ex uaudio devices).
ok mpi
|
|
no other process which could free it. Better panic in malloc(9)
or pool_get(9) instead of sleeping forever.
tested by visa@ patrick@ Jan Klemkow
suggested by kettenis@; OK deraadt@
|
|
Ensures that when URL redirection takes place, the rest of the install/upgrade
session will stick to that redirected location. This should help ensure that
the fetched set of files is coherent when a load balancer emits 301/302.
With feedback from rpe@
OK sthen@ rpe@
|
|
callback is just going to be fired again straight away to go through all
the lines, it is better just to use the last one straight away.
|
|
than one core available.
When bsd.mp is used, bsd.sp is not copied to the boot partition.
Old installations might not have enough space for all bsd.mp, bsd.sp
and bsd.rd.
OK deraadt@, krw@, pirofti@
|
|
the value matches with the number of cores that GENERIC.MP uses.
OK deraadt@, krw@, pirofti@
|
|
From canacar@, thanks!
|
|
|
|
but at that time ffs function calls were generated instead of the
compiler inlining the code. Now that ffs is marked protected in
libc this is handled better. Thanks to kshe who prompted me to
look at this again.
|
|
dance, mark it protected. This works better for both gcc and clang: gcc
blocks overriding of internal calls, while clang permits inlining again.
ok otto@
|
|
skip symbol lookup on protected symbols. Add visibility #defines to
<sys/exec_elf.h> to support that.
ok kettenis@ visa@
|
|
|
|
|
|
OK deraadt@, visa@, mpi@
|
|
ok mibek
|
|
debug printfs about LDC link state changes. No functional change.
ok kettenis@
|
|
we switched to compiled regexes. Fixes fw_update(1).
Noticed by deraadt@
ok sthen@
|
|
This will help for round-robin sites in anchoring a pkg_add session to
a single site, thus removing shearing effects.
Initial idea from job@, reworked a lot by me,
enthusiastic okay sthen@
|
|
attacks by flushing the branch predictor cache (BTB) on context switches
and page faults in kkernel address space. Note that this relies on the
presence of firmware (such as Arm Trusted Firmware) that provides PSCI
services that flush the BTB on entry as described in Arm Trusted Firmware
Security Advisory TFV 6.
ok patrick@, visa@
|
|
we can use it to defend against branch target injection attacks.
ok patrick@, visa@
|
|
in bootloader printfs.
Feedback from and ok gerhard@
|
|
Diff from jcs@, ok pirofti@
|
|
EABI aliases these can lead to duplicate symbol errors, especially when
mixing code compiled with gcc and clang. The intention is to revert this
once we convert armv7 to hardfloat.
ok patrick@, guenther@
|
|
This helps tools like scan-build, and follows the example of warningf()
and errorf(). ok anton@
|
|
ok tedu@
|
|
it's rarely subject to change. While here, unifdef RLIMIT_VMEM.
ok jca@
|
|
ok dlg@
|
|
emacs.c:1041:2: warning: Value stored to 'cp' is never read
|
|
Ensures the correct duration is logged even if the system
time is changed during the session.
ok jca@
|
|
accidentally broken in last commit.
|
|
ok deraadt@
|
|
|
|
|
|
|
|
an invalid source address. As pf route-to happens after IP source
selection based on a different route, the address should be corrected
after pf has allocated the new route. Especially loopback addresses
must not appear at outgoing packets.
OK sashan@ visa@
|
|
of IFF* flags.
inputs from jmc@, ok bluhm@, visa@
|
|
by Kristof Kovacs in GitHub issue 1186.
|
|
ok mlarkin@
|
|
device "cp -r" and "rm -rf" are working in the background while
"umount -f" or "mount -ur" are executed. Then "fsck -f" is used
to verify that the file system is clean.
Currently this triggers a bug quite reliably. After the mount point
has been changed to read-only, the file system check shows that not
everything was flushed and synced properly.
|
|
str_val returns null, not NULL, if the variable isn't set. The
erroneous check means that we later tried to open("").
ok millert@ tb@ anton@ benno@
|
|
several quirks that wre absent in the old OpenBSD-specific reimplementation.
Fixes several issues with external connectors on several generations of
hardware.
ok deraadt@, benno@
|