| Age | Commit message (Collapse) | Author |
|---|
|
Allocate a 4k temp buffer on the heap instead of the stack.
Problem was introduced in arch/amd64/stand/libsa/softraid_amd64.c r1.3.
Reported by Andreas Bartelt on bugs@
ok krw@ kettenis@ yasuoka@
|
|
From pulse.purge at gmail.com
|
|
disks.
test gonzalo
ok tom krw jsing
|
|
sector size, not in 512 byte blocks. tested by gonzalo.
|
|
in 512 byte blocks. This fixes efiboot to read disklabel on 4K sector
size disk properly. tested by gonzalo@
|
|
512. The problem actually had happened on macbookair7,1. reported
and tested by gonzalo@.
|
|
defer their creation to later, so that they are owned by BUILDUSER.
This eliminates the last root-owned files in obj/ from 'make build'.
In addition, place a MACHINE == hppa test in hppa/stand/Makefile.inc
to avoid creating bogus symlinks on all other archs.
joint work with & ok natano, "let's try it" deraadt
|
|
|
|
Based on a diff from djm@
|
|
Currently, if a buffer crosses a 64KB boundary, a bounce buffer is
allocated, however it is assumed that this new buffer does not cross the
64KB boundary. In the case of i386 fdboot, it just so happens that (due to
the size of fdboot and heap allocations) UFS gets allocated a 4KB buffer
that crosses a 64KB boundary, then biosd_io() allocates a bounce buffer,
which also crosses a 64KB boundary. At this point the BIOS gets grumpy and
refuses to read from the disk.
Further clean up to come from tom@.
|
|
as discussed with jsing@ it's easier this way to ensure people have
bootblocks capable of loading the section
|
|
|
|
|
|
will allow us to bring in a MI softraid.{c,h} in lib/libsa.
|
|
sr_crypto_pbkdf (since it is useable for more than just pkcs5_pbkdf2) and
embed a struct sr_crypto_genkdf within it, rather than redeclaring the same
fields. Rename SR_CRYPTOKDFT_PBKDF2 to SR_CRYPTOKDFT_PCKS5_PBKDF2 and add
SR_CRYPTOKDFT_BCRYPT_PBKDF for upcoming changes.
|
|
|
|
|
|
couldn't interate the ConsOut modes properly on qemu.
input NONAKA Kimihiro
|
|
ok millert@ kettenis@
|
|
contains -g by default anyway
problem noted by Edgar Pettijohn (edgar (at) pettijohn-web.com)
ok millert@ kettenis@ deraadt@
|
|
from Tom Cosgrove
|
|
the com_init fix. ok beck deraadt
|
|
a short delay so that baud rate changes on the console have
a chance of working. Lots of help from theo on this one.
Makes my serial console on my APU work for an unattended
reboot instead of hanging when garbage is echoed to the tty.
ok deraadt@
|
|
tell which one we're using.
ok deraadt@, krw@
|
|
architectures besides amd64.
ok deraadt@, krw@
|
|
Report and diff from Vegar Linge Haaland
ok krw
|
|
for the disklabel. Tweak a few comments to make it more clear what
is happening. No intentional functional change.
ok yasuoka@
|
|
a 64-bit boundary (in constrast to the System V i386 ABI). Use an alignment
attribute to force 64-bit alignment of the INT64 and UINT64 types. This
makes the definitions of the EFI data structures match the expectations of
the firmware.
Fixes the issue where the 32-bit UEFI bootloader (BOOTIA32.EFI) would not
detect a GPT leading to a failure to boot.
Cluestick from John Troy.
ok krw@, yasuoka@
|
|
|
|
correctly - logically complete that now by removing MLINKS from base;
authors need only to ensure there is an entry in NAME for any function/
util being added. MLINKS will still work, and remain for perl to ease
upgrades;
ok nicm (curses) bcook (ssl)
ok schwarze, who provided a lot of feedback and assistance
ok tb natano jung
|
|
ok beck@ deraadt@
|
|
ok kettenis@ mpi@
|
|
|
|
CPU features is enough, and this enables the use of RDRAND and the
TSC fallback on CPUs from other vendors, notably AMD.
Tested in snapshots for a month.
|
|
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@
|
|
work as EFI_GUID initialisers.
ok yasuoka@ krw@
|
|
being 0; after multiply there is no perturbance to the from-disk entropy
buffer. Further investigation showed perturbance was biased towards
the lower bytes of a word. Compensate for this with a hocus pocus
bit-spreading operation which applies a result byte by byte.
discussion with kettenis, tb, mlarkin, naddy
ok naddy
|
|
|
|
|
|
the same error messages on GPT processing failures.
Add a message about failed GPT entry checksum error to both. This
was inadvertantly deleted in the switch to one entry at a time
processing.
tweaks & ok yasuoka@ jsing@
|
|
for findopenbsd_gpt().
|
|
OpenBSD partition. Just one 128-byte stack copy of a single entry.
Problems with EFI installs involving many disks reported by and fix
tested by Thomas Bohl.
Stefan Kempf pointed finger of suspicion at the addition of the 2nd
16K static buffer for softraid GPT and suggested/reviewed several
versions of the diff.
Tweaked and ok jsing@
|
|
starts at sector 1. Instead return -1, as for all other failures during
the header check.
|
|
non-static) by passing disk size as 2nd parameter instead of the
different structures holding the disk size info. The fifth copy of
gpt_chk_mbr() in fdisk is a little specialer. No functional change.
|
|
declarations of gpt_chk_mbr() to match prototypes in efidev.c and
softraid.c.
|
|
Use the size info in struct sr_boot_volume, and not the uninitialized
info in the disklabel, when checking the GPT protective MBR. Now
both GPT and MBR formatted softraid volumes actually boot.
As a bonus add prototype for gpt_chk_mbr().
|
|
MBR/GPT.
Problem found the hard way by Glenn Faustino on tech@
|
|
boot from such volumes to match being able to install to them.
Problem spotted by jcs@, fix tested by rpe@, ok jsing@
|
|
|
|
|
