For aarch64 and amd64, we don't pull in anything from libkern, so we
don't need it in our path. However, powerpc needs ucmpdi2 from libkern,
so bring it into libkboot's build and omit it from loader.kboot.
Sponsored by: Netflix
Move the startup code from kboot/kboot to kboot/libkboot and add the
necessary infrastructure for it to build. move start_arch.h, a private
header for libkboot, over as well.
Sponsored by: Netflix
In anticipation of separating the library elements from the kboot
prorgam elements, move kboot down a level into a subdirectory. There
will be libkboot and include directories in subsequent commits,
mirroring other subsystems like i386 and efi.
Sponsored by: Netflix
The `kernel` env var provides the default kernel, usually "kernel". It
may be the case that the user doesn't have a "kernel" kernel, just
"kernel.*" kernels, but have left `kernel` to the default because we
autodetect entries by default anyways.
If we're doing autodetection, take note of whether the default kernel
exists or not and remove it from the list if it doesn't and we had found
any other kernels. We avoid it in the #kernels == 1 case because
something fishy has likely happened and we should just trust the
configuration.
Reviewed by: imp, manu
Differential Revision: https://reviews.freebsd.org/D42967
This was originally used (along with FREEBSD_AOUT) to prefer the use
of ELF in various tools instead of a.out as part of the a.out to ELF
transition in the 3.x days. The last use of it was removed from
<link.h> in commit 66422f5b7a back in
2002, but various files still #define it.
Reviewed by: kevans, imp, emaste
Sponsored by: DARPA
Differential Revision: https://reviews.freebsd.org/D42964
While we're here, enable the feature in the places we detect ACPI. This
lets us side-step the existing issues and provide a path forward for
folks upgrading from previous releases that haven't updated their ESP
yet.
Let's also fix core.setACPI: the hint already indicates that the
user's disabled it more consistently than loader.acpi_disabled_by_user.
Even more, the latter is wrong because we set it by default if we did
not detect ACPI. The ACPI hint remains even when we're setting defaults
because ACPI loaded into the kernel will make some noise if it's not
hinted off, even when we didn't detect it.
imp notes that this will result in some relatively harmless noise on
platforms that don't support ACPI but aren't using the UEFI loader, as
we would enable the ACPI module for loading on them and then loader
would not be able to find it. These are non-fatal, but should probably
be fixed by just declaring support for EARLY_ACPI in those loaders since
we know they won't have ACPI early on -- punting on this for the time
being, though, in favor of providing a safer upgrade path sooner.
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D42727
liblua now provides a loader.has_feature() function to probe the loader
binary for features advertised. name => desc mappings are provided in
loader.features to get a list of all of the features loader *can*
support. core.hasFeature is provided as a shim to loader.has_feature
so that individual consumers don't need to think about the logic of the
loader module not providing has_feature; we know that means the feature
isn't enabled.
The first consumer of this will be EARLY_ACPI to advertise that the
loader binary probes for ACPI presence before the interpreter has
started, so that we know whether we can trust the presence of acpi.rsdp
as relatively authoritative. In general, it's intended to be used to
avoid breaking new scripts on older loaders within reason.
This will be used in lua as `core.hasFeature("EARLY_ACPI")`, while the
C bits of loader will `feature_enable(FEATURE_EARLY_ACPI)`.
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D42695
It would be nice to have larger boot partitions for ESPs to live in one
day. It's trivial to carve out 5M 10M or 200M when provisioning, but
logistical issues may make it hard to do it after the fact. So only warn
when the partition is > 545k. If we ever grow the boot loader larger
than that, then it will be responsible for loading the rest anyway.
Sponsored by: Netflix
Reviewed by: tsoome
Differential Revision: https://reviews.freebsd.org/D42774
Apply the following automated changes to try to eliminate
no-longer-needed sys/cdefs.h includes as well as now-empty
blank lines in a row.
Remove /^#if.*\n#endif.*\n#include\s+<sys/cdefs.h>.*\n/
Remove /\n+#include\s+<sys/cdefs.h>.*\n+#if.*\n#endif.*\n+/
Remove /\n+#if.*\n#endif.*\n+/
Remove /^#if.*\n#endif.*\n/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/types.h>/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/param.h>/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/capsicum.h>/
Sponsored by: Netflix
Remove ancient SCCS tags from the tree, automated scripting, with two
minor fixup to keep things compiling. All the common forms in the tree
were removed with a perl script.
Sponsored by: Netflix
hw.ata.wc was disconnected as part ot the 2013 cam-ification of ata. No
need to continue setting it. It's been unused in FreeBSD 10.x and newer.
Sponsored by: Netflix
When the eisa code was removed in 2017, prior to the stable/12 branch,
setting hw.eisa_slots became a nop. The oldest supported branch doesn't
have eisa at all. The need to set it manually on boot disappeared
largely by 2000...
Sponsored by: Netflix
We have no need for 5 different copies of these.
Sponsored by: Netflix
Reviewed by: rcm, kevans, andrew
Differential Revision: https://reviews.freebsd.org/D42699
To avoid a redefinition warning... This needs to be redone correctly,
but this gets amd64 building again... My amd64 environment is polluted
with something that caues earlier failures which I ignored...
Fixes: 488bc7e9a7
Sponsored by: Netflix
Old binaries do not set acpi.rsdp early enough. So when we boot with an
older loader.efi from an ESP that's not been updated, we assume there's
no ACPI on this system. This is unwise. Put a band-aide on this until we
can implement a proper 'feature' variable that the binary reports so we
can do conditionals for things like this in the future.
This is at best a rapid-response stop-gap.
Glanced at by: kevans
Sponsored by: Netflix
This is a follow-up patch to https://reviews.freebsd.org/D42459
that modifies the loader lua to use the correct loader variables
for determining ACPI availability.
This also fixes a bug where ACPI can be inadvertently disabled when
setting System Defaults at the loader menu.
Reviewed by: imp, kevans
Approved by: kp
Sponsored by: Rubicon Communications, LLC ("Netgate")
Differential Revision: https://reviews.freebsd.org/D42483
lua was previously unable to determine ACPI presence because this
probing was postponed until the final loading and execution of the
kernel.
This patch resolves that by detecting ACPI early (similar to
the order of operations in the legacy i386 loader).
Reviewed by: kevans
Approved by: kp
Sponsored by: Rubicon Communications, LLC ("Netgate")
Differential Revision: https://reviews.freebsd.org/D42459
Preemptively address a collision with LIBFDT (to be added in the future)
from src.libnames.mk, which gets included via bsd.progs.mk. No
functional change intended.
Reviewed by: imp
MFC after: 1 week
Sponsored by: Innovate UK
Differential Revision: https://reviews.freebsd.org/D42486
This feature is marked as ZFEATURE_FLAG_READONLY_COMPAT and so
irrelevant for read-only pool imports by the loader:
"com.delphix:spacemap_v2"
This should cause no functional changes, just a code cleanup.
I'm sorry, missed it in previous commit.
MFC after: 2 months
These features are marked as ZFEATURE_FLAG_READONLY_COMPAT and so
irrelevant for read-only pool imports by the loader:
"com.datto:resilver_defer",
"com.delphix:obsolete_counts",
"com.delphix:spacemap_histogram",
"com.delphix:zpool_checkpoint",
"com.intel:allocation_classes",
"org.zfsonlinux:allocation_classes"
This should cause no functional changes, just a code cleanup.
MFC after: 2 months
To read/update the boot loader nvstore, we always need to call
zfs_attach_nvstore() regardless of whether we use bootonce key
in nvstore or the bootfs property of the pool. The call was
unintentionally left in the block of code that is processed
only when bootonce key is present.
In particular this fixes broken 'nextboot -k'.
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D41795
Fixes: e3e2681d0e
The module entries should generally allow whatever is allowed as an
env_var in the pattern table. Notably, we're missing periods which
would allow proper entries for .dtb files in loader.conf that don't need
to specify a module_name entry for it.
%d in this expression is actually redundant as %w is actually
"all alphanumerics," but I've included it for now to match the env_var
entry. We should really remove it from both.
Reported by: "aribi" on the forums via allanjude@
MFC after: 1 week
9600 was a standard baud rate decades ago, but 115200 is now more common
so choose defaults that are useful to the largest number of users.
Note that boot0sio does not support rates above 9600 so it remains
unchanged.
Reviewed by: bz, imp, manu
Relnotes: Yes
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D36295
We plan to increase the default serial rate to 115200 (see review
D36295) but early boot components that use BIOS interfaces do not
support higher rates. Add a note to that effect.
Reported by: imp
Sponsored by: The FreeBSD Foundation
By not exporting _start, we get various warnings of the form:
ld: warning: cannot find entry symbol _start
Note that in practice these don't matter because we manually construct
our PE header and use objcopy -O binary, so the entry point is set to
whatever we put explicitly in the PE header, but we should still do the
right thing and silence these warnings.
This was found in CheriBSD, where bsd.prog.mk has similar logic to
bsd.lib.mk and sets -Wl,--(no-)fatal-warnings based on LD_FATAL_WARNINGS
(unlike FreeBSD which only does so in the latter).
The removal of the sparc64 support in February 2020 obsoleted the
VTOC8 partitioning scheme as no other FreeBSD platform makes use
of it. Moreover, the code is bitrotting as nothing defines e. g.
LOADER_VTOC8_SUPPORT any more and, thus, should go now, too. With
this change, the following commits are reverted as far as VTOC8
is concerned and parts haven't already previously been deleted
along with prior sparc64 removals:
094fcb157da7d366e958ba8d50d08b
The alignment example d9711c28ef
added to the VTOC8 section of gpart.8 is folded into the MBR one.
This should finally conclude the deorbit of sparc64-specific bits.
We had joy, we had fun
we ran Unix on a Sun.
But that source and the song
of FreeBSD have all gone.
Credits to Michael Bueker for the original "Unix on a Sun" and Rod
McKuen for the "Seasons in the Sun" lyrics.
In most cases, usage does not return, so mark them as __dead2. For the
cases where they do return, they have not been marked __dead2.
Reviewed by: imp
Pull Request: https://github.com/freebsd/freebsd-src/pull/735
As a result, it is only really checking for the word Drive, making
"Drive" appended to anything else considered for efi env.
Reviewed by: imp, kevans
Pull Request: https://github.com/freebsd/freebsd-src/pull/738
Before this change we would only pass the bootonce dataset name
to the environment for the next loader, while actually reading
the next stage loader from the 'bootfs' dataset, not the bootonce
dataset.
Another problem fixed by this change is a boot from a configuration
when bootonce attribute is present, but 'bootfs' property is not set.
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D40389
First check if bootonce is configured and if it is, then change currdev
accordingly and after that do the sanity check. This fixes boot in a
situation when ZFS pool doesn't have the "bootfs" property, but has
bootonce attribute set. A strange, but legitimate case.
Reviewed by: tsoome, imp
Differential Revision: https://reviews.freebsd.org/D40388
We don't keep comc_port around anymore if the console's not present, but
some things might still try to set one of the environment variables we
hook. In particular, one need not even set efi_com_port/efi_com_speed
in loader.conf; loader may do it itself and induce the crash if ConOut
depicts an available uart.
Probably reported by: dch
OK for now: imp
If we can't find the serial port, free comc_port and return early. The
serial port just isn't there.
Sponsored by: Netflix
Reviewed by: kevans
Differential Revision: https://reviews.freebsd.org/D40223
Make the compat code more correct by probing using the eficom console
structure, not the comconsole one.
Sponsored by: Netflix
Reviewed by: kevans
Differential Revision: https://reviews.freebsd.org/D40222
If the machine isn't hyperv on amd64, then this driver fails the probe
and will do nothing further now, even if explicitly listed in a config.
Sponsored by: Netflix
Reviewed by: kevans
Differential Revision: https://reviews.freebsd.org/D40221
Don't even have a comc_port when the port doesn't exist: always free it
if we probe that it's not there. Also, when it's not present, clear the
flags indicating presence to avoid disturbing other flags.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D40220
It would be nice to make decisions early in boot, about maybe consoles,
based on smbios variables. Set them just after we setup the archsw so we
can use them everywhere.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D40219
UEFI device path may be path to one device, or concatenated list of instances
to different devices (textually represented as comma separated list).
Provide generic function to get next instance from device path.
Returns next instance or end node.
The use case is like:
EFI_DEVICE_PATH *node = (EFI_DEVICE_PATH *)buf;
while (!IsDevicePathEnd(node)) {
process(node);
node = efi_devpath_next_instance(node);
}
Where buf is pointing to either single device path or
concatenated list of device paths (such as from ConIn or ConOut).
Reviewers: imp
Differential Revision: https://reviews.freebsd.org/D40081
Commit e32fecd0c2 intended to skip installing all but one copy of
each loader variant's help file, but accidentally skipped all copies for
the userboot help file. (Other loaders install help files via the _simp
variant, but there is is no userboot_simp.)
PR: 271178
Fixes: e32fecd0c2 ("loader: install help files only once")
Sponsored by: The FreeBSD Foundation
The SPDX folks have obsoleted the BSD-2-Clause-FreeBSD identifier. Catch
up to that fact and revert to their recommended match of BSD-2-Clause.
Discussed with: pfg
MFC After: 3 days
Sponsored by: Netflix
Remove the i386 ifdefs and files. It never worked.
Sponsored by: Netflix
Reviewed by: manu, tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D40012
Only set the baudrate when it is different than what the device has
reported. In addition, pass in the args to effect no change to the other
parameters to the serial port. Some EFI firmware gets cranky when you
set them to the same value, so avoid doing so (we likely can remove the
HyperV workaround with this fix, but I kept it in place). Add comments
to the code for why we do this too.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D40010
Remove rtsdtr_off. It's basically unused. Expand its meaning, but put
changing flow control to under an ifdef. We shouldn't set it unless
we're sure we need to do so. UEFI normally initializes the device
correctly, and we should avoid needless changes that aren't user
requested.
Sponsored by: Netflix
Reviewed by: tsoome
Differential Revision: https://reviews.freebsd.org/D40009
Rename efiserialc to eficom.c and move it to libefi. Remove
loader.efi.h, since it's not needed. It's architecture independent
(though how we use it might vary). Drivers also belong in libfoo
in the boot loader: all the BIOS drivers are in i386/libi386 and
the console driver is in efi/libefi.
Sponsored by: Netflix
Reviewed by: tsoome
Differential Revision: https://reviews.freebsd.org/D40007
Add a compat shim for the "comconsole" name so that people with a
"console=comconsole" in their loader.conf on aarch64 will continue to
work (though with a warning).
This is only aarch64: it will never be there for amd64 (where comconsole
always means talk to the hardware directly). To do that is too hard.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D39983
Fix the 'renaming kludge' that we absolutely cannot do going forward
(it's cost us days of engineering time).
console=comconsole talks to the hardware directly. This is available
only on amd64. It is not available anywhere else (and so requires
changes for people doing comconsole on aarch64)
console=eficom talks to the console via EFI protocols. It's available
on amd64, aarch64 and riscv64. It's the first port that we find, though
it can be overriden by efi_com_port (which should be set to the UID of
the serial port, not the I/O port, despite the name). devinfo -v
will give the UID to uartX mapping.
This is an incompatible change for HYPER-V on amd64. It only works with
eficom console, so you'll need to change your configuration in
loader.conf. No compatibility hack will ever be provided for this (since
it requires renamig, which the loader cannot reliably do).
It's also an incompatible change for aarch64. comconsole will need to
change to eficom. There might be a comconsole "shim" for this.
All the interlock to keep only eficom and comconsole from both attaching
have been removed.
RelNotes: Yes
Sponsored by: Netflix
Discussed with: kevans
Differential Revision: https://reviews.freebsd.org/D39982
If a file is specified in loader_conf_files that ends in '.lua', lualoader
will now load and execute that file. These may be used in place of a
traditional loader.conf to use more complicated logic, where some values
may be set based on others or based on the environment that the C bits has
left us with.
Lua scripts are run in a limited environment. In particular, it does not get
access to any modules or, in-fact, anything except environment variable.
A config.buildenv hook has been added so that a local module can add
whatever it may need to to the environment.
When a global var is set in the lua script, it does not immediately alter
the loader environment. Instead, the script's environment is initially
empty and processed only if the whole script executes successfully.
Effectively, a lua configuration file either takes effect or it does not,
an error will not leave it in a half-baked state.
Reviewed by: bcr (manpages), imp
Differential Revision: https://reviews.freebsd.org/D28450
Every file should be installed exactly once by `make installworld`.
This is especially important for pkgbase.
Loader help files were being installed by each loader variant (e.g.,
the simp, lua, and 4th EFI loaders). Add a (slightly hacky) mechanism
to skip installing help files for all but one variant.
PR: 271178
Reviewed by: imp
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D40021
com.klarasystems:vdev_zaps_v2 is a new feature that the last OpenZFS
import brought in. It needs to be on the list of supported features, but
that update didn't happen so I woke up to a mailbox with multiple
complaints.
CirrusCI test to boot twice with a zpool update inbetween coming later
today.
Sponsored by: Netflix
Add some more sanity checks to make sure we don't march off the end of
the table. Typically, smbios structures are well formed, or Windows
wouldn't boot. Sometimes they aren't, and this at least fails safe.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D39794
This code was written on x86 where unaligned accesses were
easy. LinuxBoot running on aarch64 uses mmap of /dev/mem to read the
smbios table. Linux's mapping of this memory doesn't allow the normal
unaligned fixup, so we get a bus error instead. We can't use the more
natural le16dec and friends because they optimize into a single,
unaligned memory load. We don't see this issue on aarch64 UEFI because
memory is mapped such that unaligned accesses are fixed up.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D39793
Add support for getting smbios from /sys/firmware/efi/systab, if
any. Add ptov mapping that uses mmap on /dev/mem to do the mapping with
64k pages (usually we only need 1 or two mappings).
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D39792
Fix an off-by-one error that would mean we'd get stuck on the newline if
ACPI= wasn't first.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D39817
Add one ifdef to upstrem code and get rid of compiling the horrible
checked-in aarch64 assembler for the boot loader that the loader will
never use. I'll attempt to upstream this and adjust as needed.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D39897
There's plenty of stack in kboot, so use it here rather than the
malloc/free dance.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39416
We have plenty of stack in the EFI case, so use it instead of the
complicated malloc / free dance.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39415
We have way more than 8k of stack for the current value of the zfs
bootonce attribute. Allocate buf on the stack rather than the
complicated malloc / free dance.
Sponsored by: Netflix
Reviewed by: tsoome, kevans, jhb
Differential Revision: https://reviews.freebsd.org/D39414
Implement ZFS bootonce protocol. We pass zfs-bootonce=t to the next boot
stage as a command line argument. Unlike zfsboot -> loader handoff in
the BIOS case, we don't use the OS_BOOTONCE_USED. This would require
modifications to loader.efi which would only server to make it more
complicated. Instead, use the command line parsing interface for the
boot1.efi -> loader.efi to pass in the zfs-bootonce kenv that will be
needed by rc.d/zfsbe to activate the BE if boot progresses that far.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39412
Lookup the spa and pass it into zfs_get_bootonce_spa to process the boot
once protocol.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39411
Refactor zfs_set_bootenv to split out the lookup of spa from the
rest. zfs_set_bootenv_spa flushes the benv to the vdevs and updates the
cached benv.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39410
Create a new interface to zfs_get_bootenv called zfs_get_bootenv_spa
which takes a spa instead of a void * (effectively a devdesc *). Use
that in zfs_get_bootenv.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39409
zfsimpl.c doesn't know about devdesc at all, but zfs.c does. Move it to
zfs.c, which is the only user. Keep it static for now, but it could be
exposed later if something else were to need it.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39408
Sometimes filesystem modules need to pass details of the state of the
filesystem to later stages of a boot. Provide a generic method to do
so. We'll add them after any env variables set in our config files.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D39407
SEGDESC_t needs to be PACKED
there is no status in t_PXENV_UNDI_MCAST_ADDRESS
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D39799
This commentary was carried over from the x86 version of the same code,
but has actually been inaccurate for a while now. As of FreeBSD 12.x,
all environments are used unless they disable each other. See
39d44f7f15 ("kern_environment: use any provided environments [...]")
for details.
Reviewed by: imp
Differentiala Revision: https://reviews.freebsd.org/D35695
Use 1000 * major + minor when calculating the version number that
gets set in the Ficl environment or lua loader property. This allows
for more room if the minor number needs to go above 9.
Add loader.version property to lua loader.
Reviewed by: imp
Obtained from: Juniper Networks, Inc.
Differential Revision: https://reviews.freebsd.org/D39631
The latest import of openzfs broke the hacks that we used to omit the
special registers being used on arm64. Add snarky note documenting this
situation since it's a mess now since the hack was only partially
undone, leaving behind a mess.
Sponsored by: Netflix
- Mark the file as an executable in the COFF header.
- Provide separate .text and .data sections.
- Provide sane file and section alignment values. These values are the
defaults defined in the PE specification.
- Set appropriate characteristics for each of .text and .data.
This is required for the MS devkit to load our UEFI image.
Obtained from: OpenBSD via allanjude
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D37765
This shouldn't be an option (and I added it in the first place back in
4ae4202e70 and
83f4b92050). However, unlike the other
knobs I added back then, this really shouldn't be a knob since it is
hardcoded in the source.
This really shouldn't even be an option given it is hardcoded as a
constant named ORIGIN in the assembly. mbr.S also uses 0x600 and
hardcodes it in both the assembly and the Makefile.
Summary:
Include details from the SMBIOS 3 specification and some additional
details for SMBIOS 2.1.
Obtained from: Juniper Networks, Inc.
Reviewers: jmg, manu
Subscribers: imp, dab
Differential Revision: https://reviews.freebsd.org/D39635
Clang 16 introduced a warning about single bit bitfields in structs,
which is triggered by a declaration in libsa's tftp.c:
stand/libsa/tftp.c:382:20: error: implicit truncation from 'int' to a one-bit wide bit-field changes value from 1 to -1 [-Werror,-Wsingle-bit-bitfield-constant-conversion]
h->islastblock = 1; /* very short file */
^ ~
stand/libsa/tftp.c:432:18: error: implicit truncation from 'int' to a one-bit wide bit-field changes value from 1 to -1 [-Werror,-Wsingle-bit-bitfield-constant-conversion]
h->islastblock = 1; /* EOF */
^ ~
Signed one-bit bitfields can only have values -1 and 0, but the intent
here is to use the field as a boolean, so make it unsigned.
MFC after: 3 days
In the lua loader, if one sets a password in loader.conf, the
autoboot_delay timer will start before the menu is displayed. One can
interrupt the autoboot and bring up the menu by entering a keyboard
character before the timer expires.
If this is done a prompt for the password is displayed. Entering the
password will bring up the menu, but the timer will again start and
another keyboard character must be entered or autoboot will abort the
menu and boot the system.
PR: 265472
Reviewed by: kevans
MFC after: 3 days
Usually the kernel is loaded later, but there are circumstances where it
could have been loaded earlier than changing BEs. Unload anything that
is already there so that we know we're using artifacts from the proper
environment.
PR: 265471
Reviewed by: kevans
MFC after: 3 days
It may be the case that we need to set hw.uart.console manually in some
scenarios that comconsole can't necessarily support. Avoid clobbering
hw.uart.console unless we've actually selected comconsole so that one
could at least get kernel console output..
Discussed with: imp
Sponsored by: Zenith Electronics LLC
Sponsored by: Klara, Inc.
Notable upstream pull request merges:
#12194 Fix short-lived txg caused by autotrim
#13368 ZFS_IOC_COUNT_FILLED does unnecessary txg_wait_synced()
#13392 Implementation of block cloning for ZFS
#13741 SHA2 reworking and API for iterating over multiple implementations
#14282 Sync thread should avoid holding the spa config write lock
when possible
#14283 txg_sync should handle write errors in ZIL
#14359 More adaptive ARC eviction
#14469 Fix NULL pointer dereference in zio_ready()
#14479 zfs redact fails when dnodesize=auto
#14496 improve error message of zfs redact
#14500 Skip memory allocation when compressing holes
#14501 FreeBSD: don't verify recycled vnode for zfs control directory
#14502 partially revert PR 14304 (eee9362a7)
#14509 Fix per-jail zfs.mount_snapshot setting
#14514 Fix data race between zil_commit() and zil_suspend()
#14516 System-wide speculative prefetch limit
#14517 Use rw_tryupgrade() in dmu_bonus_hold_by_dnode()
#14519 Do not hold spa_config in ZIL while blocked on IO
#14523 Move dmu_buf_rele() after dsl_dataset_sync_done()
#14524 Ignore too large stack in case of dsl_deadlist_merge
#14526 Use .section .rodata instead of .rodata on FreeBSD
#14528 ICP: AES-GCM: Refactor gcm_clear_ctx()
#14529 ICP: AES-GCM: Unify gcm_init_ctx() and gmac_init_ctx()
#14532 Handle unexpected errors in zil_lwb_commit() without ASSERT()
#14544 icp: Prevent compilers from optimizing away memset()
in gcm_clear_ctx()
#14546 Revert zfeature_active() to static
#14556 Remove bad kmem_free() oversight from previous zfsdev_state_list
patch
#14563 Optimize the is_l2cacheable functions
#14565 FreeBSD: zfs_znode_alloc: lock the vnode earlier
#14566 FreeBSD: fix false assert in cache_vop_rmdir when replaying ZIL
#14567 spl: Add cmn_err_once() to log a message only on the first call
#14568 Fix incremental receive silently failing for recursive sends
#14569 Restore ASMABI and other Unify work
#14576 Fix detection of IBM Power8 machines (ISA 2.07)
#14577 Better handling for future crypto parameters
#14600 zcommon: Refactor FPU state handling in fletcher4
#14603 Fix prefetching of indirect blocks while destroying
#14633 Fixes in persistent error log
#14639 FreeBSD: Remove extra arc_reduce_target_size() call
#14641 Additional limits on hole reporting
#14649 Drop lying to the compiler in the fletcher4 code
#14652 panic loop when removing slog device
#14653 Update vdev state for spare vdev
#14655 Fix cloning into already dirty dbufs
#14678 Revert "Do not hold spa_config in ZIL while blocked on IO"
Obtained from: OpenZFS
OpenZFS commit: 431083f75b
The hang basically bricks a physical box and it can be recovered
only if you are able to boot from alternate media. This isn't a
perfect fix, but throw it in before loader experts decide on
proper one.
Submitted by: whu
Fixes: 927358dd98
UEFI provides ConIn/ConOut handles for consoles that it supports,
which include the text-video and serial ports. When the serial port
is available, use the UEFI driver instead of direct io-port accesses
to avoid conflicts between the firmware and direct hardware access, as
happens on Hyper-V (Azure) setups.
This change enables efiserialio to be built for efi-amd64 and has
higher order priority vs comconsole, and only uses efiserialio
if the hypervisor is Hyper-V. When efiserialio successfully
probes, it will set efi_comconsole_avail=true which will prevent
comconsole from probing in this setup.
Tested on Hyper-V, ESXi and Azure VMs.
PR: 264267
Reviewed by: kevans, whu
Tested by: whu
Obtained from: Rubicon Communications, LLC (Netgate)
MFC after: 2 weeks
Sponsored by: Rubicon Communications, LLC (Netgate)
When booting from a snapshot we need to follow a different code path
to turn the objset ID into the name, and for forward lookups we need
to walk the parent's snapnames_zap.
With this, it is possible to set the pools BOOTFS property to a
snapshot and boot with a read-only filesystem of that snapshot.
Reviewed by: tsoome, rew, imp
Sponsored By: Beckhoff Automation GmbH & Co. KG
Sponsored By: Klara, Inc.
Differential Revision: https://reviews.freebsd.org/D38600
If there's a kboot.conf, prase it after the command line args are
parsed. It's not always easy to get all the right command line args
depending on the environment. Allow an escape hatch. While we can't do
everything one might like in this file, we can do enough.
Sponsored by: Netflix
We were assuming that hostdisk_override was both a directory and a
file, which is not going to work very well. It's supposed to be a
single file, so recode it as such. Simplify erorr handling a little as
well and fix a return type-mismatch that doesn't matter for the
generated code (return NULL is the same as return false in this
context)
Sponsored by: Netflix
MIN works for any type, while min() is only for integers. So we were
rounding down to 0 since that's the size of 4GB truncated to an int.
Sponsored by: Netflix
When we're running on a FreeBSD host, we can't open /proc/iomem. So,
for now, just assume that we have 32GB of ram starting at 4GB.
Sponsored by: Netflix
Replace a cast '0' for a null pointers with NULL
Replace a 'goto loop' with a do-while loop in ufs and ext2fs.
Cast cp pointer to uintptr_t to test to see if it's aligned rather than long.
[ minor tweaks based on my & hps' review, reworded commit message ]
Reviewed by: imp, hps
Pull Request: https://github.com/freebsd/freebsd-src/pull/547
Summary:
Under QEMU on arm64 systems, the smbios table is above 4GB
requiring a 64-bit address to access.
Reviewers: manu
Subscribers: imp, bcran, dab
Differential Revision: https://reviews.freebsd.org/D38721
To make auditing and debugging easier, export foo_load=XXX foo_name=yyy
etc to the loader env that we export to the kernel.
Original by: dhw
Sponsored by: Netflix
Discussed with: kevans
Differential Revsion: https://reviews.freebsd.org/D38466
Merge commit '755d9301ca89f02956fd17858b9d4d821ab5c972' from the
vendor branch. This updates us from lua 5.4.2 to 5.4.4.
In addition, it switches around how we flavor liblua for the boot loader
and flua. This is done to reduce diffs with upstream and make it easier
to import new versions (the current method has too many conflicts to
resolve by hand): we include luaconf.local.h from luaconf.h (the only
change to this file is now that #include at the end). We then define
what we need to: for flua (which does very little) and one for stand
(which creates the new FLOAT type out of int64).
Sponsored by: Netflix
Address two issues with current help file logic:
The existing condition prevents the common help file from being
installed when there are no additional help files defined. This results
in no loader.help on EFI platforms, for example.
Second, due to the fact that we build and install multiple loader types,
each successive install will clobber the previous loader.help. The
result is that we could lose type-specific commands, or possibly list
them in loaders that do not have such commands.
Instead, give each loader type a uniquely named help file. The EFI
loader will look for /boot/loader.help.efi, userboot will look for
/boot/loader.help.userboot, etc. The interpreter variant has no effect
on which help file is loaded.
This leaves the old /boot/loader.help unused.
Some credit for the final approach goes to Mathieu <sigsys@gmail.com>
for their version of the fix in https://reviews.freebsd.org/D22951.
PR: 267134
Reported by: Daniel O'Connor <darius@dons.net.au>
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D28591
Keep track of how much is used in the segment as we allocate it to the
application. Set memsz to 0 first, and increment it as used. Adjust the
bufsz before we call kexec so the kernel copies the right amount (it's
an error for bufsz to be bigger than memsz, so we set them == when we
retrieve the segment). Make sure we round to the page size, otherwise
kexec_load gets cranky.
Sponsored by: Netflix
Reviewed by: tsoome
Differential Revision: https://reviews.freebsd.org/D38315
Allocate a huge segment for the first kexec_load segments. We limit the
lessor of:
allocation to the size of the remaining memory segment
45% of available memory
95% of the memory we can allocate
This allows us to have really large RAM disks. We likely need to limit
this to the amount we actually used, though, since this can be a lot of
memory.
We have to do this complicated calculation for a few reasons: First, we
need 2 copies of the loaded kernel in the memory: The kernel can copy
everything to a temporary buffer. Next, malloc (via mmap) is limited to
a certain amount due to over commit, so we have to not allocate all we
can (only most of what we can).
Sponsored by: Netflix
Reviewed by: tsoome
Differential Revision: https://reviews.freebsd.org/D38314
kboot_get_phys_load_segment is defined in kboot.h, so remove them from
the .c files.
Sponsored by: Netflix
Reviewed by: tsoome
Differential Revision: https://reviews.freebsd.org/D38310
Try to open /dev/mem to read in the UEFI memory map. If we can't, then
we'll read it in the trampoline.
Retain reading in /proc/iomem to find reserved areas in Linux. We need
to know them for good places to put the kernel. These are not reflected
in the UEFI memory map. However, we should not adjust the UEFI memory
map since these reserved areas of the Linux kernel are free to be used
once we enter the kexec trampoline...
Sponsored by: Netflix
Reviewed by: tsoome, kevans, andrew
Differential Revision: https://reviews.freebsd.org/D38264
There's no need for an arch pointer to get segments. We can call the
routine directly since we don't need this code to be called from
different context where a pointer is needed.
Sponsored by: Netflix
Reviewed by: kevans, andrew
Differential Revision: https://reviews.freebsd.org/D38266
A number of bug fixes to loading kernels and modules on aarch64 and amd64.
Fix offset calcuations.
Add a number of debugs, commented out for now (will GC them in the future)
With this, and the MD aarch64 commands, we can linux boot in qemu and on
real hardware.
Sponsored by: Netflix
Reviewed by: kevans
Differential Revision: https://reviews.freebsd.org/D38261
Copy more of the necessary state for FreeBSD to boot:
o Copy EFI memory tables
o Create custom page tables needed for the kernel to find itself
o Simplify the passing of args to the trampoline by putting them
on the stack rather than in dedicated memory.
This is only partially successful... we get only part way through the
amd64 startup code before dying. However, it's much further than before
the changes.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D38259
Update exec.c (copyied from efi/loader/arch/arm64/exec.c) to allow
execution of aarch64 kernels. This includes a new trampoline code that
handles copying the UEFI memory map, if available from the Linux FDT
provided PA. This is a complete implementation now, able to boot from
the LinuxBoot environment on an aarch64 server that only offers
LinuxBoot (though a workaround for the gicv3 inability to re-init is not
yet in FreeBSD). Many 'fit and finish' issues will be addressed in
subsequent commits.
Sponsored by: Netflix
Reviewed by: tsoome, kevans, andrew
Differential Revision: https://reviews.freebsd.org/D38258
Connect efi's bootinfo.c to the kboot build, and adjust to use
the kboot specific routines.
The getrootmount() call is independent of EFI. Remove ifdefs so it's
called for kboot too.
The differences between the kboot and efi bootinfo.c files are now tiny.
This could use some more refactoring, but this is a working checkpoint.
Sponsored by: Netflix
Reviewed by: tsoome
Differential Revision: https://reviews.freebsd.org/D38350
Since aarch64 is different, it needs a different smap. We first see if
we have the PA of the table from the FDT info. If so, we copy that and
quit. Otherwise, we do the best we can in translating the /proc/iomap
into EFI Memory Table format.
We also send the system table to the kernel.
Sponsored by: Netflix
Reviewed by: kevans
Differential Revision: https://reviews.freebsd.org/D38255
Copy the EFI memory tables we were able to get into the MODINFOMD_SMAP
metadata area for the kernel.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D38254
It's just a stub, since the kernel learns of memory via FDT.
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D38253
Each architecture will soon be required to provide this to load memory
maps as metadata for the platforms that require it (or a stub function
for those that don't).
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D38252
Now that all architectures provide this, enumerate the platform's memory
before we go to interact(). This needs to be done only once, but relies
on our ability to open host: files on some platforms, so it needs to be
done after devinit().
Sponsored by: Netflix
Reviewed by: tsoome, kevans
Differential Revision: https://reviews.freebsd.org/D38251
Move memory enumeration to the enumerate_memory_arch(), tweak the code a
bit to make that fit into that framework.
Also fix a bug in the name of the end location. The old code never found
memory (though amd64 doesn't yet work, this lead to using fallback
addresses that were good enough for QEMU...).
Sponsored by: Netflix
Reviewed by: kevans
Differential Revision: https://reviews.freebsd.org/D38250
We have an odd situation with aarch64 memory enumeration. The fdt that
we can get has a PA of the UEFI memory map, as modified by the current
running Linux kernel so it can retain those pages it needs for EFI and
other services. We have to pass in this EFI tablem but don't have access
to it in the boot loader. We do in the trampoline code, so a forthcoming
commit will copy it there for the kernel to use. All for want of /dev/mem
in the target environment sometimes.
However, we also have to find a place to load the kernel, so we have to
fallback to /proc/iomem when we can't read the UEFI memory map directly
from /dev/mem. It will give us good enough results to do this task. This
table isn't quite suitable to be converted to the EFI table, so we use
both methods. We'll fall back to this method also if there's no EFI
table advertised in the fdt. There's no /sys file on aarch64 that has
this information, hence using the old-style /proc/iomem. We're unlikely
to work if there's no EFI, though.
Note: The underlying Linux mechanism is different than the amd64 method
which seems like it should be MI, but unimplemented on aarch64.
Sponsored by: Netflix
Discussed with: kevans
Differential Revision: https://reviews.freebsd.org/D38249