Commit Graph

904 Commits

Author SHA1 Message Date
Toomas Soome
094b40659f loader: --gc-sections needs sections to work with
--gc-sections is not really useful unless we generate sections with
-ffunction-sections -fdata-sections

While there, i386/loader would win from --gc-sections too.
2019-09-08 19:39:45 +00:00
Toomas Soome
21388f5cf8 loader.efi: UEFI text mode background colors are only using 3 bits
Need to mask background colors to pass allowed value.
2019-09-06 19:30:23 +00:00
Justin Hibbits
2f6823d4c1 Loader: Add load offset to powerpc kernel entry point
Summary:
There is logic in ELF loadimage() to relocate kernels, but currently
only type ET_EXEC.  PowerPC kernels are ET_DYN, and can be relocated anywhere.
Add the load offset to kernel entry points on this platform.

Reviewed by:	imp, ian
Differential Revision:	https://reviews.freebsd.org/D21286
2019-09-06 02:28:03 +00:00
Toomas Soome
56758831fe loader: use teken teminal emulator for x86 and uefi
Replace mini cons25 emulator with teken, this does enable us proper console
terminal for loader and will make it possible to implement different
back end callbacks to draw to screen.

At this time we still only "draw" in text mode.
2019-09-05 22:15:50 +00:00
Toomas Soome
938cae327c ficl: add xemit word
While emit will output one byte on screen, the xemit will output xchar.

See: http://forth-standard.org/standard/xchar/XEMIT
2019-09-05 15:06:30 +00:00
Toomas Soome
0bd5d36798 ficl: add uIsGreater word
For some reason we have u< but not u>, fix it.
2019-09-05 09:04:48 +00:00
Rebecca Cran
9a2b7c5fce The efifat files are no longer used: remove the code to build them
Reviewed by:	imp, tsoome, emaste
Differential Revision:	https://reviews.freebsd.org/D20562
2019-09-04 20:55:48 +00:00
Toomas Soome
05b24e8617 loader.efi: use and prefer coninex interface
Add support for EFI_SIMPLE_TEXT_INPUT_EX_PROTOCOL.
2019-09-02 11:04:17 +00:00
Toomas Soome
591d5f0ef5 loader.efi: some systems do not translate scan code 0x8 to backspace
Add scancode translation for backspace.
2019-09-02 10:45:10 +00:00
Brandon Bergren
9ae631858e Move CAS check in powerpc64 ofw loader until after the PVR check.
This unbreaks using the powerpc64 loader on a 32-bit processor.

Approved by: jhibbits (mentor)
Differential Revision: https://reviews.freebsd.org/D21297
2019-09-01 18:26:21 +00:00
Toomas Soome
fd7e6079bc loader.efi: use shift 16 in efipart as it is max ashift supported by zfs 2019-08-31 06:24:09 +00:00
Toomas Soome
9f7b00f230 qemu-system-aarch64 does list block device with very large block size
Also insert Vendor device as "parent" disk (found in
qemu-system-aarch64 -cpu cortex-a57).

This does fix lsdev in loader.efi on this platform.
2019-08-30 20:54:30 +00:00
Kyle Evans
e1aa7047af stand: boot2: fix amd64-xtoolchain-gcc build
-Wno-missing-declarations is the GCC equivalent of
-Wno-missing-prototypes... this was overlooked in r351135.

MFC after:	3 days
X-MFC-With:	r351135
2019-08-23 01:16:12 +00:00
Kyle Evans
da8147c3c5 libsa: mips: fix typo that had slipped into the diff on local machine
MFC after:	1 week
X-MFC-With:	r351408
2019-08-22 21:49:13 +00:00
Kyle Evans
428925b5f2 libsa: mips: use _JB_* from machine/asm.h, remove regnum dep
This brings the libsa/mips _setjmp implementation closer to parity with the
libc version.

Reviewed by:	imp, jhb
MFC after:	1 week
Differential Revision:	https://reviews.freebsd.org/D21330
2019-08-22 21:42:11 +00:00
Toomas Soome
a15a23234d loader.efi: efipart should be more careful about constructing block device lists
The cd handles should be collected as list of partitions. Some systems also
provide base name for block device (like PciRoot(0x0)/Pci(0x5,0x0)), we need
to be careful about those.

To make out life a bit easier, we prepare not just an array of handles, but we
allocate pdinfo struct for each handle and devicepath, then we can simplify
our work to sort the devices.

Differential Revision:	https://reviews.freebsd.org/D21187
2019-08-20 19:31:11 +00:00
Tom Jones
ec18da7c04 Reduce size of EFI_STAGING_SIZE to 32 on arm
Reduce the size of the EFI_STAGING area we allocate on arm to 32. On arm SBC
such as the NanoPi-NEOLTS the staging area allocation will fail on the 256MB
model with a staging size of 64.

Reviewed by:    bcran, manu
Approved by:    bz (mentor)
MFC after:      2 weeks
Differential Revision: https://reviews.freebsd.org/D21016
2019-08-17 06:47:43 +00:00
Kyle Evans
80746f9f83 stand: gptboot: fix build with xtoolchain-llvm90
ufsread.c grows a dependency on __ashldi3 with llvm90. For gptboot, just
start pulling in ashldi3.c ashrdi3.c lshrdi3.c into libsa for all archs as
the number of archs requiring one or more of them keeps growing. qdivrem.c
and quad.h can be trivially kicked out of libsa if we start pulling these
from compiler-rt as qdivrem was only used to implement umoddi3, divdi3,
moddi3 (also in qdivrem.c).

Reviewed by:	imp
MFC after:	1 week
Differential Revision:	https://reviews.freebsd.org/D21291
2019-08-16 20:09:20 +00:00
Kyle Evans
37ac567ea7 stand: boot2: fix build with xtoolchain-llvm90
ufsread.c grows a dependency on __ashldi3 with llvm90. Grab ashldi3.c out of
compiler-rt rather than trying to link against libsa (for now).
-Wno-missing-prototypes is necessary to compile ashldi3.c standalone.

Reviewed by:	imp
MFC after:	1 week
Differential Revision:	https://reviews.freebsd.org/D21291
2019-08-16 20:07:43 +00:00
Kyle Evans
80335781ad stand: push LIBC_SRC up into defs.mk
Other parts of stand/ that don't use libsa will need to grab bits from libc
shortly. Push LIBC_SRC up to defs.mk in advance of this so that they can use
it, and rename it to LIBCSRC to match the convention of the rest of the *SRC
variables in this file.

Reviewed by:	imp
MFC after:	1 week
Differential Revision:	https://reviews.freebsd.org/D21282
2019-08-16 13:22:57 +00:00
Ed Maste
0a062e5ff6 gptzfsboot: remove CLANG_NO_IAS
Many components under stand/ had CLANG_NO_IAS added when Clang's
Integrated Assembler did not handle .codeNN directives.  Clang gained
support quite some time ago, so we can build stand/ with IAS.

In some cases there were small differences in generated object output.
In the case of gptzfsboot however using GNU as or Clang IAS to assemble
gptldr.S resulted in identical final gptzfsboot binary output.

Sponsored by:	The FreeBSD Foundation
Differential Revision:	https://reviews.freebsd.org/D11115
2019-08-15 18:43:01 +00:00
Ed Maste
aa2dd958b4 stand: remove CLANG_NO_IAS from pxeldr
Many components under stand/ had CLANG_NO_IAS added when Clang's
Integrated Assembler (IAS) did not handle .codeNN directives.  Clang
gained support quite some time ago, and we can now build stand/ with
IAS.  In most cases IAS- and GNU as-assembled boot components were
identical, and CLANG_NO_IAS was already removed from other components.

Clang IAS produces different output for some components, including
pxeldr, so CLANG_NO_IAS was not previously removed for those.

In the case of pxeldr the difference is that IAS adds a size override
prefix (67h) to three instructions to specify a 32-bit address, even
though the two high bytes are zero and the address fits in 16 bits.
this wastes three bytes per instruction and causes some additional nop
npadding to be required elsewhere in the object, but pxeboot is not
size-constrained so it doesn't matter.

Sponsored by:	The FreeBSD Foundation
2019-08-15 17:59:22 +00:00
Ed Maste
1596e9a43c stand: remove CLANG_NO_IAS from cdboot
Many components under stand/ had CLANG_NO_IAS added when Clang's
Integrated Assembler (IAS) did not handle .codeNN directives.  Clang
gained support quite some time ago, and we can now build stand/ with
IAS.  In most cases IAS- and GNU as-assembled boot components were
identical, and CLANG_NO_IAS was already removed from other components.

Clang IAS produces different output for some components, including
cdboot, so CLANG_NO_IAS was not previously removed for those.

In the case of cdboot the difference is that IAS adds a size override
prefix (67h) to many instructions to specify a 32-bit address, even
though the two high bytes are zero.  This wastes three bytes per
instance, but as cdboot is not size-constrained it doesn't matter.
Padding is also different in one case; Clang used two one-byte nops
while GNU as used a single two-byte xchg %eax, %eax.  In any case, there
is no functional change.

Sponsored by:   The FreeBSD Foundation
2019-08-15 17:32:39 +00:00
Kyle Evans
a78c1027d7 Revert r351078, r351085: stand/quad.h eviction
It did not go well; further examination is required...
2019-08-15 17:15:32 +00:00
Kyle Evans
7ddaee2559 stand: kick out quad.h
Use quad.h from libc instead for the time being. This reduces the number of
nearly-identical-quad.h we have in tree to two with only minor changes.

Prototypes for some *sh*di3 have been added to match the copy in libkern.
The differences between the two are likely few enough that they can perhaps
be merged with little additional effort to bring us down to 1.

MFC after:	3 days
2019-08-15 15:47:48 +00:00
Ed Maste
3a6c85abcb stand: remove CLANG_NO_IAS from boot2
Many components under stand/ had CLANG_NO_IAS added when Clang's
Integrated Assembler (IAS) did not handle .codeNN directives.  Clang
gained support quite some time ago, and we can now build stand/ with
IAS.  In most cases IAS- and GNU as-assembled boot components were
identical, and CLANG_NO_IAS was already removed from other components.

Clang IAS produces different output for some components, including
boot2, so CLANG_NO_IAS was not previously removed for those.

In the case of boot2 the difference is that IAS produces a larger
encoding for one instruction (the testb at the beginning of read).

GNU as produces:

    2e    f6 06 b0 08       80

while IAS includes an address size override prefix (67) and produces:

    2e 67 f6 05 b3 08 00 00 80

This results in three fewer NOPs elsewhere in boot2 but no functional
change, so switch to IAS for boot2.

(We can separately pursue improved 16-bit IAS support with the LLVM
developers.)

Sponsored by:	The FreeBSD Foundation
2019-08-15 14:54:18 +00:00
Toomas Soome
6e5555adb1 loader: add error check for vdev_indirect calls
We can error out due to memory allocation errors, we can not recover from
those and need to get out.
2019-08-09 19:09:05 +00:00
Toomas Soome
b1b9326846 loader: support com.delphix:removing
We should support removing vdev from boot pool. Update loader zfs reader
to support com.delphix:removing.

Reviewed by:	allanjude
MFC after:	2 weeks
Differential Revision:	https://reviews.freebsd.org/D18901
2019-08-08 18:08:13 +00:00
Marcin Wojtas
435672e322 Verify files loaded in chain command.
The chain command can be used to chain load another binary.
If veriexec is enabled we should verify it first.
Note that on EFI systems the verification was already done
through firmware, assuming that Secure Boot was enabled there.

Submitted by: Kornel Duleba <mindal@semihalf.com>
Reviewed by: sjg
MFC after: 1 week
Obtained from: Semihalf
Differential Revision: https://reviews.freebsd.org/D20952
2019-08-08 17:03:30 +00:00
Toomas Soome
9cb069b552 loader.efi: efilib.h and libefi.c were left out of 350654
And this is second followup for 350654
2019-08-06 20:13:28 +00:00
Toomas Soome
cefffc0b45 loader.efi: boot1.c was left out of 350654
Followup for 350654.
2019-08-06 20:08:01 +00:00
Toomas Soome
110d56cbf4 loader.efi: replace HandleProtocol() with OpenProtocol()
The HandleProtocol() is deprecated interface and we should use OpenProtocol()
instead. Moreover, in some firmware implementation(s), the HandleProtocol()
does return device path using static storage, so we can not keep the value
returned there. With same firmware, the OpenProtocol() does return data we
do not need to clone.

Reviewed by:	imp
Differential Revision:	https://reviews.freebsd.org/D21162
2019-08-06 19:27:27 +00:00
Leandro Lupori
83e356c3c9 [PPC64] Implement CAS
Guest PPC OSs running under a hypervisor may communicate the features they
support, in order for the hypervisor to expose a virtualized machine in the way
the client (guest OS) expects (see LoPAPR 1.1 - B.6.2.3).

This is done by calling the "/ibm,client-architecture-support" (CAS) method,
informing supported features in option vectors.  Until now, FreeBSD wasn't
using CAS, but instead relied on hypervisor/QEMU's defaults.

The problem is that, without CAS, it is very inconvenient to run POWER9 VMs on
a POWER9 host running with radix enabled.  This happens because, in this case,
the QEMU default is to present the guest OS a dual MMU (HPT/RPT), instead of
presenting a regular HPT MMU, as FreeBSD expects, resulting in an early panic.
The known workarounds required either changing the host to disable radix or
passing a flag to QEMU to run in a POWER8 compatible mode.

With CAS, FreeBSD is now able to communicate that it wants an HPT MMU,
independent of the host setup, which now makes FreeBSD work on POWER9/pseries,
with KVM enabled and without hugepages (support added in a previous commit).

As CAS is invoked through OpenFirmware's call-method interface, it needs to be
performed early, when OpenFirmware is still operational. Besides, now that FDT
is the default way to inspect the device tree on PPC, OFW call-method feature
will be unavailable by default, when control is passed to the kernel. Because
of this, the call to CAS is being performed at the loader, instead of at the
kernel.

To avoid regressions with old platforms, this change uses CAS only on
POWER8/POWER9.

Reviewed by:	jhibbits
Differential Revision:	https://reviews.freebsd.org/D20827
2019-07-31 20:31:36 +00:00
Ruslan Bukin
be4890f75e Fix EFI loader build when LOADER_NET_SUPPORT=no.
Sponsored by:	DARPA, AFRL
2019-07-30 14:54:18 +00:00
Simon J. Gerraty
bbac74ca3c loader: ignore some variable settings if input unverified
libsecureboot can tell us if the most recent file opened was
verfied or not.
If it's state is VE_UNVERIFIED_OK, skip if variable
matches one of the restricted prefixes.

Reviewed by:	stevek
MFC after:	1 week
Sponsored by:	Juniper Networks
Differential Revision:	https://reviews.freebsd.org//D20909
2019-07-17 23:33:14 +00:00
Justin Hibbits
1ea717577f Allow efi loader to get network params from uboot
Summary:
efi loader does not work with static network parameters. It always uses
BOOTP/DHCP and also uses RARP as a fallback.  Problems with DHCP servers can
cause the loader to fail to populate network parameters.

Submitted by:	Siddharth Tuli <siddharthtuli_gmail.com>
Reviewed by:	imp
Sponsored by:	Juniper Networks, Inc.
Differential Revision:	https://reviews.freebsd.org/D20811
2019-07-12 00:54:20 +00:00
Toomas Soome
21b2840e7d efihttp: mac and err can be used uninitialized
While there, also check if mac != NULL, and use pointer compare for ipv4
and dns.
2019-07-02 20:32:35 +00:00
Toomas Soome
8136db2894 efihttp: mark unused arguments with __unused
we do have __unused, lets use it.
2019-06-30 20:21:27 +00:00
Toomas Soome
f7ae06cb5e efihttp: comparison of integers of different signs
message.HeaderCount is UINTN (unsigned int), so should be i.
2019-06-30 20:18:31 +00:00
Toomas Soome
5bca29b882 Clean efihttp pointer-sign warnings
The Http protocol structure is using unsigned char strings, Use type casts
where needed.
2019-06-30 20:15:02 +00:00
Rebecca Cran
94e8f7c65f Increase EFI_STAGING_SIZE to 100MB on x64
To avoid failures when the large 18MB nvidia.ko module is being loaded,
increase EFI_STAGING_SIZE from 64MB to 100MB on x64 systems.
Leave the other platforms at 64MB.
2019-06-27 22:06:41 +00:00
Rebecca Cran
1ee03da25e Re-enable loader efi http boot and fix dv_open bug if dv_init failed
The code in efihttp.c was assuming that dv_open wouldn't be called if
dv_init failed. But the dv_init return value is currently ignored.

Add a new variable, `efihttp_init_done` and only proceed in dv_open if
it's true. This fixes the loader on systems without efi http support.
2019-06-26 03:06:57 +00:00
Rebecca Cran
519346ce8c Disconnect EFI HTTP support
The EFI HTTP code has been causing boot failures for people, so disable it
while a fix is being worked on.
2019-06-25 21:26:57 +00:00
Warner Losh
f5a95d9a07 Remove NAND and NANDFS support
NANDFS has been broken for years. Remove it. The NAND drivers that
remain are for ancient parts that are no longer relevant. They are
polled, have terrible performance and just for ancient arm
hardware. NAND parts have evolved significantly from this early work
and little to none of it would be relevant should someone need to
update to support raw nand. This code has been off by default for
years and has violated the vnode protocol leading to panics since it
was committed.

Numerous posts to arch@ and other locations have found no actual users
for this software.

Relnotes:	Yes
No Objection From: arch@
Differential Revision: https://reviews.freebsd.org/D20745
2019-06-25 04:50:09 +00:00
Rebecca Cran
da4961c796 loader: add HTTP support using UEFI
Add support for an HTTP "network filesystem" using the UEFI's HTTP
stack.

This also supports HTTPS, but TianoCore EDK2 implementations currently
crash while fetching loader files.
Only IPv4 is supported at the moment. IPv6 support is planned for a
follow-up changeset.

Note that we include some headers from the TianoCore EDK II project in
stand/efi/include/Protocol verbatim, including links to the license instead
of including the full text because that's their preferred way of
communicating it, despite not being normal FreeBSD project practice.

Submitted by:	scottph
Reviewed by:	imp, bcran
Differential Revision:	https://reviews.freebsd.org/D20643
2019-06-24 23:18:42 +00:00
Warner Losh
ec9abc1843 Move to using a common kernel path between the boot / laoder bits and
the kernel.
2019-06-24 20:34:53 +00:00
Matt Macy
6459a61ea7 Tell loader to ignore newer features enabled on the root pool.
There are many new features in ZoF. Most, if not all, do not effect read only usage.
Encryption in particular is enabled at the pool level but used at the dataset level.
The loader obviously will not be able to boot if the boot dataset is encrypted, but
should not care if some other dataset in the root pool is encrypted.

Reviewed by:	allanjude
MFC after:	1 week
2019-06-19 21:10:13 +00:00
Rebecca Cran
3109cebc22 efinet: Defer exclusively opening the network handles
Don't commit to exclusive access to the network device handle by
efinet until the loader has decided to load something through the
network. This allows for the possibility of other users of the
network device.

Submitted by:	scottph
Reviewed by:	tsoome, emaste
Tested by: 	tsoome, bcran
Differential Revision:	https://reviews.freebsd.org/D20642
2019-06-19 18:47:44 +00:00
Leandro Lupori
68ed5ad2d5 [PPC] Fix loader input with newer QEMU versions
At least since version 4.0.0, QEMU became bug-compatible with PowerVM's
vty, by inserting a \0 after every \r. As this confuses loader's
interpreter and as a \0 coming from the console doesn't seem reasonable,
it's now being filtered at OFW console input.

Reviewed by:	jhibbits
MFC after:	2 weeks
Differential Revision:	https://reviews.freebsd.org/D20676
2019-06-19 11:37:43 +00:00
Xin LI
f89d207279 Separate kernel crc32() implementation to its own header (gsb_crc32.h) and
rename the source to gsb_crc32.c.

This is a prerequisite of unifying kernel zlib instances.

PR:		229763
Submitted by:	Yoshihiro Ota <ota at j.email.ne.jp>
Differential Revision:	https://reviews.freebsd.org/D20193
2019-06-17 19:49:08 +00:00