Commit Graph

686 Commits

Author SHA1 Message Date
Andrew Turner
619e50a657 Remove the psci option from arm64. It is now a standard option as it is
required to boot correctly.

Sponsored by:	DARPA, AFRL
2018-06-10 19:42:44 +00:00
Andrew Turner
dc9b99a884 Clean up handling of unexpected exceptions. Previously we would issue a
breakpoint instruction, however this would lose information that may be
useful for debugging.

These are now handled in a similar way to other exceptions, however it
won't exit out of the exception handler until it is known if we can
handle these exceptions in a useful way.

Sponsored by:	DARPA, AFRL
2018-06-10 16:21:21 +00:00
Konstantin Belousov
8d59ab652b Restore release semantic for the old thread unlock on arm64.
With the introduction of pmap_switch(), the DSB instruction on the
address map switch is not necessary executed, which is fixed by
changing the unlock store to release.  Also remove comment which
documented pre-pmap_switch() code.

Reviewed by:	andrew
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2018-06-08 18:32:26 +00:00
Kevin Lo
3fff2af912 Since we don't enable BUF_TRACKING and FULL_BUF_TRACKING buffer debugging
options in GENERIC kernels on arm and arm64, there's no need to disable
them.

Sponsored by:	MSI/FUNTORO
2018-06-05 05:24:42 +00:00
Andrew Turner
c891735cf7 Move the code to print the EFI memory table to a new function and call it
in teh bootverbose path after cninit().

This allows users to see these tables when booting with boot -v.

Sponsored by:	DARPA, AFRL
2018-05-31 12:37:34 +00:00
Andrew Turner
ede605609a Fix the early spelling of bootverbose.
Sponsored by:	DARPA, AFRL
2018-05-31 11:53:46 +00:00
Andrew Turner
5428bb231f Fix the memory attribute for EFI_MD_ATTR_UC on arm64.
This is defined as Device-nGnRnE in the UEFI spec (UEFI 2.4, section
2.3.6.1). This is the VM_MEMATTR_DEVICE type in FreeBSD.

Reported by:	tuexen
Sponsored by:	DARPA, AFRL
2018-05-31 11:11:23 +00:00
Andrew Turner
9b56058160 Remove max_pa, it's unused.
Sponsored by:	DARPA, AFRL
2018-05-30 15:37:09 +00:00
Andrew Turner
e2b8bf0a18 Further limit when we call pmap_fault.
We should only call pmap_fault in the kernel when accessing a userspace
address. As this should always happen through specific functions that set
a fault handler we can use this to limit calls to pmap_fault to when this
is set.

This should help with NULL pointer dereferences when we are unable to sleep
so we fall into the correct case.

Sponsored by:	DARPA, AFRL
2018-05-30 15:25:48 +00:00
Andrew Turner
abf7742a1c Push down the locking in pmap_fault to just be around the calls to
arm64_address_translate_*. There is no need to lock around the switch
statement as we only care about a few cases.

Sponsored by:	DARPA, AFRL
2018-05-30 14:18:19 +00:00
Andrew Turner
463ac3dda0 On ThunderX2 we need to be careful to only map the memory the firmware
lists in the EFI memory map. As such we need to reduce the mappings to
restrict them to not be the full 1G block. For now reduce this to a 2M
block, however this may be further restricted to be 4k page aligned as
other SoCs may require.

This allows ThunderX2 to boot reliably to userspace without performing
any speculative memory accesses to invalid physical memory.

This is a recommit of r334035 now that we can access the EFI Runtime data
through the DMAP region.

Tested by:	tuexen
Sponsored by:	DARPA, AFRL
2018-05-29 13:52:25 +00:00
Emmanuel Vadot
44654b755d arm64: fix atomic_fcmpset_16
newval needs to be uint16_t

Reported by:	andrew
2018-05-28 21:05:00 +00:00
Emmanuel Vadot
e39ce4cafb arm64: Add atomic_fcmpset_8 and atomic_fcmpset_16
Reviewed by:	cognet
2018-05-28 20:29:03 +00:00
Andrew Turner
2dd5ae7a90 Create a new function to walk the EFI memory table & run a callback for
each entry. We can then use this to ensure the RunTime data is mapped in
the DMAP, but not in phys_avail.

Sponsored by:	DARPA, AFRL
2018-05-28 17:09:29 +00:00
Andrew Turner
b50b5555ef Remove physmap from the arm64 machdep.h. This was missed in r334162. 2018-05-24 16:07:47 +00:00
Andrew Turner
9f1a80706c Allow us to read the physmap data into our own array and use this to build
the DMAP region on arm64.

We already have the needed information to build these tables, we just need
to extract it. This significantly simplifies the code.

Obtained from:	ABT Systems Ltd
Sponsored by:	Turing Robotic Industries
2018-05-24 15:32:49 +00:00
Andrew Turner
d5591551a8 Print the physmem tables under a verbose boot.
Obtained from:	ABT Systems Ltd
Sponsored by:	Turing Robotic Industries
2018-05-24 15:07:53 +00:00
Andrew Turner
c602678b57 Exclude memory from the /reserved-memory mappings with the no-map property
set. This memory must not be mapped by the operating system other than
under control of the device driver.

Obtained from:	ABT Systems Ltd
Sponsored by:	Turing Robotic Industries
2018-05-24 14:55:50 +00:00
Emmanuel Vadot
3df266dff9 arm64: rockchip: Add proper armclock support
The core clock (armclk) on RockChip SoC is special.
It can derive it's clock from many PLLs but RockChip recommand to do it
from "apll" on old SoC and "npll" on new SoC. The reason for choosing npll
is that it's have less jitter and is more close to the arm core on the SoC.
r333314 added the core clock as a composite clock but due to it's specials
property we need to deal with it differently.
A new rk_clk_armclk type is added for this and it supports only the "npll"
as we don't run on old RockChip SoC that only have the "apll".
It will always reparent to "npll" and set the frequency according to a rate
table that is known to be good.
For now we set the "npll" to the desired frequency and just set the core clk
divider to 1 as its parent it just used for the core clk.
2018-05-23 19:07:03 +00:00
Mark Johnston
6514b4f061 Add GET_STACK_USAGE() for arm64.
Its absence meant that GEOM direct dispatch was disabled (the service
routines check the current thread's stack usage to determine whether
to hand off the request to a dedicated thread), and this change is
sufficient to enable direct dispatch by default.

Reviewed by:	allanjude
MFC after:	2 weeks
Differential Revision:	https://reviews.freebsd.org/D15527
2018-05-23 15:43:35 +00:00
Andrew Turner
84cac654ae Revert r334035 for now. It breaks the boot on some boards as er expect to
be able to read UEFI RuntimeData memory via the DMAP region.
2018-05-22 15:52:11 +00:00
Andrew Turner
89b5faf887 On ThunderX2 we need to be careful to only map the memory the firmware
lists in the EFI memory map. As such we need to reduce the mappings to
restrict them to not be the full 1G block. For now reduce this to a 2M
block, however this may be further restricted to be 4k page aligned as
other SoCs may require.

This allows ThunderX2 to boot reliably to userspace without performing
any speculative memory accesses to invalid physical memory.

Sponsored by:	DARPA, AFRL
2018-05-22 11:26:41 +00:00
Andrew Turner
9d0728e04e Stop using the DMAP region to map ACPI memory.
On some arm64 boards we need to access memory in ACPI tables that is not
mapped in the DMAP region. To handle this create the needed mappings in
pmap_mapbios in the KVA space.

Submitted by:	Michal Stanek (mst@semihalf.com)
Sponsored by:	Cavium
Differential Revision:	https://reviews.freebsd.org/D15059
2018-05-22 11:16:45 +00:00
Andrew Turner
79402150c1 Switch arm64 to use the same physmem code as 32-bit arm.
The main advantage of this is to allow us to exclude memory from being
used by the kernel. This may be from the memreserve property, or ranges
marked as no-map under the reserved-memory node.

More work is still needed to remove the physmap array. This is still used
for creating the DMAP region, however other patches need to be committed
before we can remove this.

Obtained from:	ABT Systems Ltd
Sponsored by:	Turing Robotic Industries
2018-05-22 11:07:04 +00:00
Andrew Turner
78921ae879 Restrict the faulting addresses we call pmap_fault from to just those that
may fault due to superpage mappings being changed.

Sponsored by:	DARPA, AFRL
2018-05-21 16:14:53 +00:00
Mark Johnston
892bdccca0 Enable kernel dump features in GENERIC for most platforms.
This turns on support for kernel dump encryption and compression, and
netdump. arm and mips platforms are omitted for now, since they are more
constrained and don't benefit as much from these features.

Reviewed by:	cem, manu, rgrimes
Tested by:	manu (arm64)
Relnotes:	yes
Differential Revision:	https://reviews.freebsd.org/D15465
2018-05-19 19:53:23 +00:00
Olivier Houchard
a9772d1a0e Instead of ignoring the VFP registers, set the dumppcb's pcb_fpusaved
field, so that they are saved, as they may be used in the kernel, in the
EFI and the crypto code.

Reviewed by:	andrew
2018-05-18 13:28:02 +00:00
Andrew Turner
00454c400e Enable the Qualcomm MSM UART driver. This is needed for some Qualcomm
Snapdragon SoCs.

Obtained from:	ABT Systems Ltd
Sponsored by:	Turing Robotic Industries
2018-05-18 11:32:48 +00:00
Olivier Houchard
8b2ec5a3a1 In pmap_get_tables(), check that the L2 is indeed a table before attempting
to get the l3.
2018-05-17 22:40:22 +00:00
Olivier Houchard
654a792922 In vfp_save_state(), don't bother trying to save the VFP registers if the
provided PCB doesn't have a pcb_fpusaved. All PCBs associated to a thread
should have one, but the dumppcb used when panic'ing doesn't.
2018-05-17 22:38:16 +00:00
Andriy Gapon
c9c4d38aa8 followup to r332730/r332752: set kdb_why to "trap" for fatal traps
This change updates arm, arm64 and mips achitectures.  Additionally, it
removes redundant checks for kdb_active where it already results in
kdb_reenter() and adds kdb_reenter() calls where they were missing.

Some architectures check the return value of kdb_trap(), but some don't.
I haven't changed any of that.

Some trap handling routines have a return code.  I am not sure if I
provided correct ones for returns after kdb_reenter().  kdb_reenter
should never return unless kdb_jmpbufp is NULL for some reason.

Only compile tested for all affected architectures.  There can be bugs
resulting from my poor understanding of architecture specific details.

Reported by:	jhb
Reviewed by:	jhb, eadler
MFC after:	4 weeks
Differential Revision: https://reviews.freebsd.org/D15431
2018-05-16 06:52:08 +00:00
Andrew Turner
25964cd229 Increase the number of pages we allocate in the arm64 early boot. We are
already close to the limit so increasing the kernel size may cause it to
fail to boot when it runs past the end of allocated memory.

Reported by:	manu
Sponsored by:	DARPA, AFRL
2018-05-15 16:44:35 +00:00
Emmanuel Vadot
43fd679efb arm64: Add ALT_BREAK_TO_DEBUGGER to GENERIC
It is useful to enter kdb with an escape sequence.
While here move the USB_DEBUG with the others debug options and define
nooptions USB_DEBUG for GENERIC-NODEBUG
2018-05-10 09:37:50 +00:00
Emmanuel Vadot
08f3f0f953 arm64: rockchip: cru: Call clk_set_assigned
We need to call clk_set_assigned after all the clock have been registered
to set the parents/rates described in the dtb.
2018-05-07 07:31:25 +00:00
Emmanuel Vadot
dff9720331 arm64: rockchip: clk: Add support to reparent to clk_composite
All clk_composite type have the possibility to reparent (choosing another
parent to find a better frequency), add the support for that.
2018-05-07 07:29:48 +00:00
Emmanuel Vadot
66a4c42756 arm64: rk3328: Add pll rates tables
Add the known value to be safe for the rk3328 PLLs
2018-05-07 07:28:47 +00:00
Emmanuel Vadot
78d07c93a7 arm64: rk: Add support for setting pll rate
Add support for setting pll rate. On RockChip SoC two kind of plls are
supported, integer mode and fractional mode.
The two modes are intended to support more frequencies for the core plls.
While here change the recalc method as it appears that the datasheet is
wrong on the calculation method.
2018-05-07 07:28:10 +00:00
Emmanuel Vadot
178f57b143 arm64: rockchip: rk3328: Add armclk clock
Add the clock definition for the arm clock.
While here remove the indexes in the clock table as we will need clock
with a 0 index (non-exported clocks).
2018-05-07 07:26:48 +00:00
Mark Johnston
20f85b1ddd Print the dump progress indicator after calling dump_start().
Dumpers may wish to print messages from an initialization hook; this
change ensures that such messages aren't mixed with output from the
generic dump code.

MFC after:	1 week
2018-05-01 17:32:43 +00:00
Emmanuel Vadot
590a3c9619 arm64: rockchip: rk_gpio fix rk_gpio_pin_config32
Pointy Hat to:	 me
2018-04-26 22:15:09 +00:00
Emmanuel Vadot
cbf65cf228 arm64: rockchip: Add gpio controller driver
Add a driver that match on 'rockchip,gpio-bank', this compatible
string is found on almost all RockChip SoC so this driver is compatible
with almost all of the RockChip SoCs.

The only features missing for this driver are :
- Interrupts support
- Debouncing
2018-04-26 21:44:00 +00:00
Emmanuel Vadot
d40a1b91c0 arm64: rockchip: RK3328 CRU Add gpio gates
Add the gates for the gpio controller in order to properly support them.
2018-04-26 21:40:05 +00:00
Emmanuel Vadot
a59a485e79 arm64: rockchip: Rk3328 CRU Fix some offset for gates
Some offset of some clock gates where wrong, correct them so we can
use thoses clocks.

Pointy Hat to:	me
2018-04-26 21:38:59 +00:00
Emmanuel Vadot
be2e72c286 arm64: rockchip: Add pinctrl driver
Add pinctrl driver for RockChip SoCs. This device manage which function
to set on which pin and some other properties like pull up/down, drive
strength etc ...
For now the driver only support RK3328 but it is versatile enough to
add support for other RockChip SoC in the future.
2018-04-26 21:37:38 +00:00
Emmanuel Vadot
49755fc6d2 arm64: rockchip: Add GRF driver
RockChip GRF (General Register Files) is present on almost all RockChip
SoC and is used to control some area of the system like iomuxing, gpio
or usb phy.
We need it to be probed and attached early in the boot process so
subclass syscon_generic and set the pass to BUS_PASS_BUS + BUS_PASS_ORDER_MIDDLE.
2018-04-26 21:35:04 +00:00
Ed Maste
5f2336efb4 arm64 linuxulator: add generated sysent files
From syscalls.master in r333027

Sponsored by:	Turing Robotic Industries Inc.
2018-04-26 18:46:38 +00:00
Ed Maste
2464626e2c Add arm64 Linux syscall table
This is the first step (after the recent refactoring of some common
code) to supporting the Linuxulator on arm64.

Reviewed by:	andrew
Sponsored by:	Turing Robotic Industries Inc.
Differential Revision:	https://reviews.freebsd.org/D15187
2018-04-26 18:38:59 +00:00
Brooks Davis
9c11d8d483 Remove the unused fuwintr() and suiwintr() functions.
Half of implementations always failed (returned (-1)) and they were
previously used in only one place.

Reviewed by:	kib, andrew
Obtained from:	CheriBSD
Sponsored by:	DARPA, AFRL
Differential Revision:	https://reviews.freebsd.org/D15102
2018-04-17 18:04:28 +00:00
Ruslan Bukin
5819c3eb8a Enable Qualcomm Debug Subsystem (QDSS) block on MSM8916 SoC.
This is required for ARM Coresight operation on Dragonboard 410c.

Sponsored by:	DARPA, AFRL
Differential Revision:	https://reviews.freebsd.org/D14987
2018-04-10 12:53:48 +00:00
Oleksandr Tymoshenko
b7dc17589b Fix one more OF_getprop_alloc instance missed in r332310
X-MFC-With:	r332310
2018-04-08 23:17:51 +00:00