Add the now coreboot standard MMIO read/write accessors that were
already defined for other architectures but not x86.
This leaves the old read/write{b,w,l} variants in place as was done
on the other architectures, presumably to support old payloads that
have not been updated.
BUG=chrome-os-partner:43072
BRANCH=none
TEST=emerge-glados libpayload
CQ-DEPEND=CL:294711
Change-Id: I5ae3d755adcef0f6ff27aaa7c35a5b12ddc32e22
Signed-off-by: Patrick Georgi <patrick@georgi-clan.de>
Original-Commit-Id: c09dd557050e3002fa5b8504980d72d4cb79a56c
Original-Change-Id: I58d928338335d3fe4bb7fe2bdc9c2967d8689118
Original-Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/294565
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/11405
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Alexander Couzens <lynxis@fe80.eu>
libpayload on x86 defines u32 and uint32_t as typedefs of
unsigned int. However, the readl/writel routines use long.
With alias checking this throws type punning errors. Align
the readl/writel/inl/outl types with the 32-bit fixed width
ones that are exposed.
Change-Id: Ie51cff8af4596948f6132e3cb743f1bc4ea8f204
Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-on: http://review.coreboot.org/10186
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Add support for memory barriers in arch {arm,arm64,x86}. This is required to
force strict CPU ordering. Definitions are based on FREEBSD atomic.h
definitions.
BUG=chrome-os-partner:31533
BRANCH=None
TEST=Memory barriers tested with ehci driver on arm64
Change-Id: I50060b0f33a6bd6cb95e829df079df379b2ff2a5
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Original-Commit-Id: 937d66cdab92a8521ede8307f5af8f5c20d3e552
Original-Change-Id: Ie51e3452f7a254b24111000da5dbe8714ac22223
Original-Signed-off-by: Furquan Shaikh <furquan@google.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/213916
Original-Tested-by: Furquan Shaikh <furquan@chromium.org>
Original-Reviewed-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Original-Commit-Queue: Furquan Shaikh <furquan@chromium.org>
Reviewed-on: http://review.coreboot.org/8731
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
setbits_le32() is not really arch-specific... the arch-specific part of
accessing memory is wrapped by readl() and writel(), and the endianness
can be accounted for with the right macros. Generalize the definitions,
add a be32 version and move them to endian.h so that all platforms can
use them. Also include endian.h from libpayload.h so we won't update any
payload's old use of the macros (endianness is something useful enough
to always have avalable anyway, and shouldn't clash with other things).
This also fixes a bug where these macros would only be available if
libpayload-config.h had been independently included before.
Also fix a bug with readl() macros on all archs where they refused to
work on const pointers (which they should).
CQ-DEPEND=CL:208712
BUG=None
TEST=Stuff still compiles. Built and booted on Storm.
Original-Change-Id: I01a7fbadbb5d740675657d95c1e969027562ba8c
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/208713
Original-Reviewed-by: Vadim Bendebury <vbendeb@chromium.org>
Original-Reviewed-by: David Hendricks <dhendrix@chromium.org>
(cherry picked from commit 951f8a6d77bc21bd793bf4f228a0965ade586f00)
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
Change-Id: I51c25f01b200b91abbe32c879905349bb05dc9c8
Reviewed-on: http://review.coreboot.org/8129
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Tested-by: build bot (Jenkins)
This patch makes some slight changes to the exception hook interface.
The old code provides a different handler hook for every exception
type... however, in practice all those hook functions often need to look
very similar, so this creates more boilerplate than it removes. The new
interface just allows for a single hook with the exception type passed
as an argument, and the consumer can signal whether the exception was
handled through the return value. (Right now this still only supports
one consumer, but it could easily be extended to walk through a list of
hooks if the need arises.)
Also move the excepton state from an argument to a global. This avoids a
lot of boilerplate since some consumers need to change the state from
many places, so they would have to pass the same pointer around many
times. It also removes the false suggestion that the exception state was
not global and you could have multiple copies of it (which the exception
core doesn't support for any architecture).
On the ARM side, the exception state is separated from the exception
stack for easier access. (This requires some assembly changes, and I
threw in a few comments and corrected the immediate sigils from '$' to
the official '#' while I'm there.) Since the exception state is now both
stored and loaded through an indirection pointer, this allows for some
very limited reentrance (you could point it to a different struct while
handling an exception, and while you still won't be able to return to
the outer-level exception from there, you could at least swap out the
pointer and return back to System Mode in one go).
BUG=chrome-os-partner:18390
TEST=Made sure normal exceptions still get dumped correctly on both
archs.
Original-Change-Id: I5d9a934fab7c14ccb2c9d7ee4b3465c825521fa2
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/202562
Original-Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
(cherry picked from commit 97542110f0b385b9b8d89675866e65db8ca32aeb)
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
*** Squashed to prevent build failures. ***
libpayload: align arm64 with new exception handling model
The exception handling was previously updated, however the
arm64 changes raced with hat one. Make the arm64 align with
the new model. Without these changes compilation will fail.
BUG=None
BRANCH=None
TEST=Can build libpayload for rush.
Original-Change-Id: I320b39a57b985d1f87446ea7757955664f8dba8f
Original-Signed-off-by: Aaron Durbin <adurbin@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/204402
Original-Reviewed-by: Furquan Shaikh <furquan@chromium.org>
Original-Commit-Queue: Furquan Shaikh <furquan@chromium.org>
(cherry picked from commit 0080df41b311ef20f9214b386fa4e38ee54aa1a1)
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
Change-Id: I9a0bb3848cf5286f9f4bb08172a9f4a15278348e
Reviewed-on: http://review.coreboot.org/8117
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
To support a GDB stub, it will be necessary to trap various exceptions which
will be used to implement breakpoints, single stepping, etc.
BUG=None
TEST=Built and booted on Link with hooks installed and saw that they
triggered when exceptions occurred. Built and booted on nyan.
BRANCH=None
Original-Change-Id: Iab659365864a3055159a50b8f6e5c44290d3ba2b
Original-Signed-off-by: Gabe Black <gabeblack@google.com>
Original-Reviewed-on: https://chromium-review.googlesource.com/179602
Original-Reviewed-by: Gabe Black <gabeblack@chromium.org>
Original-Tested-by: Gabe Black <gabeblack@chromium.org>
Original-Commit-Queue: Gabe Black <gabeblack@chromium.org>
(cherry picked from commit 8db0897b1ddad600e247cb4df147c757a8187626)
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
Change-Id: I5e7f724b99988cd259909dd3bd01166fa52317ec
Reviewed-on: http://review.coreboot.org/7656
Tested-by: build bot (Jenkins)
Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
This patch changes several cache-related pieces to be cleaner, faster or
more correct. The largest point is removing the old
arm_invalidate_caches() function and surrounding bootblock code to
initialize SCTLR and replace it with an all-assembly function that takes
care of cache and SCTLR initialization to bring the system to a known
state. It runs without stack and before coreboot makes any write
accesses to be as compatible as possible with whatever state the system
was left in by preceeding code. This also finally fixes the dreaded
icache bug that wasted hundreds of milliseconds during boot.
Old-Change-Id: I7bb4995af8184f6383f8e3b1b870b0662bde8bd4
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/183890
(cherry picked from commit 07a35925dc957919bf88dfc90515971a36e81b97)
nyan_big: apply cache-related changes from nyan
This applies the same changes from 07a3592 that were applied to nyan.
Old-Change-Id: Idcbe85436d7a2f65fcd751954012eb5f4bec0b6c
Reviewed-on: https://chromium-review.googlesource.com/184551
Commit-Queue: David Hendricks <dhendrix@chromium.org>
Tested-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
(cherry picked from commit 4af27f02614da41c611aee2c6d175b1b948428ea)
Squashed the followup patch for nyan_big into the original patch.
Change-Id: Id14aef7846355ea2da496e55da227b635aca409e
Signed-off-by: Isaac Christensen <isaac.christensen@se-eng.com>
(cherry picked from commit 4cbf25f8eca3a12bbfec5b015953c0fc2b69c877)
Signed-off-by: Marc Jones <marc.jones@se-eng.com>
Reviewed-on: http://review.coreboot.org/6993
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
This patch adds another cache invalidation stub to the x86 arch to
make it usable in common code. This whole stuff should probably be
redesigned anyway but I just want to get it working and unblock my CL
for now... more cleanups coming later.
Change-Id: I2e8bdd8aa0e6723209384c24042f053f2e993fe6
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/182534
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
(cherry picked from commit cafce5182a7a2a9ce17ad40d9d893a40ebd5aafd)
Signed-off-by: Isaac Christensen <isaac.christensen@se-eng.com>
Reviewed-on: http://review.coreboot.org/6919
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
This patch makes the EHCI driver work on ARM platforms which usually do
not support automatic cache snooping. It uses the new DMA memory
mechanism (which needs to be correctly set up in the Coreboot mainboard
code) to allocate all EHCI-internal communication structures in
cache-coherent memory, and cleans/invalidates the externally supplied
transfer buffers in Bulk and Control functions with explicit calls as
necessary.
Old-Change-Id: Ie8a62545d905b7a4fdd2a56b9405774be69779e5
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/167339
(cherry picked from commit 322338934add36a5372ffe7d2a45e61a4fdd4a54)
libpayload: ehci: Cache management is hard, let's go copying...
It turns out that my previous commit to make the EHCI stack cache aware
on ARM devices wasn't quite correct, and the problem is actually much
trickier than I thought. After having some fun with more weird transfer
problems that appear/disappear based on stack alignment, this is my
current worst-case threat model that any cache managing implementation
would need to handle correctly:
Some upper layer calls ehci_bulk() with a transfer buffer on its stack.
Due to stack alignment, it happens to start just at the top of a cache
line, so up to 64 - 4 bytes of ehci_bulk's stack will share that line.
ehci_bulk() calls dcache_clean() and initializes the USB transfer.
Between that point and the call to dcache_invalidate() at the end of
ehci_bulk(), any access to the stack variables in that cache line (even
a speculative prefetch) will refetch the line into the cache. Afterwards
any other access to a random memory location that just happens to get
aliased to the same cache line may evict it again, causing the processor
to write out stale data to the transfer buffer and possibly overwrite
data that has already been received over USB.
In short, any dcache_clean/dcache_invalidate-based implementation that
preserves correctness while allowing any arbitrary (non cache-aligned)
memory location as a transfer buffer is presumed to be impossible.
Instead, this patch causes all transfer data to be copied to/from a
cache-coherent bounce buffer. It will still transfer directly if the
supplied buffer is already cache-coherent, which can be used by callers
to optimize their transfers (and is true by default on x86).
Old-Change-Id: I112908410bdbc8ca028d44f2f5d388c529f8057f
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/169231
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
(cherry picked from commit 702dc50f1d56fe206442079fa443437f4336daed)
Squashed the initial commit and a follow up fix.
Change-Id: Idf7e5aa855b4f0221f82fa380a76049f273e4c88
Signed-off-by: Isaac Christensen <isaac.christensen@se-eng.com>
Reviewed-on: http://review.coreboot.org/6633
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
So far this is used by the USB driver, and instead of
having ifdefs all throughout that code, implement the same
API on x86 and ARM.
Change-Id: I8093ad818ad2e38a0901787aa8674faf591d580c
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: https://gerrit.chromium.org/gerrit/56105
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Commit-Queue: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Stefan Reinauer <reinauer@chromium.org>
Reviewed-on: http://review.coreboot.org/4320
Tested-by: build bot (Jenkins)
The way we got to include the compiler includes was kind of whacky.
Instead of mixing in potentially problematic headers, make libpayload
self-contained by adding some missing header files. Also clean up
conflicting definitions of size_t throughout the tree.
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Change-Id: I0ad1194de1a00b7133c5477c00eb167d63a2ee85
Reviewed-on: https://gerrit.chromium.org/gerrit/47608
Reviewed-by: Ronald G. Minnich <rminnich@chromium.org>
Commit-Queue: Stefan Reinauer <reinauer@google.com>
Tested-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/3058
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
There were a number of type issues in libpayload that sneaked in
with 903f8e0.
- size_t and ssize_t were conflicting with gcc builtins
- some stdint types were used in libpayload but not defined
in our stdint.h
With this patch it's possible to compile libpayload with the
reference toolchain again.
Change-Id: Idd5ccfdd9f3536b36bceca2d101e7405883b10bc
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/2903
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Some new TPM drivers in depthcharge require that type. I added it to
arch/types.h which seemed appropriate, but I'm not sure that's exactly the
right header to use, or in other words if you'd get that type from libpayload
the same way you'd get it if you were building a standard Linux program.
Also, I attempted to determine what underlying types gcc would use, and while
I think I picked the right ones I'm not 100% certain of that either.
Change-Id: Ic5c0b4173c8565ede3bfce8870976d596d69e51d
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: http://review.coreboot.org/2669
Tested-by: build bot (Jenkins)
Reviewed-by: Marc Jones <marc.jones@se-eng.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Duncan Laurie