I always thought the support for multiple logical SCSI units in the USB
mass storage class was a dead feature. Turns out that it's actually used
by SD card readers that provide multiple slots (e.g. one regular sized
and one micro-SD). Implementing perfect support for that would require a
major redesign of the whole MSC stack, since the one device -> one disk
assumption is deeply embedded in our data structures.
Instead, this patch implements a poor man's LUN support that will just
cycle through all available LUNs (in multiple calls to usb_msc_poll())
until it finds a connected device. This should be reasonable enough to
allow these card readers to be usable while only requiring superficial
changes.
Also removes the unused 'protocol' attribute of usb_msc_inst_t.
BRANCH=rambi?,nyan
BUG=chrome-os-partner:28437
TEST=Alternatively plug an SD or micro-SD card (or both) into my card
reader, confirm that one of them is correctly detected at all times.
Change-Id: I3df4ca88afe2dcf7928b823aa2a73c2b0f599cf2
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/198101
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
So I was debugging this faulty USB SD card reader that would just fail
it's REQUEST SENSE response for some reason (sending the CSW immediately
without the data), cursing those damn device vendors for building
non-compliant crap like I always do... when I noticed that we do not
actually set the Allocation Length field in our REQUEST SENSE command
block at all! We set a length in the CBW, but the SCSI command still has
its own length field and the SCSI spec specifically says that the device
has to return the exact amount of bytes listed there (even if it's 0). I
don't know what's more suprising: that we had such a blatant bug in this
stack for so long, or that this card reader is really the first device
to actually be spec compliant in that regard.
This patch fixes the bug and changes the command block structures to be
a little easier to read (why that field was called 'lun' before is
beyond me... LUN is a transport level thing and should never appear in
the command block at all, for any command). It also fixes a memcpy() in
wrap_cbw() to avoid a read buffer overflow that might expose stack frame
data to the device.
BRANCH=rambi?,nyan
BUG=chrome-os-partner:28437
TEST=The card reader works now (for it's first LUN at least).
Change-Id: I86fdcae2ea4d2e2939e3676d31d8b6a4e797873b
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/198100
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
We recently changed the USB stack to detach devices aggressively that we
don't intend to use. This alone is not really a problem, but it
exarcerbates the fact that our device detachment itself is not very
good. We destroy any local info about the device, but we don't properly
disable the offending port. The device keeps thinking that it's active,
and if we later try to reuse that device address for another device
things become confused.
The real fix would be to properly disable all ports that we don't intend
to use. Unfortunately, this isn't really possible in our current
device/hub polymorphism structure, and I don't want to hack a new
disable_port() callback into usbdev_t that really doesn't belong there.
We will only be able to fix this cleanly after we ported all root hubs
to the generic_hub interface.
Until then, an easy workaround is to just avoid reusing addresses as
long as possible. This is firmware, so the chance that we'll ever run
through 127 devices is really small in practice. Even if we ever fix the
underlying issue, it's probably a smart precaution to keep.
BRANCH=nyan,rambi
BUG=chrome-os-partner:28328
TEST=Boot from a hub that has an "unknown" device in an earlier port
than the stick you want to boot from, make sure you can still boot.
Change-Id: I9b522dd8cbcd441e8c3b8781fcecd2effa0f23ee
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/197420
Reviewed-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
The console output driver framework in libpayload is currently built on
the putchar primitive, meaning that every driver's function gets called
one character at a time. This becomes an issue when we add drivers that
could output multiple characters at a time, but have a high constant
overhead per invocation (such as the planned GDB stub, which needs to
wrap a special frame around output strings and wait for an
acknowledgement from the server).
This patch adds a new 'write' function pointer to the
console_output_driver structure as an alternative to 'putchar'. Output
drivers need to provide at least one of the two ('write' is preferred if
available). The CBMEM console driver is ported as a proof of concept
(since it's our most performace-critical driver and should in theory
benefit the most from less function pointer invocations, although it's
probably still negligible compared to the big sprawling mess that is
printf()).
Even with this fix, the problem remains that printf() was written with
the putchar primitive in mind. Even though normal text already contains
an optimization to allow multiple characters at a time, almost all
formatting directives cause their output (including things like
padding whitespace) to be putchar()ed one character at a time.
Therefore, this patch reworks parts of the output code (especially
number printing) to all but remove that inefficiency (directives still
invoke an extra write() call, but at least not one per character). Since
I'm touching printf() core code anyway, I also tried to salvage what I
could from that weird, broken "return negative on error" code path (not
that any of our current output drivers can trigger it anyway).
A final consequence of this patch is that the responsibility to prepend
line feeds with carriage returns is moved into the output driver
implementations. Doing this only makes sense for drivers with explicit
cursor position control (i.e. serial or video), and things like the
CBMEM console that appears like a normal file to the system really have
no business containing carriage returns (we don't want people to
accidentally associate us with Windows, now, do we?).
BUG=chrome-os-partner:18390
TEST=Made sure video and CBMEM console still look good, tried printf()
with as many weird edge-case strings as I could find and compared serial
output as well as sprintf() return value.
Change-Id: Ie05ae489332a0103461620f5348774b6d4afd91a
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/196384
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
If a payload decides not to use a USB device then the device can be
detached. This prevents the device from interfering with normal
operation on some platforms. Also, it aligns the behavior of
usb_generic_init with class-specific init functions such as
usb_msc_init, which will detach unsupported devices.
BUG=None
TEST=Manual on Squawks. Test recovery boot w/ USB 2.0 media, verify
that media boots and no babble error is encountered.
BRANCH=rambi
Change-Id: I8fb30951d273e4144cda214a30a2e86df90f2c1c
Original-Change-Id: Iee522344558749603defb2966e18765aa195dae2
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/195401
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Some kconfig options have changed but the configs haven't been updated. When
make oldconfig is run from the ebuild through emake, that doesn't seem to
causea problem for whatever reason. When run manually, however, kconfig stops
to ask for each config it doesn't have a value for. This change updates the
configs to correct that issue.
BUG=None
TEST=Ran make oldconfig manually for all boards. Built and booted on big.
BRANCH=None
Change-Id: I138f8ad9239a3834d28dc492e0a90a94442b2af1
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/194908
Reviewed-by: Julius Werner <jwerner@chromium.org>
Tested-by: Julius Werner <jwerner@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
These drivers are needed right away and never really fit into depthcharge's
driver model anyway.
CQ-DEPEND=CL:194064
BUG=None
TEST=Built and booted nyan, link, and peach_pit and verified that timer values
in cbmem were reasonable. Built for nyan_big, nyan_blaze and daisy.
BRANCH=None
Change-Id: Ia7953cfece57524262a6c7d6537082af7a00f4d6
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/194058
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
These drivers need to be ready right away and never really fit into the
depthcharge driver model anyway.
CQ-DEPEND=CL:194063
BUG=None
TEST=Built and booted on nyan and peach_pit. Built for nyan_big, nyan_blaze,
and daisy.
BRANCH=None
Change-Id: I9570dee53c57d42ef4cd956f66a878ce39a2dc20
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/194057
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
I don't even know why these were off (at least the XHCI one)... well,
now they aren't. As far as coreboot is concerned, Snow is just a test
platform anyway, and these are useful to have.
BUG=None
TEST=It works.
Change-Id: Ibef65e8be6cf752276ffc87df6dda04869dd97cd
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/193733
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
We've recently fixed a problem where an external hard drive would choke
due to one too many CLEAR_FEATURE(HALT) commands in the XHCI stack with
"libpayload: usb: xhci: Fix STALL endpoint handling". Clearing stall
conditions from within the transfer function is wrong in general... this
is really something that is host controller agnostic and should be left
to the higher-level driver to decide. The mass storage driver (the only
one that should really encounter stalls right now) already contains the
proper amount of clear_stall() calls... any more than that is redundant
and as we found out potentially dangerous.
This patch removes automatic clear stalls from UHCI and OHCI drivers as
well to make things consistent between host controllers.
BUG=chromium:192866
TEST=None. I could borrow the original hard drive from Shawn and compile
a Snow to only use the OHCI driver to reproduce/verify this, but alas, I
am lazy (and it's really not that important).
Change-Id: Ie1e4d4d2d70fa4abf8b4dabd33b10d6d4012048a
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/193732
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
This patch combines a few minor fixes and refactoring to the various
host controller and root hub drivers to ensure they all do the right
thing on a call to usb_exit(). It puts a usb_detach_device(0) call
into detach_controller() so that the HCD doesn't need to remember to
tear down the root hub itself, and makes sure all root hubs properly
detach the subtree of devices connected to their ports first (as
generic_hub and by extension XHCI had already been doing).
It also fixes up some missing free() calls and replaces most 'ptr =
malloc(); if (!ptr) fatal()' idioms with the new x(z)alloc().
BUG=chromium:343415
TEST=Tested EHCI on Big and OHCI, EHCI, and XHCI on Snow. Could not test
UHCI (unless anyone volunteers to port coreboot to a ZGB? ;) ), but the
changes are really tame.
Change-Id: I6eca51ff2685d0946fe4267ad7d3ec48ad7fc510
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/193731
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
This patch enables the OHCI driver to use DMA memory, which is necessary
for ARM systems where DMA devices are not cache coherent. I really only
need this to test some later OHCI changes, but it was easy enough...
copied almost verbatim from ehci.c.
BUG=chromium:343415
TEST=Works on Snow.
Change-Id: Ia717eef28340bd6182a6782e83bfdd0693cf0db1
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/193730
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
Turns out that when you clear 28 bits starting with bit 3, you leave bit
31 standing. Ooops...
This shouldn't really matter since that bit is reserved/SBZ in CLIDR
anyway, but it's still nice to fix it. This whole thing should really be
an AND for clarity anyway in my opinion.
Bug found in upstream NetBSD (who would've thought...).
BUG=None
TEST=Still boots.
Change-Id: Ic826e82d58fd1ce984971afea3dfa9296f746d9f
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/193300
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: Gabe Black <gabeblack@chromium.org>
This patch adds the 10ms TRSTRCY delay between a reset and the following
Set Address command that is required by the USB 2.0 specification to the
EHCI root hub driver. The generic_hub driver that's used for XHCI and
external hubs already included this delay. This is such a glaring
violation of the spec that I'm really amazed how many USB 2.0 devices
we tested before seemed perfectly fine with responding to a Set Address
within 2 microframes of the reset...
It also increases the port reset hold delay by one millisecond to avoid
an ugly race condition on Tegra SoCs: they decided to time the 50ms
themselves instead of relying on the CPU to do it (fair enough), and to
automatically transition Port Reset to 0 and Port Enable to 1 after that
(bad idea). If the CPU's read-modify-write to clear Port Reset races
exactly with the host controller setting Port Enable, we may end up
clearing the bit again and going into the companion controller handoff
path later on. The added millisecond shouldn't cause any problems for
other host controllers and is not a big deal compared to other delays in
this code path.
BUG=chrome-os-partner:26749
TEST=Run several dozen reboot loops with The USB Stick of Death (TM) (a
blue Patriot XT 13fe:5200 with bcdDevice = 1.00), make sure it always
gets detected correctly.
Change-Id: Idd3329ae6d7e5e1c07a84a5475549b3459836b31
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/189872
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Reviewed-by: Jim Lin <jilin@nvidia.com>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
- Remove the call to clear_stall in xhci_reset_endpoint because we will
call clear_stall from the mass-storage driver.
- Remove the xhci_reset_endpoint call from xhci_bulk on STALL since we
will reset on the next transfer anyway.
- Remove the clear_halt parameter from xhci_bulk since it's now unused.
BUG=chrome-os-partner:26687
TEST=Manual on Rambi w/ USB_DEBUG enabled in libpayload. Boot with SanDisk
Extreme USB 3.0 drive in USB 3.0 port, verify that after STALL is
encountered reset succeeds and device is initialized without extra
delay.
BRANCH=Rambi
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Change-Id: I852b87621861109e596ec24b78a8f036d796ff14
Reviewed-on: https://chromium-review.googlesource.com/192866
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Julius Werner <jwerner@chromium.org>
BUG=None
TEST=emerge-nyan_blaze chromeos-coreboot-nyan builds OK
Change-Id: I707a5efdbdbc573ef73cd366bb7c90fa7c4e74c2
Signed-off-by: Tom Warren <twarren@nvidia.com>
Reviewed-on: https://chromium-review.googlesource.com/190722
Reviewed-by: Julius Werner <jwerner@chromium.org>
This creates a new PL011 config variable which avoids the
infinite busy wait on serial_putchar() because the register
mapping is not compatible with current implementation.
BUG=None
BRANCH=none
TEST=printf() works on the PL011 based ARMv8 foundation model
Change-Id: I9feda35a50a3488fc504d1561444161e0889deda
Signed-off-by: Marcelo Povoa <marcelogp@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/187020
Reviewed-by: David Hendricks <dhendrix@chromium.org>
If a port is connected before and after an xhci controller reset, the
PORTSC CSC bit may not be asserted. Add an additional check in
xhci_rh_port_status_changed for the PRC bit so we can correctly handle
ports in such a state.
BUG=chrome-os-partner:24090
TEST=Manual on Rambi:
- Boot Chromium OS from USB 3.0 drive
- Issue 'reboot' on command line
- Boot from USB 3.0 drive again successfully
Also --
- Boot Chromium OS from USB 3.0 drive
- Issue 'reboot' on command line
- Boot Chromium OS from eMMC
- Issue 'reboot' on command line
- Boot from USB 3.0 drive again successfully
Also, verify that USB ports continue to function correctly, and USB 3.0
device is always detected in Chromium OS as a superspeed device.
BRANCH=Rambi
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Change-Id: I2d623aae647ab13711badd7211ab467afdc69548
Reviewed-on: https://chromium-review.googlesource.com/189394
Reviewed-by: Julius Werner <jwerner@chromium.org>
The generic roothub reset port function is overly broad and does some
things which may be undesirable, such as issuing multiple resets to a
port if the reset is deemed to have finished too quickly. Remove the
generic function and replace it with a controller-specific function,
currently only implemented for xhci.
BUG=chrome-os-partner:24090
TEST=Manual on Rambi. Verify that USB 3.0 media is found + bootable on
cold boot.
BRANCH=Rambi.
Change-Id: Id46f73ea3341d4d01d2b517c6bf687402022d272
Signed-off-by: Shawn Nematbakhsh <shawnn@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/189495
Reviewed-by: Julius Werner <jwerner@chromium.org>
TEST=Booted nyan in normal and recovery mode. Created a map, filled it with some
chars, then verified they can be read from the pointer returned.
BUG=chrome-os-partner:25587
BRANCH=None
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Tested-by: Daisuke Nojiri <dnojiri@chromium.org>
Change-Id: Id1f1be4f6d2d5734d87bf3452d4806d0fe3fda88
Reviewed-on: https://chromium-review.googlesource.com/188894
Reviewed-by: Julius Werner <jwerner@chromium.org>
Commit-Queue: Daisuke Nojiri <dnojiri@chromium.org>
Tested-by: Daisuke Nojiri <dnojiri@chromium.org>
Use the SPSR to extract and inject CPSR values when an exception happens and
pass that information to exception hooks.
The register structure GDB expects when using its remote protocol has a spot
for the CPSR.
BUG=None
TEST=Built and booted on link, nyan.
BRANCH=None
Change-Id: Id950fb09d72fb0f81e4eef2489c0849ce5dd8aca
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/180253
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Pass -ggdb3 to the compiler when building libpayload, -ggdb so that it uses
"the most expressive format available", and 3 so that the debugging level is
set to 3, the highest value currently supported. The debugging information can
be stripped by the payload consuming the library, and will definitely be
stripped by cbfstool when installing that payload into an image.
BUG=None
TEST=Built and booted on link, nyan.
BRANCH=None
Change-Id: Ifd6c4a928fbb0b9fa9b3b2e0ea298abff31baf3b
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/180252
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.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
Change-Id: Iab659365864a3055159a50b8f6e5c44290d3ba2b
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/179602
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
To find the coreboot tables, the payload has historically searched for their
signature in a predefined region of memory. This is a little clumsy on x86,
but it works because you can assume certain regions are RAM. Also, there are
areas which are set aside for the firmware by convention. On x86 there's a
forwarding entry which goes in one of those fairly small conventional areas
and which points to the CBMEM area at the end of memory.
On ARM there aren't areas like that, so we've left out the forwarding entry and
gone directly to CBMEM. RAM may not start at the beginning of the address space
or go to its end, and that means there isn't really anywhere fixed you can put
the coreboot tables. That's meant that libpayload has to be configured on a
per board basis to know where to look for CBMEM.
Now that we have boards that don't have fixed amounts of memory, the location
of the end of RAM isn't fixed even on a per board level which means even that
workaround will no longer cut it.
This change makes coreboot pass the location of the coreboot tables to
libpayload using r0, the first argument register. That means we'll be able to
find them no matter where CBMEM is, and we can get rid of the per board search
ranges.
We can extend this mechanism to x86 as well, but there may be more
complications and it's less necessary there. It would be a good thing to do
eventually though.
BUG=None
TEST=Built and booted on nyan. Changed the size of memory and saw that the
payload could still find the coreboot tables where before it couldn't. Built
for pit, snow, and big.
BRANCH=None
Change-Id: I7218afd999da1662b0db8172fd8125670ceac471
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/185572
Reviewed-by: Julius Werner <jwerner@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
This patch changes the ENTRY() macro in asm.h to create a new section
for every assembler function, thus providing dcache_clean/invalidate_all
and friends with the same --gc-sections goodness that our C functions
have. This requires a few minor changes of moving around data (to make
sure it ends up in the right section) and changing some libgcc functions
(which apparently need to have two names?), but nothing serious.
(You may note that some of our assembly functions have data, sometimes
even writable, within the same .text section. This has been this way
before and I'm not looking to change it for now, although it's not
totally clean. Since we don't enforce read-only sections through paging,
it doesn't really hurt.)
BUG=None
TEST=Nyan and Snow still boot. Confirm dcache_invalidate_all is not
output into any binary anymore since no one actually uses it.
Change-Id: I247b29d6173ba516c8dff59126c93b66f7dc4b8d
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/183891
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.
CQ-DEPEND=CL:183877
BUG=None
TEST=Snow and Nyan still boot. Time between entering romstage main() and
the configure_l2ctlr() call on Nyan drops from 390ms to 0.3ms. Even with
icache turned on the old implementation took 7.8ms since it cleared the
cache multiple times with a slow algorithm.
Change-Id: I7bb4995af8184f6383f8e3b1b870b0662bde8bd4
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/183890
This patch fixes the remaining few bugs in our shiny new cache iteration
by set/way/level algorithm to actually make it work: It makes it start
from cache level 0 (previously it would always start at LoC and be
"done" instantly), fixes up the two shifts that isolate the set bits at
the end (which didn't seem to account for the fact that the first shift
affects the second), and throws an S bit on that last shift so that it
actually affects the conditionals after it.
In addition, also moves the next_level block to the top so that we can
share (and thus eliminate) some code at initialization, and turns the
whole thing into a thrice-instantiated macro to create functions that
fit our existing interface.
BUG=None
TEST=Ran with cache_test code (see separate CL) and closely examined the
resulting output. Made sure results look as expected (iterating through
all sets (inner) and ways (outer) for L1 and then L2 cache, extracting
the right numbers from CLIDR and CCSIDR, not touching anything twice).
Time for a single dcache_clean_invalidate_all() on Nyan drops from 3.7ms
to 0.3ms.
Change-Id: I1338a589cbb37d74ea6e7a3d4f67ff827e24edbe
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/183879
Reviewed-by: Stefan Reinauer <reinauer@google.com>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
This change makes it possible for vboot to avoid an
exploit that could cause involuntary switch to dev mode.
It gives depthcharge/vboot some information on the
type of input device that generated a key.
BUG=chrome-os-partner:21729
TEST=manually tested for panther
BRANCH=none
CQ-DEPEND=CL:182420,CL:182241,CL:182946
Change-Id: I87bdac34bfc50f3adb0b35a2c57a8f95f4fbc35b
Reviewed-on: https://chromium-review.googlesource.com/182357
Reviewed-by: Luigi Semenzato <semenzato@chromium.org>
Tested-by: Luigi Semenzato <semenzato@chromium.org>
Commit-Queue: Luigi Semenzato <semenzato@chromium.org>
This code should have been BSD licensed but was checked in with a GPL license.
BUG=chrome-os-partner:24957
TEST=Built libpayload for nyan.
BRANCH=None
Change-Id: Ic939fd21710c1d52b57b84d3038ec0c0ce4443cd
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/182344
Reviewed-by: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.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.
BUG=None
TEST=Builds on Falco.
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>
This patch switches every last part of Coreboot on ARM over to Thumb
mode: libpayload, the internal libgcc, and assorted assembly files. In
combination with the respective depthcharge patch, this will switch to
Thumb mode right after the entry point of the bootblock and not switch
back to ARM until the final assembly stub that jumps to the kernel.
The required changes to make this work include some new headers and
Makefile flags to handle assembly files (using the unified syntax and
the same helper macros as Linux), modifying our custom-written libgcc
code for 64-bit division to support Thumb (removing some stale old files
that were never really used for clarity), and flipping the general
CFLAGS to Thumb (some more cleanup there as well while I'm at it).
BUG=None
TEST=Snow and Nyan still boot.
Change-Id: I80c04281e3adbf74f9f477486a96b9fafeb455b3
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/182212
Reviewed-by: Gabe Black <gabeblack@chromium.org>
The memcpy/memset/memmove assembly implementations have been taken from
U-Boot, which originally got them from Linux. I turns out that they are
actually not that bad, but they could use an update. This patch pulls in
the current Linux upstream versions of those files, removing some old
U-Boot cruft such as checking whether the two pointers in a memcpy() are
equal (really now?) or side-stepping the R8 register because it was used
for special purposes. It also returns to the good old Linux
ENTRY/ENDPROC macros since we have them now anyway, and straightens out
the W() macro in preparation for unified thumb support.
BUG=None
TEST=Snow still boots.
Change-Id: I138af269b423bef0a237759ac29f1ee58ca206a0
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/182179
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
libgcc/macros.h contains some useful assembly macros that are common in
Linux kernel code and facilitate things such as unified ARM/THUMB
assembly. This patch moves it to a more general place where it can be
used by other code as well.
BUG=None
TEST=Snow still boots.
Change-Id: If68e8930aaafa706c54cf9a156fac826b31bb193
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/182178
Reviewed-by: Vincent Palatin <vpalatin@chromium.org>
Some files were accidentally made GPL when they were added to libpayload. This
change changes them over to a BSD license to be in line with the intended
license of libpayload.
BUG=chrome-os-partner:24957
TEST=Built libpayload for nyan.
BRANCH=None
Change-Id: Ia95ac4951b173dcb93cb489705680e7313df3c92
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/182202
Reviewed-by: Ronald Minnich <rminnich@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
This functionality is already available for ARM, so lets add it to x86 as
well. We'll want to be able to hook exceptions when running as a remote GDB
target.
BUG=None
TEST=Booted depthcharge on link with this test code added to its main function:
__asm__ __volatile__(
"pushl %eax\n"
"mov $0, %eax\n"
"mov %eax, %ss\n"
"popl %eax\n"
);
Saw that the state at the point of the exception was printed, and that %eax
and other registers which should have known values had those values. Modified
the exception handler to change %eax in the saved state so that the above code
was correct and return, and saw that depthcharge continued on to boot the
kernel
BRANCH=None
Change-Id: I42f640b08eb9eb86a1bcab3c327f7780191a2eb5
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/179601
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
This function returns the number of microseconds scaled from the number of raw
timer ticks. It accepts a base parameter which is subtracted from the current
time, which makes it easy to keep track of relative times.
BUG=None
TEST=With a corresponding change in depthcharge, built and booted on link.
BRANCH=None
Change-Id: I55f2f9e90c0e12cda430bbe88b044f12b0b563c8
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/179600
Reviewed-by: Ronald Minnich <rminnich@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
This is needed by depthcharge on ARM if coreboot is loading its
ramstage from the RW section of the ROM.
BUG=none
BRANCH=none
TEST=boot depthcharge on pit
Change-Id: I96c6c04a0cee39854b45f2eda169e93461da0694
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: https://chromium-review.googlesource.com/176757
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Stefan Reinauer <reinauer@chromium.org>
Tested-by: Stefan Reinauer <reinauer@chromium.org>
You might want to use the serial hardware for something other than a console,
or you might want to intercede in the serial stream to wrap it in another
protocol. This is what you'd do to send output to GDB while using it to debug
the payload.
BUG=None
TEST=Built and booted on nyan and saw that there was serial output. Built for
pit.
BRANCH=None
Change-Id: I2218c0dbb988dacb64e5bdaf5d92138828eff8b6
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/179559
Reviewed-by: Ronald Minnich <rminnich@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Pull the ACPI GNVS pointer from CBMEM and expose it in
the sysinfo structure for use by payloads.
BUG=chrome-os-partner:24380
BRANCH=none
TEST=build and boot rambi with emmc in ACPI mode
Change-Id: I47c358f33c464a4a01080268fb553705218c940c
Signed-off-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/179900
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
This additional CFLAGS makes it build without have to wrap the make in magic.
BUG=None
TEST=libpayload builds for ARM
BRANCH=None
Change-Id: Ie9a6239e2864734788c5b72f65a7523635ccf75c
Signed-off-by: Ronald G. Minnich <rminnich@google.com>
Reviewed-on: https://chromium-review.googlesource.com/178757
Reviewed-by: Ronald Minnich <rminnich@chromium.org>
Commit-Queue: Ronald Minnich <rminnich@chromium.org>
Tested-by: Ronald Minnich <rminnich@chromium.org>
The C standard considers it legal to return a NULL pointer for zero
length memory allocations, and our malloc implementation does in fact
make use of that. xmalloc() and xzmalloc() should therefore not consider
this case a failure.
Also fixed a minor formatting issue.
BUG=None
TEST=Made sure xmalloc(0) and xmalloc(1000) succeed and
xmalloc(0xffffffff) still dies.
Change-Id: Ib9b75df9458ce2ba75fd0bc0af9814a3323298eb
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/178725
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
The xmalloc wrapper checks whether the malloc succeeded, and if not stops
execution and prints a message. xmalloc always returns a valid pointer. The
xzalloc wrapper does the same thing, but also zeroes the memory before
returning it.
BUG=None
TEST=Used this function in nyan, built and booted on it.
BRANCH=None
Change-Id: I00e7de04a5c368ab3603530b98bd3e3596e10632
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/178001
Reviewed-by: Julius Werner <jwerner@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
If a programming error is detected, die can be used to print a message and
stop execution similar to failing an assert. There's also a "die_if" function
which is conditional.
die functions, like asserts, should be used to trap programming errors and not
when the hardware does something wrong. If all code was written perfectly, no
die function would ever be called. In other words, it would be appropriate to
use die if a function was called with a value that was out of bounds or if
malloc failed. It wouldn't be appropriate if an external device doesn't
respond.
In the future, the die family of functions might print a stack trace or show
other debugging info.
BUG=None
TEST=Used the die_if function in other code and verified that it stops
execution, prints messages like printf, shows file, line, and function
information, and is correctly gated by its condition.
BRANCH=None
Change-Id: I653fc8cb0b4e459522f1b86f7fac280836d57916
Signed-off-by: Gabe Black <gabeblack@google.com>
Reviewed-on: https://chromium-review.googlesource.com/178000
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
The exception_test() mechanism might have been useful when exceptions
were first implemented, but now that they are pretty stable it's really
not necessary anymore (especially not on every single boot in production
Chromebooks). It forces a simple unaligned access, and as we start
having exceptions in stages that might not have paging turned on yet,
it's better to remove that completely.
Also removed the duplicated implementations of SCTLR-stuff and switched
to the existing ones in cache.h.
BUG=None
TEST=Made sure Pit and Nyan still boot and can trigger exceptions in all
stages.
Change-Id: I85e66269f5e2f2dfd3e8aaaa18441493514b62f8
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/177101
Reviewed-by: David Hendricks <dhendrix@chromium.org>
Reviewed-by: Gabe Black <gabeblack@chromium.org>
"Hey guys, I have this awesome idea! How about we put a huge array
filled with 0xa5 into the data segment of our uncompressed romstage
for no particular reason? Give our SPI driver something to do so it
doesn't get too bored, you know?"
Guess it pays off to just hexdump our image and sanity-check it top to
bottom every once in a while...
Also reduces the size because 8K is crazy just to print a bunch of
registers (256 bytes ought to be enough for anybody).
BUG=None
TEST=Triggered an exception, still works as expected (and verified
romstage load size on Nyan is notably smaller now).
Change-Id: Icec0a711a1b5140d2ebcd98338ec638a4b6262fa
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/176762
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Reviewed-by: Ronald Minnich <rminnich@chromium.org>
Reviewed-by: David Hendricks <dhendrix@chromium.org>
This patch removes the -ffixed-r8 CFLAG from the coreboot and libpayload
Makefiles. This seems to be a relic from U-Boot, which uses that
register to keep it's global data structure pointer. There's no reason
for us to throw away a perfectly fine register on this already pretty
constrained architecture.
Also removed a config.h inclusion from the Makefile because that should
really be done inside the C files.
BUG=None
TEST=Nyan still boots.
Change-Id: Ia176c0f323c1be07cddf88fa5488788786a27cdf
Signed-off-by: Julius Werner <jwerner@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/177110
Reviewed-by: Gabe Black <gabeblack@chromium.org>
