Stephen Warren [Thu, 31 Jul 2014 19:40:07 +0000 (13:40 -0600)]
lib: lmb: fix overflow in __lmb_alloc_base w/ large RAM
If a 32-bit system has 2GB of RAM, and the base address of that RAM is
2GB, then start+size will overflow a 32-bit value (to a value of 0).
__lmb_alloc_base is affected by this; it calculates the minimum of
(start+size of RAM) and max_addr. However, when start+size is 0, it
is always less than max_addr, which causes the value of max_addr not
to be taken into account when restricting the allocation's location.
Fix this by calculating start+size separately, and if that calculation
underflows, using -1 (interpreted as the max unsigned value) as the
value instead, and then taking the min of that and max_addr. Now that
start+size doesn't overflow, it's typically large, and max_addr
dominates the min() call, and is taken into account.
The user-visible symptom of this bug is that CONFIG_BOOTMAP_SZ is ignored
on Tegra124 systems with 2GB of RAM, which in turn causes the DT to be
relocated at the very end of RAM, which the ARM Linux kernel doesn't map
during early boot, and which causes boot failures. With this fix,
CONFIG_BOOTMAP_SZ correctly restricts the relocated DT to a much lower
address, and everything works.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Wed, 30 Jul 2014 22:37:16 +0000 (16:37 -0600)]
ARM: rpi_b: use new generic $bootcmd
Replace the custom $bootcmd with that from <config_distro_bootcmd.h>.
There should be no functional change, since the new generic $bootcmd was
derived strongly from tegra-common-post.h, after which this part of
rpi_b.h was modelled.
The #defines to enable/disable U-Boot commands/features were moved
earlier in rpi_b.h, so they are set up before config_distro_bootcmd.h
is included, since it tests whether certain features should be included
based on those defines.
Signed-off-by: Stephen Warren <swarren@nvidia.com> Acked-by: Simon Glass <sjg@chromium.org>
Stephen Warren [Wed, 30 Jul 2014 22:37:15 +0000 (16:37 -0600)]
ARM: tegra: use new generic $bootcmd
Replace the custom $bootcmd with that from <config_distro_bootcmd.h>.
There should be no functional change, since the new generic $bootcmd was
derived strongly from tegra-common-post.h.
Signed-off-by: Stephen Warren <swarren@nvidia.com> Acked-by: Simon Glass <sjg@chromium.org>
Dennis Gilmore [Wed, 30 Jul 2014 22:37:14 +0000 (16:37 -0600)]
config: introduce a generic $bootcmd
This generic $bootcmd, and associated support macros, automatically
searches a defined set of storage devices (or network protocols) for an
extlinux configuration file or U-Boot boot script in various standardized
locations. Distros that install such a boot config file/script in those
standard locations will get easy-to-set-up booting on HW that enables
this generic $bootcmd.
Boards can define the set of devices from which boot is attempted, and
the order in which they are attempted. Users may later customize this
set/order by edting $boot_targets.
Users may interrupt the boot process and boot from a specific device
simply by executing e.g.:
$ run bootcmd_mmc1
or:
$ run bootcmd_pxe
This patch was originally written by Dennis Gilmore based on Tegra and
rpi_b boot scripts. I have made the following modifications since then:
* Boards must define the BOOT_TARGET_DEVICES macro in order to specify
the set of devices (and order) from which to attempt boot. If needed,
we can define a default directly in config_distro_bootcmd.h.
* Removed $env_import and related variables; nothing used them, and I
think it's better for boards to pre-load an environment customization
file using CONFIG_PREBOOT if they need.
* Renamed a bunch of variables to suit my whims:-)
Signed-off-by: Dennis Gilmore <dennis@ausil.us> Signed-off-by: Stephen Warren <swarren@nvidia.com> Reviewed-by: Marek Vasut <marex@denx.de> Acked-by: Simon Glass <sjg@chromium.org>
- Add 'p1023rds' to the list since commit d0bc5140 dropped
the board support but missed to update this file
- Fill the Commit and Removed Date fields for boards removed
by earlier commits
- Move 'incaip' to keep the list sorted in reverse
chronological order
- Describe the soring rule in the comment block:
"The list should be sorted in reverse chronological order."
- Fix typos in the comment block
Stephen Warren [Fri, 25 Jul 2014 23:30:48 +0000 (17:30 -0600)]
net: BOOTP retry timeout improvements
Currently, the BOOTP code sends out its initial request as soon as the
Ethernet driver indicates "link up". If this packet is lost or not
replied to for some reason, the code waits for a 1s timeout before
retrying. For some reason, such early packets are often lost on my
system, so this causes an annoying delay.
To optimize this, modify the BOOTP code to have very short timeouts for
the first packet transmitted, but gradually increase the timeout each
time a timeout occurs. This way, if the first packet is lost, the second
packet is transmitted quite quickly and hence the overall delay is low.
However, if there's still no response, we don't keep spewing out packets
at an insane speed.
It's arguably more correct to try and find out why the first packet is
lost. However, it seems to disappear inside my Ethenet chip; the TX chip
indicates no error during TX (not that it has much in the way of
reporting...), yet wireshark on the RX side doesn't see any packet.
FWIW, I'm using an ASIX USB Ethernet adapter. Perhaps "link up" is
reported too early or based on the wrong condition in HW, and we should
add some fixed extra delay into the driver. However, this would slow down
every link up event even if it ends up not being needed in some cases.
Having BOOTP retry quickly applies the fix/WAR to every possible
Ethernet device, and is quite simple to implement, so seems a better
solution.
Signed-off-by: Stephen Warren <swarren@nvidia.com> Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Stephen Warren [Wed, 23 Jul 2014 00:06:46 +0000 (18:06 -0600)]
pxe: clear Bootfile before returning
When "pxe boot" downloads the initrd/kernel/DTB, netboot_common() saves
the downloaded filename to global variable BootFile. If the boot
operation is aborted, this global state is not cleared. If "dhcp" is
executed later without any arguments, BootFile is not cleared, and when
the DHCP response is received, BootpCopyNetParams() writes the value into
environment variable bootfile.
This causes the following scenario:
* Boot script executes dhcp; pxe get; pxe boot
* User CTRL-C's the PXE menu, which causes the first menu item to be
booted, which causes some file to be downloaded.
(This boot-on-CTRL-C behaviour is arguably a bug too, but it's a
separate bug and the bug this patch fixes would still exist if the user
simply waited to press CTRL-C until "pxe boot" started downloading
files)
* User CTRL-C's the file downloads, but the filename is still written to
the bootfile environment variable.
* User re-runs the boot command, which in my case executes "dhcp; pxe get;
pxe boot" again, and "dhcp" picks up the saved bootfile environment
variable and proceeds to download a file that it shouldn't.
To solve this, modify the implementation of "pxe get" to clear BootFile
if the whole boot operation fails, which avoids this whole mess.
An alternative would be to modify netboot_common() such that the no-
arguments case explicitly clears the global variable BootFile. However,
that would prevent the following command sequences from working:
doc: README.android-fastboot: Add note about vendor ID
The Android fastboot client only communicates with specific vendor IDs.
This addition to the documentation points out that fact so everyone is
aware that not just any vendor ID will work and where to find the IDs
that will.
Signed-off-by: Clifton Barnes <cabarnes@indesign-llc.com> Cc: Rob Herring <robh@kernel.org>
Implement generalised RSA public exponents for verified boot
Remove the verified boot limitation that only allows a single
RSA public exponent of 65537 (F4). This change allows use with
existing PKI infrastructure and has been tested with HSM-based
PKI.
Change the configuration OF tree format to store the RSA public
exponent as a 64 bit integer and implement backward compatibility
for verified boot configuration trees without this extra field.
Parameterise vboot_test.sh to test different public exponents.
Mathematics and other hard work by Andrew Bott.
Tested with the following public exponents: 3, 5, 17, 257, 39981,
50457, 65537 and 4294967297.
Signed-off-by: Andrew Bott <Andrew.Bott@ipaccess.com> Signed-off-by: Andrew Wishart <Andrew.Wishart@ipaccess.com> Signed-off-by: Neil Piercy <Neil.Piercy@ipaccess.com> Signed-off-by: Michael van der Westhuizen <michael@smart-africa.com> Cc: Simon Glass <sjg@chromium.org>
R8A7780 and R7A7791 of rmobile supports External Clock mode, and these uses
different from Internal Clock mode registers and calculations to the baud rate
setting. This adds function for External Clock mode.
echi-rmobile does not support xHCI. This removes xHCI address
from address table. And this revise a value of CONFIG_USB_MAX_CONTROLLER_COUNT
for lager board and koelsh board.
Stephen Warren [Tue, 1 Jul 2014 18:16:56 +0000 (12:16 -0600)]
test: dfu: add some more test cases
On Tegra, the DFU buffer size is 1M. Consequently, the 8M test always
fails. Add tests for the 1M size, and one byte less as a corner case,
so that some large tests are executed and expected to pass.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Tue, 1 Jul 2014 18:16:55 +0000 (12:16 -0600)]
test: dfu: cleanup before execution
Call cleanup() before running tests too. If a previous test was CTRL-C'd
some stale files may have been left around. dfu-util refuses to receive
a file to a filename that already exists, which results in false test
failures if the files aren't cleaned up first.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Lukasz Majewski [Mon, 23 Jun 2014 07:39:16 +0000 (09:39 +0200)]
thor: defer parsing of device string to IO backend
Commit d4f5ef59cc7 "dfu: defer parsing of device string to IO backend" changed
the function signature of dfu_init_env_entities(). Adjust cmd_thordown.c
to match that change.
Also, apply the same change as commit d6d37d737b58e "dfu: free entities
when parsing fails" to cmd_thordown.c.
Fixes: d4f5ef59cc7 ("dfu: defer parsing of device string to IO backend") Signed-off-by: Lukasz Majewski <l.majewski@samsung.com> Reviewed-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Tue, 10 Jun 2014 22:28:10 +0000 (16:28 -0600)]
test: dfu: script enhancements
Various misc enhancements to dfu_gadget_test.sh:
* After every write (download), perform a write to a different file
with different data. This ensures that the DFU buffer's content is
replaced, so that if the read (upload) succeeds, we know that the
correct data was actually read from the storage device, rather than
simply being left over in the DFU buffer. This requires two alt
setting names to be passed to the script, and a dummy data file to
be generated by dfu_gadget_test_init.sh.
* Fix the assumption that dfu_gadget_test.sh is run from the directory
that contains it, by cd'ing to that directory before invoking
./dfu_gadget_test_init.sh.
* Use $DIR$RCV_DIR consistently, rather than using plain $RCV_DIR in
some places.
* Add 959, 961 test file sizes, to be consistent with having one
more than and one less than all the other "round" sizes 64, 128, and
4096.
* Remove references to $BKP_DIR from dfu_gadget_test_init.sh, since it
isn't used.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Wed, 11 Jun 2014 22:03:34 +0000 (16:03 -0600)]
dfu: allow backend to specify a maximum buffer size
CONFIG_SYS_DFU_DATA_BUF_SIZE may be large to allow for FAT/ext layouts
to transfer large files. However, this means that individual write
operations will take a long time. Allow backends to specify a maximum
buffer size, so that each write operation is limited to a smaller data
block. This prevents the DFU protocol from timing out when e.g. writing
to SPI flash. I would guess that NAND might benefit from setting this
value too, but I can't test that.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Wed, 11 Jun 2014 22:03:33 +0000 (16:03 -0600)]
dfu: defer parsing of device string to IO backend
Devices are not all identified by a single integer. To support
this, defer the parsing of the device string to the IO backed, so that
it can apply the appropriate rules.
SPI devices are specified as controller:chip_select. SPI/SF support will
be added soon.
MMC devices can also be specified as controller[.hwpart][:partition] in
many commands, although we don't support that syntax in DFU.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Wed, 11 Jun 2014 18:48:08 +0000 (12:48 -0600)]
dfu: add write error handling
Fix calls to dfu_write() and dfu_flush() to detect errors in the I/O
itself. This could happen due to problems with the storage medium, or
simply when trying to write a FAT/ext file that is larger than the buffer
dfu_mmc.c maintains for this purpose.
Signal the error by switching the DFU state/status. This will be picked
up by the DFU client when it sends the next DFU request. Note that errors
can't simply be returned from e.g. dnload_request_complete(), since that
function has no way to pass errors back to the DFU client; a call to
dnload_request_complete() simply means that a USB OUT completed.
This error state/status needs to be cleared when the next DFU client
connects. While there is a DFU_CLRSTATUS request, no DFU client seems to
send this. Hence, clear this when selecting the USB alternate setting on
the USB interface.
Finally, dfu.c relies on a call to dfu_flush() to clear up the internal
state of the write transaction. Now that errors in dfu_write() are
detected, dfu_flush() may no longer be called for every transaction.
Separate out the cleanup code into a new function, and call it whenever
dfu_write() fails, as well as from any call to dfu_flush().
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Wed, 11 Jun 2014 18:47:27 +0000 (12:47 -0600)]
dfu: fix some issues with reads/uploads
DFU read support appears to rely upon dfu->read_medium() updating the
passed-by-reference len parameter to indicate the remaining size
available for reading.
dfu_read_medium_mmc() never does this, and the implementation of
dfu_read_medium_nand() will only work if called just once; it hard-codes
the value to the total size of the NAND device irrespective of read
offset.
I believe that overloading dfu->read_medium() is confusing. As such,
this patch introduces a new function dfu->get_medium_size() which can
be used to explicitly find out the medium size, and nothing else.
dfu_read() is modified to use this function to set the initial value for
dfu->r_left, rather than attempting to use the side-effects of
dfu->read_medium() for this purpose.
Due to this change, dfu_read() must initially set dfu->b_left to 0, since
no data has been read.
dfu_read_buffer_fill() must also be modified not to adjust dfu->r_left
when simply copying data from dfu->i_buf_start to the upload request
buffer. r_left represents the amount of data left to be read from HW.
That value is not affected by the memcpy(), but only by calls to
dfu->read_medium().
After this change, I can read from either a 4MB or 1.5MB chunk of a 4MB
eMMC boot partion with CONFIG_SYS_DFU_DATA_BUF_SIZE==1MB. Without this
change, attempting to do that would result in DFU read returning no data
at all due to r_left never being set.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Stephen Warren [Wed, 11 Jun 2014 18:47:26 +0000 (12:47 -0600)]
fs: implement size/fatsize/ext4size
These commands may be used to determine the size of a file without
actually reading the whole file content into memory. This may be used
to determine if the file will fit into the memory buffer that will
contain it. In particular, the DFU code will use it for this purpose
in the next commit.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Simon Glass [Mon, 7 Jul 2014 16:16:38 +0000 (10:16 -0600)]
exynos: spi: Fix calculation of SPI transaction start time
The SPI transaction delay is supposed to be measured from the end of one
transaction to the start of the next. The code does not work that way, so
fix it.
Signed-off-by: Simon Glass <sjg@chromium.org> Tested-by: Ajay Kumar <ajaykumar.rs@samsung.com> Reviewed-by: Jagannadha Sutradharudu Teki <jaganna@xilinx.com>
Simon Glass [Fri, 11 Jul 2014 04:23:26 +0000 (22:23 -0600)]
arm: Set up global data before board_init_f()
At present arm defines CONFIG_SYS_GENERIC_GLOBAL_DATA, meaning that
the global_data pointer is set up in board_init_f(). However it is
actually set up before this, it just isn't zeroed.
If we zero the global data before calling board_init_f() then we
don't need to define CONFIG_SYS_GENERIC_GLOBAL_DATA.
Make this change (on arm32 only) to simplify the init process. I
don't have the ability to test aarch64 yet.
Signed-off-by: Simon Glass <sjg@chromium.org> Acked-by: Tom Rini <trini@ti.com>
When CONFIG_WATCHDOG is defined the board initialization just performs
a WATCHDOG_RESET, an initialization of the watchdog is not done.
This has been modified fot the MPC85xx, the board initialization calls
its watchdog initialitzation allowing for full watchdog configuration
very early in the boot phase.
Signed-off-by: Rainer Boschung <rainer.boschung@keymile.com> Reviewed-by: York Sun <yorksun@freescale.com>
For e500mc cores the watchdog timer period has to be set by means of a
6bit value, that defines the bit of the timebase counter used to signal
a watchdog timer exception on its 0 to 1 transition.
The macro used to set the watchdog period TCR_WP, was redefined for e500mc
to support 6 WP setting.
The parameter (x) given to the macro specifies the prescaling factor of
the time base clock (fTB):
watchdog_period = 1/fTB * 2^x
Signed-off-by: Rainer Boschung <rainer.boschung@keymile.com> Reviewed-by: York Sun <yorksun@freescale.com>
Chin Liang See [Tue, 10 Jun 2014 06:26:52 +0000 (01:26 -0500)]
mmc/dw_mmc: Fix clock divider calculation error for bypass mode
To fix the clock divider calculation error when the controller
clock same as the operating frequency. This is known as bypass
mode. In this mode, the divider should be 0.
Lubomir Rintel [Tue, 10 Jun 2014 18:46:43 +0000 (20:46 +0200)]
bcm2835_sdhci: Add SDHCI_QUIRK_NO_HISPD_BIT flag
Seems like the controller doesn't support the flag. None of the hi-speed cards
I've tried could be read, while they successfully worked with the quirk enabled.
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk> Tested-by: Stephen Warren <swarren@wwwdotorg.org>
- Use "make <board>_defconfig" instead of "make <board>_config".
- Invoke tools/genboardscfg.py to generate boards.cfg when it is
missing.
- Show "Building ${BOARD_NAME} board..." message.
(Prior to Kconfig, instead, mkconfig script displayed
"Configuring for ${BOARD_NAME} board..." but it was removed.)
Without this message, we cannot know which board is currently
being built.
- Do not show "# configuration written to .config".
This message is useless and just annoying for MAKEALL.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Simon Glass <sjg@chromium.org>
The old configuration script is no longer necessary.
Nor is boards.cfg a primary database.
We can generate it with the genboardscfg.py tool
based on the latest Kconfig, defconfig and MAINTAINERS.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Simon Glass <sjg@chromium.org>
Now the primary data for each board is in Kconfig, defconfig and
MAINTAINERS.
It is true boards.cfg is needed for MAKEALL and buildman and might be
useful to brouse all the supported boards in a single database.
But it would be painful to maintain the boards.cfg in sync.
So, this is the solution.
Add a tool to generate the equivalent boards.cfg file based on
the latest Kconfig, defconfig and MAINTAINERS.
We can keep all the functions of MAKEALL and buildman with it.
The best thing would be to change MAKEALL and buildman for not
depending on boards.cfg in the future, but it would take some time.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Simon Glass <sjg@chromium.org>
We have switched to Kconfig and the boards.cfg file is going to
be removed. We have to retrieve the board status and maintainers
information from it.
The MAINTAINERS format as in Linux Kernel would be nice
because we can crib the scripts/get_maintainer.pl script.
After some discussion, we chose to put a MAINTAINERS file under each
board directory, not the top-level one because we want to collect
relevant information for a board into a single place.
TODO:
Modify get_maintainer.pl to scan multiple MAINTAINERS files.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Suggested-by: Tom Rini <trini@ti.com> Acked-by: Simon Glass <sjg@chromium.org>
This commit enables Kconfig.
Going forward, we use Kconfig for the board configuration.
mkconfig will never be used. Nor will include/config.mk be generated.
Kconfig must be adjusted for U-Boot because our situation is
a little more complicated than Linux Kernel.
We have to generate multiple boot images (Normal, SPL, TPL)
from one source tree.
Each image needs its own configuration input.
Usage:
Run "make <board>_defconfig" to do the board configuration.
It will create the .config file and additionally spl/.config, tpl/.config
if SPL, TPL is enabled, respectively.
You can use "make config", "make menuconfig" etc. to create
a new .config or modify the existing one.
Use "make spl/config", "make spl/menuconfig" etc. for spl/.config
and do likewise for tpl/.config file.
The generic syntax of configuration targets for SPL, TPL is:
<target_image>/<config_command>
Here, <target_image> is either 'spl' or 'tpl'
<config_command> is 'config', 'menuconfig', 'xconfig', etc.
When the configuration is done, run "make".
(Or "make <board>_defconfig all" will do the configuration and build
in one time.)
For futher information of how Kconfig works in U-Boot,
please read the comment block of scripts/multiconfig.py.
By the way, there is another item worth remarking here:
coexistence of Kconfig and board herder files.
Prior to Kconfig, we used C headers to define a set of configs.
We expect a very long term to migrate from C headers to Kconfig.
Two different infractructure must coexist in the interim.
In our former configuration scheme, include/autoconf.mk was generated
for use in makefiles.
It is still generated under include/, spl/include/, tpl/include/ directory
for the Normal, SPL, TPL image, respectively.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Simon Glass <sjg@chromium.org>
We are about to switch to Kconfig in the next commit.
But there are something to get done beforehand.
In Kconfig, include/generated/autoconf.h defines boolean
CONFIG macros as 1.
CONFIG_SPL and CONFIG_TPL, if defined, must be set to 1.
Otherwise, when switching to Kconfig, the build log
would be sprinkled with warning messages like this:
warning: "CONFIG_SPL" redefined [enabled by default]
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Reviewed-by: Simon Glass <sjg@chromium.org>
This commit adds:
- arch/${ARCH}/Kconfig
provide a menu to select target boards
- board/${VENDOR}/${BOARD}/Kconfig or board/${BOARD}/Kconfig
set CONFIG macros to the appropriate values for each board
- configs/${TARGET_BOARD}_defconfig
default setting of each board
(This commit was automatically generated by a conversion script
based on boards.cfg)
In Linux Kernel, defconfig files are located under
arch/${ARCH}/configs/ directory.
It works in Linux Kernel since ARCH is always given from the
command line for cross compile.
But in U-Boot, ARCH is not given from the command line.
Which means we cannot know ARCH until the board configuration is done.
That is why all the "*_defconfig" files should be gathered into a
single directory ./configs/.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Simon Glass <sjg@chromium.org>
Since the command name 'make' may not be GNU Make on some platforms
such as FreeBSD, buildman should call scripts/show-gnu-make to get
the command name for GNU MAKE (and error out if it is not found).
Since the command name 'make' may not be GNU Make on some platforms
such as FreeBSD, MAKEALL should call scripts/show-gnu-make to get
the command name for GNU MAKE (and error out if it is not found).
The GNU Make should be searched after parsing options because we want
to allow "MAKEALL -h" even if GNU Make is missing on the system.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Simon Glass <sjg@chromium.org>
projects / u-boot / log