Stephen Warren [Mon, 25 Apr 2016 21:55:42 +0000 (15:55 -0600)]
malloc: improve memalign fragmentation fix
Commit 4f144a416469 "malloc: work around some memalign fragmentation
issues" enhanced memalign() so that it can succeed in more cases where
heap fragmentation is present. However, it did not solve as many cases
as it could. This patch enhances the code to cover more cases.
The alignment code works by allocating more space than the user requests,
then adjusting the returned pointer to achieve alignment. In general, one
must allocate "alignment" bytes more than the user requested in order to
guarantee that alignment is possible. This is what the original code does.
The previous enhancement attempted a second allocation if the padded
allocation failed, and succeeded if that allocation just happened to be
aligned; a fluke that happened often in practice. There are still cases
where this could fail, yet where it is still possible to honor the user's
allocation request. In particular, if the heap contains a free region that
is large enough for the user's request, and for leading padding to ensure
alignment, but has no or little space for any trailing padding. In this
case, we can make a third(!) allocation attempt after calculating exactly
the size of the leading padding required to achieve alignment, which is
the minimal over-allocation needed for the overall memalign() operation to
succeed if the third and second allocations end up at the same location.
This patch isn't checkpatch-clean, since it conforms to the existing
coding style in dlmalloc.c, which is different to the rest of U-Boot.
Signed-off-by: Stephen Warren <swarren@nvidia.com> Reviewed-by: Tom Rini <trini@konsulko.com>
Merge the parsing of layout aware and layout unaware eeprom commands into
one parsing function. With this change, layout aware commands now follow
the eeprom read and eeprom write conventions of making i2c bus and i2c address
parameters optional.
Cc: Heiko Schocher <hs@denx.de> Cc: Marek Vasut <marex@denx.de> Cc: Simon Glass <sjg@chromium.org> Cc: Igor Grinberg <grinberg@compulab.co.il> Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Nikita Kiryanov <nikita@compulab.co.il>
eeprom: use eeprom_execute_command for all eeprom functions
Update eeprom_execute_command() and related code to accommodate both layout
aware and layout unaware functions.
No functional changes.
Cc: Heiko Schocher <hs@denx.de> Cc: Marek Vasut <marex@denx.de> Cc: Simon Glass <sjg@chromium.org> Cc: Igor Grinberg <grinberg@compulab.co.il> Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Nikita Kiryanov <nikita@compulab.co.il>
[trini: Make eeprom_execute_command have ulong for i2c_addr] Signed-off-by: Tom Rini <trini@konsulko.com>
Introduce parse_i2c_bus_addr() to generalize the parsing of i2c bus number and
i2c device address. This is done in preparation for merging layout aware and
layout unaware command parsing into one function.
No functional changes.
Cc: Heiko Schocher <hs@denx.de> Cc: Marek Vasut <marex@denx.de> Cc: Simon Glass <sjg@chromium.org> Cc: Igor Grinberg <grinberg@compulab.co.il> Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Nikita Kiryanov <nikita@compulab.co.il>
cmd: eeprom: add support for layout aware commands
Introduce the (optional) eeprom print and eeprom update commands.
These commands are eeprom layout aware:
* The eeprom print command prints the contents of the eeprom in a human
readable way (eeprom layout fields, and data formatted to be fit for human
consumption).
* The eeprom update command allows user to update eeprom fields by specifying
the field name, and providing the new data in a human readable format (same
format as displayed by the eeprom print command).
* Both commands can either auto detect the layout, or be told which layout to
use.
New CONFIG options:
CONFIG_CMD_EEPROM_LAYOUT - enables commands.
CONFIG_EEPROM_LAYOUT_HELP_STRING - tells user what layout names are supported
Feature API:
__weak int parse_layout_version(char *str)
- override to provide your own layout name parsing
__weak void __eeprom_layout_assign(struct eeprom_layout *layout, int layout_version);
- override to setup the layout metadata based on the version
__weak int eeprom_layout_detect(unsigned char *data)
- override to provide your own algorithm for detecting layout version
eeprom_field.c
- contains various printing and updating functions for common types of
eeprom fields. Can be used for defining custom layouts.
Cc: Heiko Schocher <hs@denx.de> Cc: Marek Vasut <marex@denx.de> Cc: Simon Glass <sjg@chromium.org> Cc: Igor Grinberg <grinberg@compulab.co.il> Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Nikita Kiryanov <nikita@compulab.co.il>
cmd: eeprom: add bus switching support for all i2c drivers
The i2c_init function is always provided when CONFIG_SYS_I2C is
defined. No need to limit ourselves to just one supported I2C driver
(soft_i2c). Update the #ifdef conditions to support bus switching for
all I2C drivers.
Cc: Heiko Schocher <hs@denx.de> Cc: Marek Vasut <marex@denx.de> Cc: Igor Grinberg <grinberg@compulab.co.il> Cc: Tom Rini <trini@konsulko.com> Signed-off-by: Nikita Kiryanov <nikita@compulab.co.il>
Marek Vasut [Sat, 30 Apr 2016 12:45:42 +0000 (14:45 +0200)]
arm: mvebu: a38x: Weed out floating point use
For reason unknown, recently, the DDR init code writers are really fond
of hiding some small floating point operating deep in their creations.
This patch removes one from the Marvell A38x code.
Instead of returning size of chip as float from ddr3_get_device_size()
in GiB units, return it as int in MiB units. Since this would interfere
with the huge switch code in ddr3_calc_mem_cs_size(), rework the code
to match the change.
Before this patch, the cs_mem_size variable could have these values:
( { 16, 32 } x { 8, 16 } x { 0.01, 0.5, 1, 2, 4, 8 } ) / 8 =
{ 0.000000, 0.001250, 0.002500, 0.005000, 0.062500, 0.125000,
0.250000, 0.500000, 1.000000, 2.000000, 4.000000, }
The switch code checked for a subset of the resulting RAM sizes, which
is in range 128 MiB ... 2048 MiB.
With this patch, the cs_mem_size variable can have these values:
( { 16, 32 } x { 8, 16 } x { 0, 512, 1024, 2048, 4096, 8192 } ) / 8 =
{ 0, 64, 128, 256, 512, 1024, 2048, 4096 }
To retain previous behavior, filter out 0 MiB (invalid size), 64 MiB
and 4096 MiB options.
Removing the floating point stuff also saves 1.5k from text segment:
clearfog : spl/u-boot-spl:all -1592 spl/u-boot-spl:text -1592
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Dirk Eibach <dirk.eibach@gdsys.cc> Cc: Stefan Roese <sr@denx.de> Signed-off-by: Stefan Roese <sr@denx.de>
Simon Glass [Sun, 1 May 2016 19:52:44 +0000 (13:52 -0600)]
dm: mmc: test: Add tests for MMC
Add a simple test which checks that a sandbox-emulated SD card can be used
correctly. This tests plumbing through the MMC stack's block-device
implementaion.
Simon Glass [Sun, 1 May 2016 19:52:42 +0000 (13:52 -0600)]
dm: mmc: sandbox: Add an SD-card emulation
Add an emulation of an SD card to sandbox, allowing MMC to be used in tests.
The emulation is very simple, supporting only card detection and reading
test data.
Simon Glass [Sun, 1 May 2016 19:52:40 +0000 (13:52 -0600)]
dm: mmc: Add a way to bind MMC devices with driver model
Binding an MMC device when CONFIG_BLK is enabled requires that a block
device be bound as a child of the MMC device. Add a function to do this.
The mmc_create() method will be used only when DM_BLK is disabled.
Simon Glass [Sun, 1 May 2016 19:52:39 +0000 (13:52 -0600)]
dm: mmc: Implement the MMC functions for block devices
Implement the functions in mmc_legacy.c for driver-model block devices, so
that MMC can use driver model for these. This allows CONFIG_BLK to be enabled
with DM_MMC.
Simon Glass [Sun, 1 May 2016 19:52:35 +0000 (13:52 -0600)]
dm: mmc: Move the device list into a separate file
At present the MMC subsystem maintains its own list of MMC devices. This
cannot work with driver model, which needs to maintain this itself. Move the
list code into a separate 'legacy' file. The core MMC code remains, and will
be shared with the driver-model implementation.
Simon Glass [Sun, 1 May 2016 19:52:30 +0000 (13:52 -0600)]
dm: blk: Add functions to select a hardware partition
The block device uclass does not currently support selecting a particular
hardware partition but this is needed for MMC. Add it so that the blk API
can support MMC properly.
Simon Glass [Sun, 1 May 2016 19:52:27 +0000 (13:52 -0600)]
dm: mmc: Add a function to obtain the block device
The MMC block device is contained within struct mmc. But with driver model
this will not be the case. Add a function to obtain the block device. We
can later implement this for CONFIG_BLK.
Simon Glass [Sun, 1 May 2016 19:52:25 +0000 (13:52 -0600)]
dm: mmc: Move mmc_switch_part() above its callers
This function is defined after it is used. In preparation for making it
static, move it up a little. Also drop the printf() which should not appear
in a driver.
Simon Glass [Sun, 1 May 2016 17:36:29 +0000 (11:36 -0600)]
dm: blk: Add a easier way to create a named block device
Add a function that automatically builds the device name given the parent
and a supplied string. Most callers will want to do this, so putting this
functionality in one place makes more sense.
Simon Glass [Sun, 1 May 2016 17:36:28 +0000 (11:36 -0600)]
dm: blk: Allow blk_create_device() to allocate the device number
Allow a devnum parameter of -1 to indicate that the device number should be
alocated automatically. The next highest available device number for that
interface type is used.
Simon Glass [Sun, 1 May 2016 17:36:26 +0000 (11:36 -0600)]
dm: sata: Add support for driver-model block devices
Add driver-model block-device support to the SATA implementation. This is
just a dummy implementation for now, since the SATA low-level API uses
numbered devices and that doesn't fit with driver model.
Simon Glass [Sun, 1 May 2016 17:36:11 +0000 (11:36 -0600)]
dm: sata: Separate the non-command code into its own file
At present the SATA command code includes both the command-processing code
and the core SATA functions and data structures.
Separate the latter into its own file, adding functions as needed to avoid
the command code accessing data structures directly.
With this commit:
- All CONFIG option are referenced from the non-command code
- The concept of a 'current SATA device' is confined to the command code
This will make it easier to convert this code to driver model.
Simon Glass [Sun, 1 May 2016 17:36:10 +0000 (11:36 -0600)]
dm: ide: Separate the non-command code into its own file
At present the IDE command code includes both the command-processing code
and the core IDE functions and data structures.
Separate the latter into its own file, adding functions as needed to avoid
the command code accessing data structures directly.
With this commit:
- Most CONFIG option are referenced from the non-command code
- The concept of a 'current IDE device' is confined to the command code
This will make it easier to convert this code to driver model.
Simon Glass [Sun, 1 May 2016 17:36:09 +0000 (11:36 -0600)]
dm: scsi: Separate the non-command code into its own file
At present the SCSI command code includes both the command-processing code
and the core SCSI functions and data structures.
Separate the latter into its own file, adding functions as needed to avoid
the command code accessing data structures directly. This functions use the
new legacy block functions.
With this commit:
- There is no CONFIG option referenced from the command code
- The concept of a 'current SCSI device' is confined to the command code
This will make it easier to convert this code to driver model.
Simon Glass [Sun, 1 May 2016 17:36:08 +0000 (11:36 -0600)]
dm: mmc: Add an implementation of the 'devnum' functions
Now that the MMC code accesses devices by number, we can implement this same
interface for driver model, allowing MMC to support using driver model for
block devices.
Simon Glass [Sun, 1 May 2016 17:36:02 +0000 (11:36 -0600)]
dm: scsi: Rename CONFIG_CMD_SCSI to CONFIG_SCSI
This option currently enables both the command and the SCSI functionality.
Rename the existing option to CONFIG_SCSI since most of the code relates
to the feature.
Simon Glass [Sun, 1 May 2016 17:35:54 +0000 (11:35 -0600)]
sandbox: Add string and 16-bit I/O functions
Add outsw() and insw() functions for sandbox, as these are needed by the IDE
code. The functions will not do anything useful if called, but allow the
code to be compiled.
Also add out16() and in16(), required by systemace.
Simon Glass [Sun, 1 May 2016 17:35:53 +0000 (11:35 -0600)]
Allow iotrace byte access to use an address of any size
If an address is used with readb() and writeb() which is smaller than the
expected size (e.g. 32-bit value on a machine with 64-bit addresses), a
warning results. Fix this by adding a cast.
Simon Glass [Sun, 1 May 2016 17:35:52 +0000 (11:35 -0600)]
dm: Rename disk uclass to ahci
This started as 'ahci' and was renamed to 'disk' during code review. But it
seems that this is too generic. Now that we have a 'blk' uclass, we can use
that as the generic piece, and revert to ahci for this.
Simon Glass [Sun, 1 May 2016 17:35:50 +0000 (11:35 -0600)]
dm: sandbox: Add a board for sandbox without CONFIG_BLK
While the driver-model block device support is in progress, it is useful to
build sandbox both with and without CONFIG_BLK. Add a separate board for
the latter.
Simon Glass [Sun, 1 May 2016 17:35:49 +0000 (11:35 -0600)]
Revert "dm: sandbox: Drop the pre-DM host implementation"
Bring this support back so that sandbox can be compiled with CONFIG_BLK. This
allows sandbox to have greater build coverage during the block-device
transition. This can be removed again later.
Peng Fan [Tue, 3 May 2016 02:02:23 +0000 (10:02 +0800)]
dm: gpio: introduce 74x164 driver
Introduce driver to support "fairchild,74hc595" devices.
1. Take linux drivers/drivers/gpio/gpio-74x164.c as reference.
2. Following the naming used in Linux driver with gen_7x164 as the prefix.
3. Enable CONFIG_DM_74X164 to use this driver.
4. Follow Documentation/devicetree/bindings/gpio/gpio-74x164.txt to add device
nodes
5. Tested on i.MX6 UltraLite with 74LV595 using gpio command and oscillograph.
Signed-off-by: Peng Fan <van.freenix@gmail.com> Cc: Simon Glass <sjg@chromium.org> Cc: Masahiro Yamada <yamada.masahiro@socionext.com> Cc: Chin Liang See <clsee@altera.com> Cc: Bhuvanchandra DV <bhuvanchandra.dv@toradex.com> Cc: Daniel Schwierzeck <daniel.schwierzeck@gmail.com> Cc: Fabio Estevam <fabio.estevam@nxp.com> Cc: Stefano Babic <sbabic@denx.de> Reviewed-by: Simon Glass <sjg@chromium.org>
Peng Fan [Tue, 3 May 2016 02:02:22 +0000 (10:02 +0800)]
dm: spi: introduce dm api
Introduce dm_spi_claim_bus, dm_spi_release_bus and dm_spi_xfer
Convert spi_claim_bus, spi_release_bus and spi_xfer to use
the new API.
Signed-off-by: Peng Fan <van.freenix@gmail.com> Cc: Simon Glass <sjg@chromium.org> Cc: Jagan Teki <jteki@openedev.com> Acked-by: Simon Glass <sjg@chromium.org>
Peng Fan [Tue, 3 May 2016 02:02:21 +0000 (10:02 +0800)]
dm: spi: soft_spi: switch to use linux compatible string
1. Support compatible string "spi-gpio" which is used by Linux
Linux use different bindings, so use UBOOT_COMPAT and
LINUX_COMPAT to differentiate them.
2. Introduce SPI_MASTER_NO_RX and SPI_MASTER_NO_TX to handle
no rx or no tx case.
3. Tested on i.MX6 UltraLite board with 74LV595 spi-gpio chip.
Signed-off-by: Peng Fan <van.freenix@gmail.com> Cc: Simon Glass <sjg@chromium.org> Cc: Przemyslaw Marczak <p.marczak@samsung.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Peng Fan [Tue, 3 May 2016 02:02:20 +0000 (10:02 +0800)]
dm: spi: soft_spi bug fix
When doing xfer, should use device->parent, but not device
When doing bit xfer, should use "!!(tmpdout & 0x80)", but not
"(tmpdout & 0x80)"
Signed-off-by: Peng Fan <van.freenix@gmail.com> Cc: Simon Glass <sjg@chromium.org> Cc: Jagan Teki <jteki@openedev.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Stephen Warren [Tue, 19 Apr 2016 22:19:30 +0000 (16:19 -0600)]
video: tegra: refuse to bind to disabled dcs
This prevents the following boot-time message on any board where only the
first DC is in use, yet the DC's DT node is enabled:
stdio_add_devices: Video device failed (ret=-22)
(This happens on at least Harmony, Ventana, and likely any other Tegra20
board with display enabled other than Seaboard).
The Tegra DC's DT node represents a display controller. It may itself
drive an integrated RGB display output, or be used by some other display
controller such as HDMI. For this reason the DC node itself is not
enabled/disabled in DT; the DC itself is considered a shared resource, not
the final (board-specific) display output. The node should instantiate a
display output driver only if the rgb subnode is enabled. Other output
drivers are free to use the DC if they are enabled and their DT node
references the DC's DT node. Adapt the Tegra display drivers' bind()
routine to only bind to the DC's DT node if the RGB subnode is enabled.
Now that the display driver does the right thing, remove the workaround
for this issue from Seaboard's DT file.
Cc: Thierry Reding <treding@nvidia.com> Signed-off-by: Stephen Warren <swarren@nvidia.com> Acked-by: Thierry Reding <treding@nvidia.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Stephen Warren [Tue, 19 Apr 2016 22:19:29 +0000 (16:19 -0600)]
dm: core: allow drivers to refuse to bind
In some cases, drivers may not want to bind to a device. Allow bind() to
return -ENODEV in this case, and don't treat this as an error. This can
be useful in situations where some information source other than the DT
node's main status property indicates whether the device should be
enabled, for example other DT properties might indicate this, or the
driver might query non-DT sources such as system fuses or a version number
register.
Signed-off-by: Stephen Warren <swarren@nvidia.com> Reviewed-by: Simon Glass <sjg@chromium.org>
Stephen Warren [Mon, 11 Apr 2016 16:48:44 +0000 (10:48 -0600)]
buildman: allow more incremental building
One use-case for buildman is to continually run it interactively after
each small step in a large refactoring operation. This gives more
immediate feedback than making a number of commits and then going back and
testing them. For this to work well, buildman needs to be extremely fast.
At present, a couple issues prevent it being as fast as it could be:
1) Each time buildman runs "make %_defconfig", it runs "make mrproper"
first. This throws away all previous build results, requiring a
from-scratch build. Optionally avoiding this would speed up the build, at
the cost of potentially causing or missing some build issues.
2) A build tree is created per thread rather than per board. When a thread
switches between building different boards, this often causes many files
to be rebuilt due to changing config options. Using a separate build tree
for each board would avoid this. This does put more strain on the system's
disk cache, but it is worth it on my system at least.
This commit adds two command-line options to implement the changes
described above; -I ("--incremental") turns of "make mrproper" and -P
("--per-board-out-dir") creats a build directory per board rather than per
thread.
... each once after deleting the buildman result/work directory, and once
"incrementally" after a previous identical invocation.
Signed-off-by: Stephen Warren <swarren@nvidia.com> Reviewed-by: Tom Rini <trini@konsulko.com> Acked-by: Simon Glass <sjg@chromium.org> # v1 Tested-by: Simon Glass <sjg@chromium.org> # v1 Acked-by: Simon Glass <sjg@chromium.org>
projects / u-boot.git / log