Paul Burton [Mon, 7 Apr 2014 15:41:48 +0000 (16:41 +0100)]
pcnet: force ordering of descriptor accesses
The ordering of accesses to the rx & tx descriptors is important, yet
the send & recv functions accessed them via regular structure accesses.
This leaves the compiler with the opportunity to reorder those accesses
or to hoist them outside of loops. Prevent that from happening by using
readl & writel to access the descriptors. As a nice bonus, this removes
the need for the driver to care about endianness.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Paul Burton [Mon, 7 Apr 2014 15:41:47 +0000 (16:41 +0100)]
pcnet: align rx buffers for cache invalidation
The RX buffers are invalidated when a packet is received, however they
were not suitably cache-line aligned. Allocate them seperately to the
pcnet_priv structure and align to ARCH_DMA_MINALIGN in order to ensure
suitable alignment for the cache invalidation, preventing anything else
being placed in the same lines & lost.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
The prior accesses to the descriptor rings & init block via cached
memory had a few issues:
- The memory needs cache flushes or invalidation at the appropriate
times, but was not necessarily aligned on cache line boundaries.
This could lead to data being incorrectly lost or written back to
RAM at the wrong time.
- There are points where ordering of writes to the memory is
important, but because it's cached memory the pcnet controller
would see cache lines written back ordered by address. This could
occasionally lead to hardware seeing descriptors in an incorrect
state.
- Flushing the cache constantly is inefficient.
So, to avoid all of those issues simply access the descriptors & init
block via uncached memory. The MIPS-specific UNCACHED_SDRAM macro is
used to do this (retrieving an address in kseg1) as I could see no
existing generic solution. Since the MIPS Malta board is the only user
of the pcnet driver, hopefully this doesn't matter.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
doc:power:pmic: Add doc entry for PMIC(v2) framework
Well written documentation for PMIC framework was missing and hence
it has been probably difficult and time consuming for other developers
to understand rationale for key design decisions and overall design
structure.
This commit provides proper documentation entry.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com> Acked-by: Simon Glass <sjg@chromium.org>
Enough time has passed since this board was moved to Orphan. Remove.
- Remove board/genietv/*
- Remove include/configs/GENIETV.h
- Clean-up if defined(CONFIG_GENIETV)
- Move the entry from boards.cfg to doc/README.scrapyard
Enough time has passed since these boards were moved to Orphan. Remove.
- Remove board/mbx8xx/*
- Remove include/configs/{MBX.h,MBX860T.h}
- Clean-up if defined(CONFIG_MBX)
- Move the entries from boards.cfg to doc/README.scrapyard
Ian Campbell [Thu, 27 Mar 2014 20:34:13 +0000 (20:34 +0000)]
blackfin: make name_to_gpio be a weak symbol
This required moving it into a C file from the header.
The only user of a non-default name_to_gpio is blackfin, therefore build tested
with the blackfin bct-brettl2 build, which is one I picked at random. Also
tested with a build for the ARM tec board which uses the default/fallback
implementation. Inspection with objdump shows that both have done the right
thing.
This change was requested by Marek during review of the sunxi patch series.
Signed-off-by: Ian Campbell <ijc@hellion.org.uk> Cc: Marek Vasut <marex@denx.de> Cc: Wolfgang Denk <wd@denx.de> Cc: Sonic Zhang <sonic.adi@gmail.com>
Masahiro Yamada [Mon, 31 Mar 2014 03:59:58 +0000 (12:59 +0900)]
board: esd: remove remainders of dead boards
Commit 99bcad18 deleted ADCIOP and DASA_SIM board support
but missed to delete board/esd/adciop and board/esd/dasa_sim.
It also missed to add entries to doc/README.scrapyard.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Cc: Stefan Roese <sr@denx.de> Acked-by: Stefan Roese <sr@denx.de> Acked-by: Matthias Fuchs <matthias.fuchs@esd.eu>
sh_i2c_probe() calls sh_i2c_read(), but read length is 0. So acutally it does not read device at all. This patch prepares dummy buffer and read data into it.
Andreas Bießmann [Thu, 10 Apr 2014 10:52:52 +0000 (12:52 +0200)]
board:tricorder: fixup SPL OOB layout
Commit d016dc42cedbf6102e100fa9ecb58462edfb14f8 changed the layout of BCH8 SW
on omap3 boards. We need to adopt the ecc layout for the nand_spl_simle
driver to avoid wrong ecc errors.
Signed-off-by: Andreas Bießmann <andreas.biessmann@corscience.de> Cc: Thomas Weber <thomas.weber@corscience.de> Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
Andreas Bießmann [Thu, 10 Apr 2014 10:52:51 +0000 (12:52 +0200)]
board:tricorder: enable omap_gpio clocks
Signed-off-by: Andreas Bießmann <andreas.biessmann@corscience.de> Cc: Thomas Weber <thomas.weber@corscience.de> Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
Andreas Bießmann [Thu, 10 Apr 2014 10:52:50 +0000 (12:52 +0200)]
board:tricorder: always work with valid eeprom data
Commit 890880583d84607e36b52a785a96b167728bbf73 introduced EEPROM parsing and
board detection but faild to return a valid tricorder_eeprom struct for backup
case. When pressing S200 while reading EEPROM we ignore the value. We
returned falsely a tricorder_eeprom struct with uninitialized data which is
just garbage.
Initialize it by zeroing the whole structure.
Signed-off-by: Andreas Bießmann <andreas.biessmann@corscience.de> Cc: Thomas Weber <thomas.weber@corscience.de> Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
David Feng [Fri, 14 Mar 2014 06:26:27 +0000 (14:26 +0800)]
arm64 patch: gicv3 support
This patch add gicv3 support to uboot armv8 platform.
Changes for v2:
- rename arm/cpu/armv8/gic.S with arm/lib/gic_64.S
- move smp_kick_all_cpus() from gic.S to start.S, it would be
implementation dependent.
- Each core initialize it's own ReDistributor instead of master
initializeing all ReDistributors. This is advised by arnab.basu
<arnab.basu@freescale.com>.
Signed-off-by: David Feng <fenghua@phytium.com.cn>
Leo Yan [Mon, 31 Mar 2014 01:50:35 +0000 (09:50 +0800)]
ARMv8: fix bug for flush data cache by set/way
When flush the d$ with set/way instruction, it need calculate the way's
offset = log2(Associativity); but in current uboot's code, it use below
formula to calculate the offset: log2(Associativity * 2 - 1), so finally
it cannot flush data cache properly.
York Sun [Mon, 31 Mar 2014 21:40:32 +0000 (14:40 -0700)]
armv8: Flush dcache before switching to EL2
For ARMv8, U-boot has been running at EL3 with cache and MMU enabled.
Without proper setup for EL2, cache and MMU are both disabled (out of
reset). Before switching, we need to flush the dcache to make sure the
data is in the main memory.
Signed-off-by: York Sun <yorksun@freescale.com> Acked-by: David.Feng <fenghua@phytium.com.cn>
Marcel Ziswiler [Tue, 11 Mar 2014 17:43:59 +0000 (18:43 +0100)]
arm: vf610: add enet1 support
This patch contains several changes required for second Ethernet
(enet1/RMII1) port on vf610
- ANADIG PLL5 control definitions required for Ethernet RMII1 clock
- Secondary Ethernet (enet1) MAC RMII1 base address definition
- RMII1 iomux definitions
- VF610_PAD_PTA6__RMII0_CLKOUT iomux definition required for
internal (e.g. crystal-less) Ethernet clocking.
Marcel Ziswiler [Tue, 11 Mar 2014 17:43:57 +0000 (18:43 +0100)]
arm: vf610: fix anadig register struct
The anadig_reg structure started at the wrong offset (fixed by adding
reserved_0x000[4]), was missing some reserved field required for
alignment purpose (reserved_0x094[3] between pll4_denom and pll6_ctrl)
and further contained a too short reserved field causing further miss-
alignment (reserved_0x0C4[7]). Also, rename all the reserved fields
and using a memory offset based scheme for.
Discovered and tested by temporarily putting the following debug
instrumentation into board_init():
struct anadig_reg *anadig = (struct anadig_reg *)ANADIG_BASE_ADDR;
printf("&anadig->pll3_ctrl=0x%p\n", &anadig->pll3_ctrl);
printf("&anadig->pll5_ctrl=0x%p\n", &anadig->pll5_ctrl);
A short loop including a DMB instruction might cause a denial of
service on another processor which executes a CP15 broadcast operation.
Exists on r1, r2, r3, r4 revisions.
Andreas Färber [Mon, 27 Jan 2014 04:48:11 +0000 (05:48 +0100)]
arm: Handle .gnu.hash section in ldscripts
Avoids "could not find output section .gnu.hash" ld.bfd errors on openSUSE.
Cc: Albert Aribaud <albert.u.boot@aribaud.net> Cc: Tom Rini <trini@ti.com> Signed-off-by: Andreas Färber <afaerber@suse.de> Acked-by: Simon Glass <sjg@chromium.org> Tested-by: Simon Glass <sjg@chromium.org>
Chin Liang See [Wed, 5 Mar 2014 04:13:53 +0000 (22:13 -0600)]
socfpga: Adding Clock Manager driver
Clock Manager driver will be called to reconfigure all the
clocks setting based on user input. The input are passed to
Preloader through handoff files
Signed-off-by: Chin Liang See <clsee@altera.com> Cc: Albert Aribaud <albert.u.boot@aribaud.net> Cc: Tom Rini <trini@ti.com> Cc: Wolfgang Denk <wd@denx.de> CC: Pavel Machek <pavel@denx.de> Cc: Dinh Nguyen <dinguyen@altera.com> Acked-by: Pavel Machek <pavel@denx.de>
Marek Vasut [Thu, 3 Apr 2014 17:12:21 +0000 (19:12 +0200)]
arm: mxs: Add support for generating signed BootStream
This patch adds the groundwork for generating signed BootStream, which
can be used by the HAB library in i.MX28. We are adding a new target,
u-boot-signed.sb , since the process for generating regular non-signed
BootStream is much easier. Moreover, the signed bootstream depends on
external _proprietary_ _binary-only_ tool from Freescale called 'cst',
which is available only under NDA.
To make things even uglier, the CST or HAB mandates a kind-of circular
dependency. The problem is, unlike the regular IVT, which is generated
by mxsimage, the IVT for signed boot must be generated by hand here due
to special demands of the CST. The U-Boot binary (or SPL binary) and IVT
are then signed by the CST as a one block. But here is the problem. The
size of the entire image (U-Boot, IVT, CST blocks) must be appended at
the end of IVT. But the size of the entire image is not known until the
CST has finished signing the U-Boot and IVT. We solve this by expecting
the CST block to be always 3904B (which it is in case two files, U-Boot
and the hand-made IVT, are signed in the CST block).
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Stefano Babic <sbabic@denx.de>
This change enables automatically uuid generation by command gpt.
In case of updating partitions layout user don't need to care about
generate uuid manually.
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com> Cc: Minkyu Kang <mk7.kang@samsung.com> Cc: Piotr Wilczek <p.wilczek@samsung.com> Cc: Stephen Warren <swarren@nvidia.com> Cc: Lukasz Majewski <l.majewski@samsung.com> Cc: trini@ti.com
cmd:gpt: randomly generate each partition uuid if undefined
Changes:
- randomly generate partition uuid if any is undefined and CONFIG_RAND_UUID
is defined
- print debug info about set/unset/generated uuid
- update doc/README.gpt
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com> Acked-by: Lukasz Majewski <l.majewski@samsung.com> Cc: Piotr Wilczek <p.wilczek@samsung.com> Cc: Tom Rini <trini@ti.com> Cc: Stephen Warren <swarren@nvidia.com> Cc: Lukasz Majewski <l.majewski@samsung.com>
new commands: uuid and guid - generate random unique identifier
Those commands basis on implementation of random UUID generator version 4
which is described in RFC4122. The same algorithm is used for generation
both ids but string representation is different as below.
char: 0 9 14 19 24 36
xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
UUID: be be be be be
GUID: le le le be be
lib: uuid: add functions to generate UUID version 4
This patch adds support to generate UUID (Universally Unique Identifier)
in version 4 based on RFC4122, which is randomly.
Source: https://www.ietf.org/rfc/rfc4122.txt
Changes:
- new configs:
- CONFIG_LIB_UUID for compile lib/uuid.c
- CONFIG_RANDOM_UUID for functions gen_rand_uuid() and gen_rand_uuid_str()
- add configs dependency to include/config_fallbacks.h for lib uuid.
lib/uuid.c:
- add gen_rand_uuid() - this function writes 16 bytes len binary representation
of UUID v4 to the memory at given address.
- add gen_rand_uuid_str() - this function writes 37 bytes len hexadecimal
ASCII string representation of UUID v4 to the memory at given address.
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com> Cc: Stephen Warren <swarren@nvidia.com> Cc: Lukasz Majewski <l.majewski@samsung.com>
[trini: Add CONFIG_EFI_PARTITION to fallbacks] Signed-off-by: Tom Rini <trini@ti.com>
lib: uuid: code refactor for proper maintain between uuid bin and string
Changes in lib/uuid.c to:
- uuid_str_to_bin()
- uuid_bin_to_str()
New parameter is added to specify input/output string format in listed functions
This change allows easy recognize which UUID type is or should be stored in given
string array. Binary data of UUID and GUID is always stored in big endian, only
string representations are different as follows.
String byte: 0 36
String char: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
string UUID: be be be be be
string GUID: le le le be be
This patch also updates functions calls and declarations in a whole code.
part_efi: move uuid<->string conversion functions into lib/uuid.c
This commit introduces cleanup for uuid library.
Changes:
- move uuid<->string conversion functions into lib/uuid.c so they can be
used by code outside part_efi.c.
- rename uuid_string() to uuid_bin_to_str() for consistency with existing
uuid_str_to_bin()
- add an error return code to uuid_str_to_bin()
- update existing code to the new library functions.
Roger Quadros [Tue, 1 Apr 2014 14:26:40 +0000 (17:26 +0300)]
ahci: Fix data abort on multiple scsi resets.
Commit 2faf5fb82ed6 introduced a regression that causes a data
abort when running scsi init followed by scsi reset.
There are 2 problems with the original commit
1) ALLOC_CACHE_ALIGN_BUFFER() allocates memory on the stack but is
assigned to ataid[port] and used by other functions.
2) The function ata_scsiop_inquiry() tries to free memory which was
never allocated on the heap.
Fix these problems by using tmpid as a temporary cache aligned buffer.
Allocate memory separately for ataid[port] and re-use it if required.
Fixes: 2faf5fb82ed6 (ahci: Fix cache align error messages) Reported-by: Eli Nidam <elini@marvell.com> Signed-off-by: Roger Quadros <rogerq@ti.com>
Haijun.Zhang [Tue, 18 Mar 2014 09:04:23 +0000 (17:04 +0800)]
mmc:eSDHC: Workaround for data timeout issue on Txxx SoC
1. The Data timeout counter value in eSDHC_SYSCTL register is
not working as it should be, so add quirks to enable this
workaround to fix it to the max value 0xE.
2. Add CONFIG_SYS_FSL_ERRATUM_ESDHC111 to enable its workaround.
* Update of patch for change mmc interface by
Pantelis Antoniou <panto@antoniou-consulting.com>
Andrew Gabbasov [Mon, 24 Mar 2014 07:40:41 +0000 (02:40 -0500)]
mmc: fsl_esdhc: fix calculation of timeout for data transactions
Calculation of the timeout value should be based on actual clock value,
written to controller registers. Since mmc->tran_speed is either the
maximum allowed speed, or the preliminary value, that is be not yet
set to registers, the actual timeout, taken by the controller, based
on its clock settings, may be much longer than expected, based on
mmc->tran_speed value. In particular it happens at early initialization
stage, when typical value of mmc->tran_speed is 20MHz or 26MHz, while
actual clock setting, configured in the controller, is 400kHz.
It's more correct to use mmc->clock value for timeout calculation instead.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com> Acked-by: Pantelis Antoniou <panto@antoniou-consulting.com>
Tom Rini [Fri, 7 Feb 2014 19:15:20 +0000 (14:15 -0500)]
mmc: Add 'mmc rst-function' sub-command
Some eMMC chips may need the RST_n_FUNCTION bit set to a non-zero value
in order for warm reset of the system to work. Details on this being
required will be part of the eMMC datasheet. Also add using this
command to the dra7xx README.
* Whitespace fix by panto
Signed-off-by: Tom Rini <trini@ti.com> Acked-by: Pantelis Antoniou <panto@antoniou-consulting.com>
Marek Vasut [Wed, 19 Mar 2014 01:21:36 +0000 (02:21 +0100)]
arm: mxs: Add serial console support into SPL
Add support for serial console into the i.MX23/i.MX28 SPL. A full,
uncrippled serial console support comes very helpful when debugging
various spectacular hardware bringup issues early in the process.
Because we do not use SPL framework, but have our own minimalistic
SPL, which is compatible with the i.MX23/i.MX28 BootROM, we do not
use preloader_console_init(), but instead use a similar function to
start the console. Nonetheless, to avoid blowing up the size of the
SPL binary, this support is enabled only if CONFIG_SPL_SERIAL_SUPPORT
is defined, which is disabled by default.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Stefano Babic <sbabic@denx.de>
Marek Vasut [Fri, 28 Mar 2014 07:31:01 +0000 (08:31 +0100)]
arm: mx5: Avoid hardcoding memory sizes on M53EVK
The DRAM size can be easily detected at runtime on i.MX53. Implement
such detection on M53EVK and adjust the rest of the macros accordingly
to use the detected values.
An important thing to note here is that we had to override the function
for trimming the effective DRAM address, get_effective_memsize(). That
is because the function uses CONFIG_MAX_MEM_MAPPED as the upper bound of
the available DRAM and we don't have gd->bd->bi_dram[0].size set up at
the time the function is called, thus we cannot put this into the macro
CONFIG_MAX_MEM_MAPPED . Instead, we use custom override where we use the
size of the first DRAM block which we just detected.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Fabio Estevam <fabio.estevam@freescale.com> Cc: Stefano Babic <sbabic@denx.de> Cc: Wolfgang Denk <wd@denx.de>
Marek Vasut [Fri, 28 Mar 2014 07:31:00 +0000 (08:31 +0100)]
arm: mx5: Fix memory slowness on M53EVK
Fix memory access slowness on i.MX53 M53EVK board. Let us inspect the
issue: First of all, the i.MX53 CPU has two memory banks mapped at
0x7000_0000 and 0xb000_0000 and each of those can hold up to 1GiB of
DRAM memory. Notice that the memory area is not continuous. On M53EVK,
each of the banks contain 512MiB of DRAM, which makes a total of 1GiB
of memory available to the system.
The problem is how the relocation of U-Boot is treated on i.MX53 . The
U-Boot is placed at the ((start of first DRAM partition) + (gd->ram_size)) .
This in turn poses a problem, since in our case, the gd->ram_size is 1GiB,
the first DRAM bank starts at 0x7000_0000 and contains 512MiB of memory.
Thus, with this algorithm, U-Boot is placed at offset:
0x7000_0000 + 1GiB - sizeof(u-boot and some small margin)
This is past the DRAM available in the first bank on M53EVK, but is still
within the address range of the first DRAM bank. Because of the memory
wrap-around, the data can still be read and written to this area, but the
access is much slower.
There were two ideas how to solve this problem, first was to map both of
the available DRAM chunks next to one another by using MMU, second was to
define CONFIG_VERY_BIG_RAM and CONFIG_MAX_MEM_MAPPED to size of the memory
in the first DRAM bank. We choose the later because it turns out the former
is not applicable afterall. The former cannot be used in case Linux kernel
was loaded into the second DRAM bank area, which would be remapped and one
would try booting the kernel, since at some point before the kernel is started,
the MMU would be turned off, which would destroy the mapping and hang the
system.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Fabio Estevam <fabio.estevam@freescale.com> Cc: Stefano Babic <sbabic@denx.de> Cc: Wolfgang Denk <wd@denx.de>
Marek Vasut [Fri, 28 Mar 2014 07:30:59 +0000 (08:30 +0100)]
arm: mx5: Avoid hardcoding memory sizes on MX53QSB
The DRAM size can be easily detected at runtime on i.MX53. Implement
such detection on MX53QSB and adjust the rest of the macros accordingly
to use the detected values.
An important thing to note here is that we had to override the function
for trimming the effective DRAM address, get_effective_memsize(). That
is because the function uses CONFIG_MAX_MEM_MAPPED as the upper bound of
the available DRAM and we don't have gd->bd->bi_dram[0].size set up at
the time the function is called, thus we cannot put this into the macro
CONFIG_MAX_MEM_MAPPED . Instead, we use custom override where we use the
size of the first DRAM block which we just detected.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Fabio Estevam <fabio.estevam@freescale.com> Cc: Stefano Babic <sbabic@denx.de> Cc: Wolfgang Denk <wd@denx.de>
Marek Vasut [Fri, 28 Mar 2014 07:30:58 +0000 (08:30 +0100)]
arm: mx5: Fix memory slowness on MX53QSB
Fix memory access slowness on i.MX53 MX53QSB board. Let us inspect the
issue: First of all, the i.MX53 CPU has two memory banks mapped at
0x7000_0000 and 0xb000_0000 and each of those can hold up to 1GiB of
DRAM memory. Notice that the memory area is not continuous. On MX53QSB,
each of the banks contain 512MiB of DRAM, which makes a total of 1GiB
of memory available to the system.
The problem is how the relocation of U-Boot is treated on i.MX53 . The
U-Boot is placed at the ((start of first DRAM partition) + (gd->ram_size)) .
This in turn poses a problem, since in our case, the gd->ram_size is 1GiB,
the first DRAM bank starts at 0x7000_0000 and contains 512MiB of memory.
Thus, with this algorithm, U-Boot is placed at offset:
0x7000_0000 + 1GiB - sizeof(u-boot and some small margin)
This is past the DRAM available in the first bank on MX53QSB, but is still
within the address range of the first DRAM bank. Because of the memory
wrap-around, the data can still be read and written to this area, but the
access is much slower.
There were two ideas how to solve this problem, first was to map both of
the available DRAM chunks next to one another by using MMU, second was to
define CONFIG_VERY_BIG_RAM and CONFIG_MAX_MEM_MAPPED to size of the memory
in the first DRAM bank. We choose the later because it turns out the former
is not applicable afterall. The former cannot be used in case Linux kernel
was loaded into the second DRAM bank area, which would be remapped and one
would try booting the kernel, since at some point before the kernel is started,
the MMU would be turned off, which would destroy the mapping and hang the
system.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Fabio Estevam <fabio.estevam@freescale.com> Cc: Stefano Babic <sbabic@denx.de> Cc: Wolfgang Denk <wd@denx.de>
Marek Vasut [Sun, 23 Mar 2014 21:45:40 +0000 (22:45 +0100)]
pci: mx6: Implement power callback
Implement a callback to toggle the slot power supply. The callback
can be overriden in case some more complex power supply for the slot
was implemented in hardware, yet for the usual case, one can define
a GPIO which toggles the power to the slot.
Signed-off-by: Marek Vasut <marex@denx.de> Cc: Stefano Babic <sbabic@denx.de> Cc: Fabio Estevam <fabio.estevam@freescale.com> Cc: Liu Ying <Ying.Liu@freescale.com>
projects / u-boot.git / log