Kuo-Jung Su [Thu, 4 Jul 2013 03:40:36 +0000 (11:40 +0800)]
cfi_flash: use buffer length in unmap_physmem()
While the flash_detect_legacy() of drivers/mtd/cfi_flash.c
feed unmap_physmem() with MAP_NOCACHE as 2nd parameter,
the do_spi_flash_read_write() of common/cmd_sf.c
feed unmap_physmem() with the length of the mapped buffer
as 2nd parameter.
It's apparently a bug, and I personally think the 2nd parameter
should be the length of the mapped buffer.
Signed-off-by: Kuo-Jung Su <dantesu@faraday-tech.com> CC: Albert Aribaud <albert.u.boot@aribaud.net> CC: Stefan Roese <sr@denx.de> Signed-off-by: Stefan Roese <sr@denx.de>
ken kuo [Wed, 24 Jul 2013 18:17:11 +0000 (02:17 +0800)]
nds32: Enable FPU if the version of CPU supported
Some version of Andes core support FPU coprocessor,
if this is the case, and toolchain support FPU instruction set,
we should enable it at low level initialization time.
Signed-off-by: Kuan-Yu Kuo <ken.kuoky@gmail.com> Cc: Macpaul Lin <macpaul@gmail.com>
Checking mips32/time.c with checkpatch.pl shows this:
arch/mips/cpu/mips32/time.c:30: WARNING: line over 80 characters
arch/mips/cpu/mips32/time.c:57: ERROR: return is not a function, parentheses are not required
total: 1 errors, 1 warnings, 0 checks, 85 lines checked
Fix the code to make checkpatch.pl happy.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org> Cc: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
Gabor Juhos [Wed, 22 May 2013 03:57:44 +0000 (03:57 +0000)]
MIPS: qemu-malta: bring up ethernet
Qemu emulates a PCNET PCI card for the Malta CoreLV board.
Enable the pcnet driver and add board specific ethernet
initialization function to bring it up. Also enable the
CONFIG_CMD_NET and CONFIG_CMD_PING options.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org> Cc: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
Gabor Juhos [Wed, 22 May 2013 03:57:46 +0000 (03:57 +0000)]
MIPS: start.S: emulate REVISION register for qemu-malta
On the origial Malta boards the REVISION register is
accessible at the 0x1fc00010 address. The contents of
this register gives information about the revision
of the Malta and Core Boards.
This register is used by the Linux kernel to identify
the actual board it is running on. However the register
is not emulated properly by Qemu, so put a hardcoded
value into the flash to make Linux work.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org> Cc: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
Gabor Juhos [Wed, 22 May 2013 03:57:43 +0000 (03:57 +0000)]
net: pcnet: use pci_virt_to_mem to obtain buffer addresses
The pcnet driver uses the pci_phys_to_mem function
to get the memory address of the DMA buffers. This
This assumes an 1:1 mapping between the PCI and
physical memory which is not true on all platforms.
On MIPS platform U-Boot is running within a mapped
memory region, and the pci_phys_to_mem macro can't
be used to obtain the memory address of the buffers.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org> Cc: Daniel Schwierzeck <daniel.schwierzeck@googlemail.com>
Like many other projects, U-Boot has a tradition of including big
blocks of License headers in all files. This not only blows up the
source code with mostly redundant information, but also makes it very
difficult to generate License Clearing Reports. An additional problem
is that even the same lincenses are referred to by a number of
slightly varying text blocks (full, abbreviated, different
indentation, line wrapping and/or white space, with obsolete address
information, ...) which makes automatic processing a nightmare.
To make this easier, such license headers in the source files will be
replaced with a single line reference to Unique Lincense Identifiers
as defined by the Linux Foundation's SPDX project [1]. For example,
in a source file the full "GPL v2.0 or later" header text will be
replaced by a single line:
SPDX-License-Identifier: GPL-2.0+
We use the SPDX Unique Lincense Identifiers here; these are available
at [2].
Note: From the legal point of view, this patch is supposed to be only
a change to the textual representation of the license information,
but in no way any change to the actual license terms. With this patch
applied, all files will still be licensed under the same terms they
were before.
Note 2: The apparent difference between the old "COPYING" and the new
"Licenses/gpl-2.0.txt" only results from switching to the upstream
version of the license which is differently formatted; there are not
any actual changes to the content.
Note 3: There are some recurring questions about linense issues, such
as:
- Is a "All Rights Reserved" clause a problem in GPL code?
- Are files without any license header a problem?
- Do we need license headers at all?
The following excerpt from an e-mail by Daniel B. Ravicher should help
with these:
| Message-ID: <4ADF8CAA.5030808@softwarefreedom.org>
| Date: Wed, 21 Oct 2009 18:35:22 -0400
| From: "Daniel B. Ravicher" <ravicher@softwarefreedom.org>
| To: Wolfgang Denk <wd@denx.de>
| Subject: Re: GPL and license cleanup questions
|
| Mr. Denk,
|
| Wolfgang Denk wrote:
| > - There are a number of files which do not include any specific
| > license information at all. Is it correct to assume that these files
| > are automatically covered by the "GPL v2 or later" clause as
| > specified by the COPYING file in the top level directory of the
| > U-Boot source tree?
|
| That is a very fact specific analysis and could be different across the
| various files. However, if the contributor could reasonably be expected
| to have known that the project was licensed GPLv2 or later at the time
| she made her contribution, then a reasonably implication is that she
| consented to her contributions being distributed under those terms.
|
| > - Do such files need any clean up, for example should we add GPL
| > headers to them, or is this not needed?
|
| If the project as a whole is licensed under clear terms, you need not
| identify those same terms in each file, although there is no harm in
| doing so.
|
| > - There are other files, which include both a GPL license header
| > _plus_ some copyright note with an "All Rights Reserved" clause. It
| > has been my understanding that this is a conflict, and me must ask
| > the copyright holders to remove such "All Rights Reserved" clauses.
| > But then, some people claim that "All Rights Reserved" is a no-op
| > nowadays. License checking tools (like OSLC) seem to indicate this is
| > a problem, but then we see quite a lot of "All rights reserved" in
| > BSD-licensed files in gcc and glibc. So what is the correct way to
| > deal with such files?
|
| It is not a conflict to grant a license and also reserve all rights, as
| implicit in that language is that you are reserving all "other" rights
| not granted in the license. Thus, a file with "Licensed under GPL, All
| Rights Reserved" would mean that it is licensed under the GPL, but no
| other rights are given to copy, modify or redistribute it.
|
| Warm regards,
| --Dan
|
| Daniel B. Ravicher, Legal Director
| Software Freedom Law Center (SFLC) and Moglen Ravicher LLC
| 1995 Broadway, 17th Fl., New York, NY 10023
| (212) 461-1902 direct (212) 580-0800 main (212) 580-0898 fax
| ravicher@softwarefreedom.org www.softwarefreedom.org
Axel Lin [Mon, 8 Jul 2013 06:29:52 +0000 (14:29 +0800)]
nds32: ag101/ag102: Fix setting lastdec and now values
The timer3 counter unit for lastdesc and now values are inconsistent in current
code. The unit of "readl(&tmr->timer3_counter) / (CONFIG_SYS_CLK_FREQ / 2)" is
second. However, CONFIG_SYS_HZ is defined as 1000 in board config file.
This means the accuracy of "lastdec" and "now" should be in millisecond,
thus fix the equation to set lastdec and now variables accordingly.
ken kuo [Sat, 8 Jun 2013 03:14:12 +0000 (11:14 +0800)]
nds32: Enable the function of passing parameters to Linux
Add a header file, setup.h, which copy from Linux source code,
this file contain structures are used to pass initialisation parameters
to Linux. Enable this function on adp-ag101/adp-ag101p target
Signed-off-by: Kuan-Yu Kuo <ken.kuoky@gmail.com> Cc: Macpaul Lin <macpaul@gmail.com>
ken kuo [Sat, 8 Jun 2013 03:14:09 +0000 (11:14 +0800)]
nds32: Enable two banks of SDRAM on Andes board
The original adp-ag101/adp-ag101p initialize only one bank(64MB)
by default at boot time, but it is not enough for some application,
so increasing to two banks(128M).
Signed-off-by: Kuan-Yu Kuo <ken.kuoky@gmail.com> Cc: Macpaul Lin <macpaul@gmail.com>
Gabor Juhos [Sun, 26 May 2013 10:11:28 +0000 (12:11 +0200)]
block: constify sect_buf argument of ide_write_data
Add a const keyword to the sect_buf argument of
ide_write_data to fix the following warning:
cmd_ide.c: In function '__ide_output_data':
cmd_ide.c:548: warning: passing argument 2 of 'ide_write_data' discards qualifiers from pointer target type
/devel/u-boot.git/include/ide.h:76: note: expected 'ulong *' but argument is of type 'const ulong *'
Also modify the driver-model documentation to
match with the new prototype.
Compile tested only.
Cc: Macpaul Lin <macpaul@andestech.com> Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Gabor Juhos [Sun, 26 May 2013 10:11:27 +0000 (12:11 +0200)]
mmc: ftsdc010_mci: fix build error if CONFIG_FTSDC010_SDIO is not defined
The FTSDC010_DCR_FIFO_RST symbol is conditionally
defined in <faraday/ftsdc010.h> and it is available
available when CONFIG_FTSDC010_SDIO is enabled.
However the actual driver code unconditionally uses
the FTSDC010_DCR_FIFO_RST constant and this causes
build error if CONFIG_FTSDC010_SDIO is not enabled.
The following error happens when compiling for the
adp-ag101 board:
ftsdc010_mci.c: In function 'ftsdc010_request':
ftsdc010_mci.c:178: error: 'FTSDC010_DCR_FIFO_RST' undeclared (first use in this function)
ftsdc010_mci.c:178: error: (Each undeclared identifier is reported only once
ftsdc010_mci.c:178: error: for each function it appears in.)
The patch ensures that the FTSDC010_DCR_FIFO_RST
symbol gets used only if CONFIG_FTSDC010_SDIO is
defined.
Compile tested only.
Cc: Kuo-Jung Su <dantesu@faraday-tech.com> Cc: Macpaul Lin <macpaul@andestech.com> Signed-off-by: Gabor Juhos <juhosg@openwrt.org> Reviewed-by: Kuo-Jung Su <dantesu@faraday-tech.com>
The driver assumes that the bit manipulation macros
are provided by all architectures. This is not the
case for nds32 and it causes a build error like this:
ftsdc010_mci.c: In function 'ftsdc010_clkset':
ftsdc010_mci.c:118: warning: implicit declaration of function 'setbits_le32'
ftsdc010_mci.c:123: warning: implicit declaration of function 'clrbits_le32'
drivers/mmc/libmmc.o: In function `ftsdc010_request':
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:234: undefined reference to `setbits_le32'
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:243: undefined reference to `clrbits_le32'
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:234: undefined reference to `clrbits_le32'
drivers/mmc/libmmc.o: In function `ftsdc010_clkset':
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:118: undefined reference to `clrbits_le32'
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:118: undefined reference to `clrbits_le32'
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:121: undefined reference to `setbits_le32'
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:123: undefined reference to `setbits_le32'
/devel/u-boot.git/drivers/mmc/ftsdc010_mci.c:123: undefined reference to `setbits_le32'
The patch adds bit manipulation macros for the
nds32 architecture to avoid the errors. The macros
are copied from the ARM implementation.
Compile tested only.
Cc: Kuo-Jung Su <dantesu@faraday-tech.com> Cc: Macpaul Lin <macpaul@andestech.com> Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Alison Wang [Mon, 17 Jun 2013 07:30:39 +0000 (15:30 +0800)]
I2C: mxc_i2c: Add support for Vybrid VF610 platform
This patch adds support for Vybrid VF610 platform.
There are some differences between i.MX6 and Vybrid for I2C controller.
(1) The registers' offset are different.
(2) The I2C clock divider values are different.
(3) In I2C control register, the enable/disable/reset bit is inverted for Vybrid comparing to i.MX6.
(4) In I2C status register, the interrupt flag bit is cleared by writing "1" for Vybrid.
For i.MX6, this bit is cleared by writing "0".
(5) In I2C status register, the arbitration lost flag bit is cleared by writing "1" for Vybrid.
For i.MX6, this bit is cleared by writing "0".
Holger Brunck [Mon, 6 May 2013 02:54:38 +0000 (02:54 +0000)]
arm/km: fix u-boot update functionality
Due to the new I2C framework we need to adapt the u-boot update
function. Due to the new framework all i2c leafs behind a mux are
present in the system and not only those who are defined and used. So it
is bus number 5 after the rework.
Heiko Schocher [Thu, 25 Oct 2012 08:32:14 +0000 (10:32 +0200)]
i2c, multibus: get rid of CONFIG_I2C_MUX
CONFIG_I2C_MUX is replaced through the new i2c multibus/multiadapter
framework, configured through CONFIG_SYS_I2C. As CONFIG_I2C_MUX
is only used on the keymile boards, and they are now completely
moved to the new framework, remove CONFIG_I2C_MUX.
Heiko Schocher [Mon, 16 Jan 2012 21:12:23 +0000 (21:12 +0000)]
i2c: add i2c_core and prepare for new multibus support
This Patch introduce the new i2c_core file, which holds
the I2C core functions, for the rework of the multibus/
multiadapter support.
Also adds changes in i2c.h for the new I2C multibus/multiadapter
support. This new support can be activated with the
CONFIG_SYS_I2C define.
Signed-off-by: Heiko Schocher <hs@denx.de> Signed-off-by: Simon Glass <sjg@chromium.org> Cc: Mike Frysinger <vapier@gentoo.org> Cc: Stephen Warren <swarren@wwwdotorg.org>
fs/ext4: fix log2blksz un-initialized error, by cacaulating its value from blksz
The problem here is that uboot can't mount ext4 filesystem with
commit "50ce4c07df1" applied. We use hard-coded "SECTOR_SIZE"(512)
before this commit, now we introduce (block_dev_desc_t *)->log2blksz
to replace this macro. And after we calling do_ls()->fs_set_blk_dev(),
the variable log2blksz is not initialized, which it's not correct.
And this patch try to solve the problem by caculating the value of
log2blksz from variable blksz.
Fix reading ext4_extent_header struture on BE machines. Some 16 bit
fields where converted to 32 bit fields, due to the byte swap on BE
machines the containing value was corrupted. Therefore reading ext4
filesystems on BE machines where broken before.
Signed-off-by: Rommel Custodio <sessyargc+uboot@gmail.com>
[sent via git-send-email; rework commit message] Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com> Reviewed-by: Simon Glass <sjg@chromium.org> Tested-by: Simon Glass <sjg@chromium.org> Tested-by: Lukasz Majewski <l.majewski@samsung.com>
Bo Shen [Wed, 17 Jul 2013 09:14:17 +0000 (17:14 +0800)]
arm: at91sam9n12: change EBI IO to high drive mode
As both the DDR SDRAM and NAND flash connect to EBI on at91sam9n12
and share the lower 8 bits data line. If use low drive of the data
line, it will cause DDR data access corrupt in lower address, so
change the data line to high drive mode
This will fix the Linux kernel boot issue when use Lower address
Signed-off-by: Bo Shen <voice.shen@atmel.com> Acked-by: Andreas Bießmann <andreas.devel@googlemail.com>
The old value of 0x000e0030 will cause ethernet
timeout issues on the sabrelite and possibly other
boards using the KSZ9021.
I have no explanation as to why.
But this is a correct change, the TRM will be updated
to show that 00b is the only valid setting for bits
19-18 of DRAM_RESET.
My thanks go to Liu Hui(Jason) for this information.
net: phy: Set SUPPORTED_1000baseX_Half flag in ESTATUS_1000_XHALF case
Commit de1d786e (add support for Xilinx 1000BASE-X phy (GTX)) introduced the
checking for ESTATUS_1000_XHALF, but it incorrectly sets the
SUPPORTED_1000baseX_Full flag in this case.
Set the SUPPORTED_1000baseX_Half flag instead.
Acked-by: Charles Coldwell <coldwell@gmail.com> Reviewed-By: Sascha Silbe <t-uboot@infra-silbe.de> Reviewed-by: Joe Hershberger <joe.hershberger@ni.com> Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
phy: fix 10/100Mbps operation on 1Gbps-capable links
de1d786 [add support for Xilinx 1000BASE-X phy (GTX)] introduced a
check for the extended status register in order to support
1Gbps-capable PHYs that don't have the 1000BASE-T registers. Since
Extended Status only indicates what the PHY (i.e. the local side) is
capable of, this broke communication with non-1Gbps peers.
Only check the extended status if the 1000BASE-T registers are
actually missing so we don't end up setting speed to 1Gbps even though
the previous test (for the combination of local and peer support for
1Gbps) already indicated we can't do 1Gbps with the current peer.
Simon Glass [Wed, 17 Jul 2013 22:10:59 +0000 (15:10 -0700)]
mkimage: Use board config to get CONFIG_FIT_SIGNATURE value
The value of this config variable is not available to image.h on the host,
since the board config is not actually included. Bring this in so that
mkimage will be built with image-signing support for sandbox at least.
Simon Glass [Wed, 17 Jul 2013 03:09:59 +0000 (20:09 -0700)]
bootm: Move fixup_silent_linux() earlier in the bootm stages
Before the bootm refactor, fixup_silent_linux() was done only in the
monolithic bootm case, not in the subcommand case. With the refactor, it
is done always, which is good. Unfortunately it is done too late, since it
is the PREP or CMDLINE stages that set up the command line for Linux.
Move fixup_silent_linux() into the LOADOS stage, which is where we find
out the OS being used, and can thus decide whether to perform this step.
Dealing with the sys ctrl register should touch only the
relevant bits and not accidently the whole register. On i.MX6,
the sys control register contains bits which shouldn't be reset to
0, e.g. SYS_CTRL[3-0] and IPP_RST_N (SYS_CTRL[23]).
Do this by read/modify/write instead of just a 32bit write.
Dirk Eibach [Wed, 26 Jun 2013 13:55:17 +0000 (15:55 +0200)]
mpc85xx: Add gdsys ControlCenter Digital board
The gdsys ControlCenter Digital board is based on a Freescale P1022 QorIQ SOC.
It boots from SPI-Flash but can be configured to boot from SD-card for
factory programming and testing.
On board peripherals include:
- 2x GbE
- Lattice ECP3 FPGA connected via PCIe
- mSATA RAID1
- USB host
- DisplayPort video output
- Atmel TPM
Reinhard Pfau [Wed, 26 Jun 2013 13:55:13 +0000 (15:55 +0200)]
tpm: add AUTH1 cmds for LoadKey2 and GetPubKey
Extend the tpm library with support for single authorized (AUTH1) commands
as specified in the TCG Main Specification 1.2. (The internally used helper
functions are implemented in a way that they could also be used for double
authorized commands if someone needs it.)
Provide enums with the return codes from the TCG Main specification.
For now only a single OIAP session is supported.
OIAP authorized version of the commands TPM_LoadKey2 and TPM_GetPubKey are
provided. Both features are available using the 'tpm' command, too.
Authorized commands are enabled with CONFIG_TPM_AUTH_SESSIONS. (Note that
this also requires CONFIG_SHA1 to be enabled.)
This commit fixed udelay(), which is necessary (due to HW LCD controller
oddity) for mipi-dsi correct operation. As a result the display orientation
has been switched.
As a follow up, the hwrevision() function has been removed, since it was
used only in this particular place.
Test HW: Trats Exynos4210 rev 0.
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com> Cc: Minkyu Kang <mk7.kang@samsung.com> Acked-by: Minkyu Kang <mk7.kang@samsung.com>
arm:exynos:fix: Fix clock calculation for Exynos4210 based targets.
Provide proper setting for the APLL fout frequency calculation for
Exynos4 based targets (especially Exynos4210 - Trats board).
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com> Cc: Minkyu Kang <mk7.kang@samsung.com> Acked-by: Minkyu Kang <mk7.kang@samsung.com> Acked-by: Simon Glass <sjg@chromium.org> Tested-by: Simon Glass <sjg@chromium.org>
Wolfgang Denk [Sun, 14 Jul 2013 17:42:40 +0000 (19:42 +0200)]
PPC MPC83xx: Fix MPC8323ERDB build warning
Fix:
mpc8323erdb.c: In function 'mac_read_from_eeprom':
mpc8323erdb.c:198:3: warning: dereferencing type-punned pointer will
break strict-aliasing rules [-Wstrict-aliasing]
Signed-off-by: Wolfgang Denk <wd@denx.de>
cc: Timur Tabi <timur@tabi.org>
cc: Kim Phillips <kim.phillips@freescale.com>
Frederic Leroy [Wed, 26 Jun 2013 16:11:25 +0000 (18:11 +0200)]
Fix ext2/ext4 filesystem accesses beyond 2TiB
With CONFIG_SYS_64BIT_LBA, lbaint_t gets defined as a 64-bit type,
which is required to represent block numbers for storage devices that
exceed 2TiB (the block size usually is 512B), e.g. recent hard drives
We now use lbaint_t for partition offset to reflect the lbaint_t change,
and access partitions beyond or crossing the 2.1TiB limit.
This required changes to signature of ext4fs_devread(), and type of all
variables relatives to block sector.
ext2/ext4 fs uses logical block represented by a 32 bit value. Logical
block is a multiple of device block sector. To avoid overflow problem
when calling ext4fs_devread(), we need to cast the sector parameter.
Gerhard Sittig [Sun, 14 Jul 2013 09:27:45 +0000 (11:27 +0200)]
patman: README documentation nits (unit test)
adjust instructions for the invocation of Patman's self test: the -t
flag appears to have a different meaning now, refer to the --test option
for the builtin unit test; adjust a directory location and make sure to
run the file which resides in the source directory
Signed-off-by: Gerhard Sittig <gsi@denx.de> Acked-by: Simon Glass <sjg@chromium.org>
Simon Glass [Wed, 3 Jul 2013 14:11:41 +0000 (07:11 -0700)]
scsi: Correct types of scsi_read/write()
The block device expects to see lbaint_t for the blknr parameter. Change
the SCSI read/write functions to suit.
This fixes the following build warnings for coreboot:
cmd_scsi.c: In function ‘scsi_scan’:
cmd_scsi.c:119:30: error: assignment from incompatible pointer type [-Werror]
cmd_scsi.c:120:32: error: assignment from incompatible pointer type [-Werror]
Tom Rini [Mon, 15 Jul 2013 13:19:39 +0000 (09:19 -0400)]
Revert "MIPS: Jz4740: Add qi_lb60 board support"
The files board/qi/qi_lb60/qi_lb60.c and include/configs/qi_lb60.h were
licensed under the GPL v3 or later, and not v2 or later. As this is
incompatible with the project, revert this board support until the
responsible parties are available to re-license (if so desired) under
GPL v2.
Simon Glass [Wed, 10 Jul 2013 13:25:33 +0000 (09:25 -0400)]
bootm: Handle errors consistently
A recent bootm fix left the error path incomplete. If CONFIG_TRACE is
set it may still not be a supported command, so cover that with the
unsupported subcommand print. Once we handle BOOTM_STATE_OS_GO, we can
just move into the error handler itself, no need for a goto there.
Signed-off-by: Simon Glass <sjg@chromium.org>
[trini: Update slightly based on Simon's changes to also cover
CONFIG_TRACE/BOOTM_STATE_FAKE_OS_GO] Signed-off-by: Tom Rini <trini@ti.com>
common/image.c: Fix regression with ramdisk load/entry points in FIT
A FIT image with a ramdisk that sets the entry or load points to 0x0
must be treated as meaning "leave in place" and NOT "relocate to 0x0".
This regression was introduced in a51ec63.
projects / u-boot.git / log