The gd->cpu pointer is set to an address located in flash when the
probecpu() function is called while U-Boot is executing from flash.
This pointer needs to be updated to point to an address in RAM after
relocation has occurred otherwise Linux may not be able to boot due to
"fdt board" crashing if flash has been erased or changed.
This bug was introduced in commit
a0e2066f39.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Reported-by: Ed Swarthout <Ed.Swarthout@freescale.com>
Tested-by: Kumar Gala <galak@kernel.crashing.org>
Tested on MPC8527DS.
Tested by: Ed Swarthout <Ed.Swarthout@freescale.com>
Neither the MVBLUE nor its underlying architecture implement the
do_irqinfo() function which is required when CONFIG_CMD_IRQ is defined.
This change fixes the following MVBLUE compiler error:
-> ./MAKEALL MVBLUE
Configuring for MVBLUE board...
common/libcommon.a(cmd_irq.o):(.u_boot_cmd+0x24): undefined reference to `do_irqinfo'
make: *** [u-boot] Error 1
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Acked-by: Andre Schwarz <andre.schwarz@matrix-vision.de>
The support for this was silently dropped by a configuration
split during the merge of the imx27lite board support in commit
864aa034f3 (cmd_mtdparts: Move to common
handling of FLASH devices via MTD layer).
Signed-off-by: Detlev Zundel <dzu@denx.de>
Currently, the last block of NAND devices can't be accessed. This patch
fixes this issue by correcting the boundary checking (off-by-one error).
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Scott Wood <scottwood@freescale.com>
Cc: Wolfgang Denk <wd@denx.de>
Commit c7190f02 (retain POR values of non-configured ACR, SPCR, SCCR,
and LCRR bitfields) moved the LCRR assignment to after relocation
to RAM because of the potential problem with changing the local bus
clock while executing from flash.
This change unfortunately adversely affects the boot time, as running
all code up to cpu_init_r can cause significant slowdown.
E.G. on a 8347 board a bootup time increase of ~600ms has been observed:
0.020 CPU: e300c1, MPC8347_PBGA_EA, Rev: 3.0 at 400 MHz, CSB: 266.667 MHz
0.168 RS: 232
0.172 I2C: ready
0.176 DRAM: 64 MB
1.236 FLASH: 32 MB
Versus:
0.016 CPU: e300c1, MPC8347_PBGA_EA, Rev: 3.0 at 400 MHz, CSB: 266.667 MHz
0.092 RS: 232
0.092 I2C: ready
0.096 DRAM: 64 MB
0.644 FLASH: 32 MB
So far no boards have needed the late LCRR setup, so simply revert it
for now - If it is needed at a later time, those boards can either do
their own final LCRR setup in board code (E.G. in board_early_init_r),
or we can introduce a CONFIG_SYS_LCRR_LATE config option to only do
the setup in cpu_init_r.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
There is more and more usage of printing 64bit values,
so enable this feature generally, and delete the
CONFIG_SYS_64BIT_VSPRINTF and CONFIG_SYS_64BIT_STRTOUL
defines.
Signed-off-by: Heiko Schocher <hs@denx.de>
u-boot updates, before starting Linux, the memory node in the
DTS. As this is a "standard" feature, move this functionality
to the cpu.c file for mpc5xxx and mpc512x processors.
Signed-off-by: Heiko Schocher <hs@denx.de>
We are using generic implementation of ffs. This should
be part of Simon's commit 0413cfecea
Here is warning message which this patch removes.
In file included from /tmp/u-boot-microblaze/include/common.h:38,
from cmd_mtdparts.c:87:
/tmp/u-boot-microblaze/include/linux/bitops.h:123:1: warning: "ffs" redefined
In file included from /tmp/u-boot-microblaze/include/linux/bitops.h:110,
from /tmp/u-boot-microblaze/include/common.h:38,
from cmd_mtdparts.c:87:
/tmp/u-boot-microblaze/include/asm/bitops.h:269:1:
warning: this is the location of the previous definition
Signed-off-by: Michal Simek <monstr@monstr.eu>
A typo caused the stack and malloc regions to overlap, which prevented
mem_malloc_init() from returning. This commit makes the memory layout match
the example described in include/configs/microblaze-generic.h
Signed-off-by: Graeme Smecher <graeme.smecher@mail.mcgill.ca>
Signed-off-by: Michal Simek <monstr@monstr.eu>
Previously MAKEALL would always return a value of 0, even if 1 or more
boards did not compile. This change causes MAKEALL to return 0 if all
boards were able to build, otherwise 1.
This change also requires changing the script interpreter from sh to
bash to support bash's PIPESTATUS variable.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Recent commits 1a99de2cb4 and
6a590c5f5f both fixed the same bug in the
same manner. Unfortunately git was "smart" enough to merge both changes
which resulted in some duplicate code.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Reordered code and comment a bit.
Signed-off-by: Wolfgang Denk <wd@denx.de>
commit bd3784df94 deleted some unused
code in do_i2c_mw(), but missed to also remove the respective
commment. This patch fixes this.
Signed-off-by: Heiko Schocher <hs@denx.de>
add the following commands for the manroland boards:
CONFIG_CMDLINE_EDITING
CONFIG_COMMAND_HISTORY
CONFIG_AUTO_COMPLETE
Signed-off-by: Heiko Schocher <hs@denx.de>
This patch introduces a weak default function for post_hotkey_pressed(),
returning 0, for boards without hotkey support. The long-running tests
won't be started on those boards. This default function was implemented
in many board directories. By implementing this weak default we can
remove all those duplicate versions.
Boards with hotkey support, can override this weak default function
by defining one in their board specific code.
Signed-off-by: Stefan Roese <sr@denx.de>
To update the real memory size in the memory node on the
uc101 and mucmc52 boards call fdt_fixup_memory() in
ft_board_setup().
Signed-off-by: Heiko Schocher <hs@denx.de>
Depending on offset, flash size and the number of bad blocks,
get_len_incl_bad may return a too small value which may lead to:
1) If there are no bad blocks, nand_{read,write}_skip_bad chooses the
bad block aware read/write code. This may hurt performance, but does
not have any adverse effects.
2) If there are bad blocks, the nand_{read,write}_skip_bad may choose
the bad block unaware read/write code (if len_incl_bad == *length)
which leads to corrupted data.
Signed-off-by: Daniel Hobi <daniel.hobi@schmid-telecom.ch>
- NfsTimeout() does not correctly update the NFS timeout value which
results in NfsTimeout() only being called once in certain situations.
This can result in the 'nfs' command hanging indefinetly. For
example, the command:
nfs 192.168.0.1:/home/user/file
will not exit until ctrl-c is pressed if 192.168.0.1 does not have an
NFS server running.
This issue is resolved by reinitializting the NFS timeout value inside
NfsTimeout() when a timeout occurs.
- Make the 'nfs' command print the 'T' character when a timeout occurs.
Previously there was no indication that timeouts were occuring.
- Mimic the 'tftpboot' command and when a download fails print "Retry
count exceeded; starting again", and restart the download taking the
'netretry' environment variable into account.
Signed-off-by: Evan Samanas <esamanas@xes-inc.com>
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Tested on TQM8xxL.
Tested by: Wolfgang Denk <wd@denx.de>
Tested on MPC8527DS.
Tested by: Ed Swarthout <Ed.Swarthout@freescale.com>
The help message for the 'bootm' command listed the 'cmdline' and 'bdt'
sub-commands in the wrong order which resulted in the error below when
following the 'help' command's instructions:
"Trying to execute a command out of order"
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Since commit 736fead8fd "Convert SMC911X
Ethernet driver to CONFIG_NET_MULTI API" SMC911X configration options
are called CONFIG_SMC911X rather than CONFIG_DRIVER_SMC911X. Update
README to reflect that change.
Signed-off-by: Mike Rapoport <mike@compulab.co.il>
An off by one error may cause nfs readlink lookup fail if
nfs_path_buff has non-zero data from a previous use.
Loading: *** ERROR: File lookup fail
Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
Commit dce3d79710 updated the zlib code to v0.95; this caused
conflicts when building for the "cradle" board, because the (pretty
generic) preprocessor variable "OFF" was used in multiple files.
Make sure to avoid further conflicts by #undef'ing it in zlib.c
before redefining it.
Signed-off-by: Wolfgang Denk <wd@denx.de>
cc: Giuseppe Condorelli <giuseppe.condorelli@st.com>
cc: Angelo Castello <angelo.castello@st.com>
cc: Alessandro Rubini <rubini-list@gnudd.com>
The Linux kernel build system changed how it compresses things with LZMA
such that the header no longer contains the filesize (it is instead set to
all F's). So if we get a LZMA image that has -1 for the 64bit field,
let's just assume that the decompressed size is unknown and continue on.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Two later additions to the Configuration Option section unfortunately
split the description of Show boot progress and the list of its call outs.
Signed-off-by: Detlev Zundel <dzu@denx.de>
This patch optimizes the direct copy procedure.
Uses get_unaligned() but only in one place.
The copy loop just above this one can also use this
optimization, but I havn't done so as I have not tested if it
is a win there too.
On my MPC8321 this is about 17% faster on my JFFS2 root FS
than the original. No speed test has been performed in u-boot.
Size increase on ppc: 484 bytes
Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>
Add lzop decompression support to the existing lzo bitstream handling
(think gzip versus zlib), and support it for uImage decompression if
CONFIG_LZO is enabled.
Lzop doesn't compress as good as gzip (~10% worse), but decompression
is very fast (~0.7s faster here on a slow ppc). The lzop decompression
code is based on Albin Tonnerre's recent ARM Linux lzo support patch.
Cc: albin.tonnerre@free-electrons.com
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Previously, there was no indication to the user that a FIT image was
successfully created after executing mkimage. For example:
$ mkimage -f uImage.its uImage.itb
DTC: dts->dtb on file "uImage.its"
Adding some additional output after creating a FIT image lets the user
know exactly what is contained in their image, eg:
$ mkimage -f uImage.its uImage.itb
DTC: dts->dtb on file "uImage.its"
FIT description: Linux kernel 2.6.32-rc7-00201-g7550d6f-dirty
Created: Tue Nov 24 15:43:01 2009
Image 0 (kernel@1)
Description: Linux Kernel 2.6.32-rc7-00201-g7550d6f-dirty
Type: Kernel Image
Compression: gzip compressed
Data Size: 2707311 Bytes = 2643.86 kB = 2.58 MB
Architecture: PowerPC
OS: Linux
Load Address: 0x00000000
Entry Point: 0x00000000
Hash algo: crc32
Hash value: efe0798b
Hash algo: sha1
Hash value: ecafba8c95684f2c8fec67e33c41ec88df1534d7
Image 1 (fdt@1)
Description: Flattened Device Tree blob
Type: Flat Device Tree
Compression: uncompressed
Data Size: 12288 Bytes = 12.00 kB = 0.01 MB
Architecture: PowerPC
Hash algo: crc32
Hash value: a5cab676
Hash algo: sha1
Hash value: 168722b13e305283cfd6603dfe8248cc329adea6
Default Configuration: 'config@1'
Configuration 0 (config@1)
Description: Default Linux kernel
Kernel: kernel@1
FDT: fdt@1
This brings the behavior of creating a FIT image in line with creating a
standard uImage, which also prints out the uImage contents after
creation.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
The FIT fit_set_header() function was copied from the standard uImage's
image_set_header() function during mkimage reorganization. However, the
fit_set_header() function is not used since FIT images use a standard
device tree blob header.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
When building a Flattened Image Tree (FIT) the image type needs to be
"flat_dt". Commit 89a4d6b12f introduced a
regression which caused the user to need to specify the "-T flat_dt"
parameter on the command line when building a FIT image. The "-T
flat_dt" parameter should not be needed and is at odds with the current
FIT image documentation.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
According to the PPC reference implementation the udelay() function is
responsible for resetting the watchdog timer as frequently as needed.
Most other architectures do not meet that requirement, so long-running
operations might result in a watchdog reset.
This patch adds a generic udelay() function which takes care of
resetting the watchdog before calling an architecture-specific
__udelay().
Signed-off-by: Ingo van Lil <inguin@gmx.de>
Wolfgang Denk
Peter Tyser
Detlev Zundel
Stefan Roese
Peter Korsgaard
Heiko Schocher
Michal Simek
Graeme Smecher
Daniel Hobi
Evan Samanas
Robert P. J. Day
Mike Rapoport
Ed Swarthout
Mike Frysinger
Joakim Tjernlund
Ingo van Lil
Graeme Russ