ARM (ARM926ejs): add data cache support, tested on magnesium and tx25 board
Enable "cache" command on tx25 and magnesium board and test performance.
Test 1: Loading 127 MB of data from NAND flash into RAM:
Instr. Cache off on on
Data Cache off off on
--------------------------------------------------
magnesium 32,6s 22,5s 30s = x 1,09
tx25 (29MB only) 9,69s 5,05s 8,16s = x 1,19
Test 2: uncompressing a gzipped image from RAM to RAM
(size compressed: 6.5 MiB, uncompressed: 35 MiB):
Instr. Cache off on on
Data Cache off off on
--------------------------------------------------
magnesium 4,25s 2,08s 1,72s = x 2,47
tx25 4,82s 2,04s 1,84s = x 2,62
Portions of this work were supported by funding from
the CE Linux Forum.
This has been tested on at91sam9263 and STN8815.
Again, I didn't check if it has bad effects
on non-arm926 cores.
Initially I had a "done" bit to only set up page tables
at the beginning. However, since the aligmnent requirement
was for the whole object file, this extra integer tool 16kB
in BSS, so I chose to remove it.
Also, note not all boards use PHYS_SDRAM, but it looks like
it's the most used name (more than CONFIG_SYS_DRAM_BASE for
example).
Portions of this work were supported by funding from
the CE Linux Forum.
Mike Frysinger [Thu, 19 Aug 2010 17:05:06 +0000 (13:05 -0400)]
MAKEALL: cut down on duplication of targets
Merge the other significant source of board name duplication with the new
boards.cfg file. I cleaned up most targets, but the ARM and MIPS trees
are such a mess than I didn't bother. If those maintainers care, they can
take are of it.
While we're at it, we can be a bit more clever in the LIST_xxx handling
and avoid duplicating the list names too.
Signed-off-by: Mike Frysinger <vapier@gentoo.org> Reviewed-by: Ben Gardiner <bengardiner@nanometrics.ca> Tested-by: Thomas Chou <thomas@wytron.com.tw>
Mike Frysinger [Wed, 11 Aug 2010 22:52:36 +0000 (18:52 -0400)]
Makefile: restore support for board shortcut targets
The helpful shortcut for doing a board config + make was dropped recently
as it conflicted with some new build code. However, the reason for using
pattern targets originally was to avoid managing a duplicate list of
boards. Since we now have one centralized place for a list of boards (the
new boards.cfg), we don't need a pattern target -- we can generate the
exact list of boards on the fly. So do just that.
When cleaning things up, the top level gitignore file ignores all things
that end with ".depend", but the clean target only deletes files that are
named exactly ".depend". Keep these in sync by having the clean target
punt all files that match the pattern that gitignore is using.
Signed-off-by: Mike Frysinger <vapier@gentoo.org> Acked-by: Detlev Zundel <dzu@denx.de>
Wolfgang Denk [Sun, 20 Jun 2010 21:33:59 +0000 (23:33 +0200)]
New implementation for internal handling of environment variables.
Motivation:
* Old environment code used a pessimizing implementation:
- variable lookup used linear search => slow
- changed/added variables were added at the end, i. e. most
frequently used variables had the slowest access times => slow
- each setenv() would calculate the CRC32 checksum over the whole
environment block => slow
* "redundant" envrionment was locked down to two copies
* No easy way to implement features like "reset to factory defaults",
or to select one out of several pre-defined (previously saved) sets
of environment settings ("profiles")
* No easy way to import or export environment settings
- Variable names starting with '#' are no longer allowed
I didn't find any such variable names being used; it is highly
recommended to follow standard conventions and start variable names
with an alphanumeric character
- "printenv" will now print a backslash at the end of all but the last
lines of a multi-line variable value.
Multi-line variables have never been formally defined, allthough
there is no reason not to use them. Now we define rules how to deal
with them, allowing for import and export.
- Function forceenv() and the related code in saveenv() was removed.
At the moment this is causing build problems for the only user of
this code (schmoogie - which has no entry in MAINTAINERS); may be
fixed later by implementing the "env set -f" feature.
Inconsistencies:
- "printenv" will '\\'-escape the '\n' in multi-line variables, while
"printenv var" will not do that.
- "printenv" output much better readable (sorted)
- faster!
- extendable (additional variable properties can be added)
- new, powerful features like "factory reset" or easy switching
between several different environment settings ("profiles")
Disadvantages:
- Image size grows by typically 5...7 KiB (might shrink a bit again on
systems with redundant environment with a following patch series)
same as "setenv": set (and delete) environment variables
["-f" - force setting even for read-only variables - not
implemented yet.]
- end delete [-f] name
not implemented yet
["-f" - force delete even for read-only variables]
- env save
same as "saveenv": save environment
- env export [-t | -b | -c] addr [size]
export internal representation (hash table) in formats usable for
persistent storage or processing:
-t: export as text format; if size is given, data will be
padded with '\0' bytes; if not, one terminating '\0'
will be added (which is included in the "filesize"
setting so you can for exmple copy this to flash and
keep the termination).
-b: export as binary format (name=value pairs separated by
'\0', list end marked by double "\0\0")
-c: export as checksum protected environment format as
used for example by "saveenv" command
addr: memory address where environment gets stored
size: size of output buffer
With "-c" and size is NOT given, then the export command will
format the data as currently used for the persistent storage,
i. e. it will use CONFIG_ENV_SECT_SIZE as output block size and
prepend a valid CRC32 checksum and, in case of resundant
environment, a "current" redundancy flag. If size is given, this
value will be used instead of CONFIG_ENV_SECT_SIZE; again, CRC32
checksum and redundancy flag will be inserted.
With "-b" and "-t", always only the real data (including a
terminating '\0' byte) will be written; here the optional size
argument will be used to make sure not to overflow the user
provided buffer; the command will abort if the size is not
sufficient. Any remainign space will be '\0' padded.
On successful return, the variable "filesize" will be set.
Note that filesize includes the trailing/terminating '\0'
byte(s).
Usage szenario: create a text snapshot/backup of the current
settings:
Re-import this snapshot, deleting all other settings:
=> env import -d -t ${backup_addr}
- env import [-d] [-t | -b | -c] addr [size]
import external format (text or binary) into hash table,
optionally deleting existing values:
-d: delete existing environment before importing;
otherwise overwrite / append to existion definitions
-t: assume text format; either "size" must be given or the
text data must be '\0' terminated
-b: assume binary format ('\0' separated, "\0\0" terminated)
-c: assume checksum protected environment format
addr: memory address to read from
size: length of input data; if missing, proper '\0'
termination is mandatory
- env default -f
reset default environment: drop all environment settings and load
default environment
- env ask name [message] [size]
same as "askenv": ask for environment variable
- env edit name
same as "editenv": edit environment variable
- env run
same as "run": run commands in an environment variable
- drop default env as implemented now; provide a text file based
initialization instead (eventually using several text files to
incrementally build it from common blocks) and a tool to convert it
into a binary blob / object file.
- It would be nice if we could add wildcard support for environment
variables; this is needed for variable name auto-completion,
but it would also be nice to be able to say "printenv ip*" or
"printenv *addr*"
- Some boards don't link any more due to the grown code size:
DU405, canyonlands, sequoia, socrates.
Wolfgang Denk [Sat, 24 Jul 2010 18:22:02 +0000 (20:22 +0200)]
Make getenv() work before relocation.
So far, getenv() would work before relocation is most cases, even
though it was not intended to be used that way. When switching to a
hash table based implementation, this would break a number of boards.
For convenience, we make getenv() check if it's running before
relocation and, if so, use getenv_f() internally.
Note that this is limited to simple cases, as we use a small static
buffer (32 bytes) in the global data for this purpose.
For this reason, it is also not a good idea to convert all current
uses of getenv_f() into getenv() - some of the existing use cases need
to be able to deal with longer variable values, so getenv_f() is still
needed and recommended for use before relocation.
Wolfgang Denk [Sun, 20 Jun 2010 14:03:45 +0000 (16:03 +0200)]
Remove support for CONFIG_HAS_UID and "forceenv" command
This (undocumented) concept was only in use for the MVSMR and
davinci_schmoogie Sergey Kubushyn <ksi@koi8.net> boards.
Drop it for now. If really needed, it should be reimplemented
later in the context of the new environment command set.
Signed-off-by: Wolfgang Denk <wd@denx.de> Cc: Andre Schwarz <andre.schwarz@matrix-vision.de> Cc: Sergey Kubushyn <ksi@koi8.net> Acked-by: Sergey Kubushyn <ksi@koi8.net>
Wolfgang Denk [Sun, 20 Jun 2010 11:17:12 +0000 (13:17 +0200)]
Add hash table support as base for new environment code
This implementation is based on code from uClibc-0.9.30.3 but was
modified and extended for use within U-Boot.
Major modifications and extensions:
* hsearch() [modified / extended]:
- While the standard version does not make any assumptions about
the type of the stored data objects at all, this implementation
works with NUL terminated strings only.
- Instead of storing just pointers to the original objects, we
create local copies so the caller does not need to care about the
data any more.
- The standard implementation does not provide a way to update an
existing entry. This version will create a new entry or update an
existing one when both "action == ENTER" and "item.data != NULL".
- hsearch_r(): Instead of returning 1 on success, we return the
index into the internal hash table, which is also guaranteed to be
positive. This allows us direct access to the found hash table
slot for example for functions like hdelete().
* hdelete() [added]:
- The standard implementation of hsearch(3) does not provide any way
to delete any entries from the hash table. We extend the code to
do that.
* hexport() [added]:
- Export the data stored in the hash table in linearized form:
Entries are exported as "name=value" strings, separated by an
arbitrary (non-NUL, of course) separator character. This allows to
use this function both when formatting the U-Boot environment for
external storage (using '\0' as separator), but also when using it
for the "printenv" command to print all variables, simply by using
as '\n" as separator. This can also be used for new features like
exporting the environment data as text file, including the option
for later re-import.
- The entries in the result list will be sorted by ascending key
values.
* himport() [added]:
- Import linearized data into hash table. This is the inverse
function to hexport(): it takes a linear list of "name=value"
pairs and creates hash table entries from it.
- Entries without "value", i. e. consisting of only "name" or
"name=", will cause this entry to be deleted from the hash table.
- The "flag" argument can be used to control the behaviour: when
the H_NOCLEAR bit is set, then an existing hash table will kept,
i. e. new data will be added to an existing hash table;
otherwise, old data will be discarded and a new hash table will
be created.
- The separator character for the "name=value" pairs can be
selected, so we both support importing from externally stored
environment data (separated by NUL characters) and from plain text
files (entries separated by newline characters).
- To allow for nicely formatted text input, leading white space
(sequences of SPACE and TAB chars) is ignored, and entries
starting (after removal of any leading white space) with a '#'
character are considered comments and ignored.
- NOTE: this means that a variable name cannot start with a '#'
character.
- When using a non-NUL separator character, backslash is used as
escape character in the value part, allowing for example fo
multi-line values.
- In theory, arbitrary separator characters can be used, but only
'\0' and '\n' have really been tested.
Wolfgang Denk [Sun, 25 Jul 2010 21:08:00 +0000 (23:08 +0200)]
4xx: adjust TEXT_BASE to increase U-Boot image size
On some boards (canyonlands, luan, sequoia) we need more room for the
U-Boot image to allow for new features like the new environment code.
Shift TEXT_BASE as needed.
Signed-off-by: Wolfgang Denk <wd@denx.de> Cc: Stefan Roese <sr@denx.de> Acked-by: Stefan Roese <sr@denx.de>
Wolfgang Denk [Wed, 11 Aug 2010 07:38:31 +0000 (09:38 +0200)]
Cleanup use of CONFIG_SYS_SRAM_BASE and CONFIG_SYS_SRAM_SIZE
Traditionally many boards used local definitions for SRAM base address
and size (like SRAM_BASE, SRAM_LEN and/or SRAM_SIZE), while the (now)
"official" names are CONFIG_SYS_SRAM_BASE and CONFIG_SYS_SRAM_SIZE.
The corresponding code in arch/powerpc/lib/board.c was board specific,
and has never actually been maintained well. Replace this by feature-
specific code and adapt the boards that actually use this.
NOTE: there is still a ton of boards using the old #defines, which
therefor contain incorrect values in bi_sramstart and bi_sramsize.
All respective board maintainers are requested to clean up their
respective configurations. Thanks.
Signed-off-by: Wolfgang Denk <wd@denx.de> Cc: Josef Wagner <Wagner@Microsys.de> Cc: Stefan Roese <sr@denx.de> Cc: Heiko Schocher <hs@denx.de> Cc: Wolfgang Denk <wd@denx.de>
Wolfgang Denk [Fri, 10 Sep 2010 21:04:05 +0000 (23:04 +0200)]
4xx: add missing CONFIG_SYS_SRAM_SIZE definition
A number of boards define CONFIG_SYS_SRAM_BASE but fail to define
CONFIG_SYS_SRAM_SIZE which is needed when cleaning up the code that
prints this information with the bdinfo command.
Add the missing deinitions.
Signed-off-by: Wolfgang Denk <wd@denx.de> Cc: Stefan Roese <sr@denx.de> Acked-by: Stefan Roese <sr@denx.de>
Wolfgang Denk [Sat, 18 Sep 2010 23:03:20 +0000 (01:03 +0200)]
mmc: fix compiler warnings
Commit d2bf29e3 caused a number of compiler warnings:
mmc.c: In function 'mmc_bwrite':
mmc.c:97: warning: format '%x' expects type 'unsigned int', but argument 2 has type 'long unsigned int'
mmc.c:97: warning: format '%x' expects type 'unsigned int', but argument 3 has type 'lbaint_t'
mmc.c: In function 'mmc_bread':
mmc.c:229: warning: format '%x' expects type 'unsigned int', but argument 2 has type 'long unsigned int'
mmc.c:229: warning: format '%x' expects type 'unsigned int', but argument 3 has type 'lbaint_t'
Fix these.
Signed-off-by: Wolfgang Denk <wd@denx.de> Cc: Lei Wen <leiwen@marvell.com>
Daniel Hobi [Wed, 15 Sep 2010 17:46:26 +0000 (19:46 +0200)]
tools/env: allow option "-n" for fw_printenv
In commit bd7b26f8 (Tools: set multiple variable with fw_setenv utility),
the option parsing was changed to getopt_long(3), but option "-n"
of fw_printenv was not included.
This leads to an error message "invalid option -- 'n'" on stderr,
although the output on stdout is correct.
Signed-off-by: Daniel Hobi <daniel.hobi@schmid-telecom.ch>
Graeme Russ [Sun, 22 Aug 2010 06:26:00 +0000 (16:26 +1000)]
x86: Remove Unmaintained Boards
The SC520 CDP boards originally implemented by Daniel Engström are now
very broken. Attempts to contact Daniel via the email address on the
copyright notice have failed. Remove these boards from mainline
Graeme Russ [Sun, 22 Aug 2010 06:25:59 +0000 (16:25 +1000)]
x86: Fix x86 Cold Boot
Commit 077e1958ca4afe12d88043b123ded058c51b89f7 broke the ability of the
x86 port to boot from a cold-reset by removing the initial IDT. Re-
instate the initial IDT to allow cold-booting of x86 boards
Graeme Russ [Sun, 22 Aug 2010 06:25:58 +0000 (16:25 +1000)]
x86: Fix do_go_exec() - const argv[]
Commit 54841ab50c20d6fa6c9cc3eb826989da3a22d934 made the argv parameter
to do_go_exec() const but did not allow for the fact that argv[-1] is
set to point to the global data structure and relies on argv being non-
const.
With this patch, do_go_exec() creates a new copy of the argv array with
an extra element to store global data pointer rather than simply
clobbering an arbitrary memory location.
Wolfgang Denk [Wed, 18 Aug 2010 22:27:33 +0000 (00:27 +0200)]
bedbug_860.c, bedbug_603e.c: Fix return type to silence compile warnings.
commit 47e26b1b "cmd_usage(): simplify return code handling" caused
the following compile warnings:
bedbug_860.c: In function 'bedbug860_do_break':
bedbug_860.c:73: warning: 'return' with a value, in function returning void
bedbug_860.c:121: warning: 'return' with a value, in function returning void
Fix the return type.
Actually these files could need some cleanup - commands should
return proper error codes, and there are coding style issues.
=> To be fixed later.
Correctly set PAD1_FREF_CLK4_REQ and PAD0_FREF_CLK4_OUT to enable and
activate both LEDs while setting pad mux.
Since this increases the line length, this patch also adjusts the white
space in this section of code to allign the pad mux signal description
comments.
Signed-off-by: Ricardo Salveti de Araujo <ricardo.salveti@canonical.com> Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Steve Sakoman [Tue, 24 Aug 2010 17:37:29 +0000 (10:37 -0700)]
ARMV7: OMAP: Overo: Autodetect presence/absence of transceiver on mmc2
An upcoming version of Overo uses a Wifi/BT module with 1.8V signaling,
eliminating the need for an external transceiver to handle the level
shifting. This patch detects whether an external transceiver is present
and adjusts the pinmux settings as appropriate.
Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Steve Sakoman [Fri, 20 Aug 2010 03:09:57 +0000 (20:09 -0700)]
ARMV7: OMAP3: Remove erroneous hard coded sdram setup for 128MB/bank
Upcoming Beagle and Overo revisions use POP memory with 256MB or 512MB
per bank. This patches uses the SDRC settings from x-load or the config
header to set up timing properly.
Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Mans Rullgard [Wed, 14 Apr 2010 15:10:28 +0000 (16:10 +0100)]
ARMV7: OMAP3: Apply Cortex-A8 errata workarounds only on affected revisions
The workarounds for errata 621766 and 725233 should only be applied
on affected Cortex-A8 revisions. Recent chips use r3px cores where
these have been fixed.
Signed-off-by: Mans Rullgard <mans@mansr.com> Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Mans Rullgard [Wed, 14 Apr 2010 14:49:57 +0000 (15:49 +0100)]
ARMV7: OMAP3: Convert setup_auxcr() to pure asm
This function consists entirely of inline asm statements, so writing
it directly in a .S file is simpler. Additionally, the inline asm is
not safe as is, since registers are not guaranteed to be preserved
between asm() statements.
Signed-off-by: Mans Rullgard <mans@mansr.com> Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Mans Rullgard [Wed, 14 Apr 2010 10:08:00 +0000 (11:08 +0100)]
ARMV7: OMAP3: Fix and clean up L2 cache enable/disable functions
On OMAP34xx ES1.0, the L2 enable bit can only be set in secure mode,
so an SMC call to the ROM monitor is required. On later versions,
and on newer devices, this bit is banked and we can set it directly.
The code checked only the ES revision of the chip, and hence incorrectly
used the ROM call on ES1.0 versions of other devices.
This patch adds a check for chip family as well as revision, and also
removes some code duplication between the enable and disable functions.
Signed-off-by: Mans Rullgard <mans@mansr.com> Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Steve Sakoman [Thu, 12 Aug 2010 22:17:37 +0000 (15:17 -0700)]
ARMV7: OMAP: Add detection and support for Beagle C4 revision
This patch enhances the revision detection function and adds
support for the C4 revision. The board revision is printed
and approriate revision specific setup is done automatically.
Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Steve Sakoman [Thu, 12 Aug 2010 22:13:02 +0000 (15:13 -0700)]
ARMV7: OMAP: Add board revision detection for Overo
The latest Overo COM modules encode their revision number on
GPIOs 115, 113, and 112. All boards to date have no pullups on these pins
and hence appear as revision 0.
This patch reads and prints the revision information.
Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Steve Sakoman [Tue, 10 Aug 2010 19:58:39 +0000 (12:58 -0700)]
ARMV7: OMAP: add convenience function to set TWL4030 regulator voltages
This patch adds a function to allow one to easily set the target
voltage for the TWL4030 regulators. It also modifies the existing
code to use this new function. Applicable definitions are moved
out of the driver file and into the header file so that they are
generally accessible
Signed-off-by: Steve Sakoman <steve@sakoman.com> Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Use the MMU hardware to set up 1:1 mappings between physical and virtual
addresses. This allows us to bypass the cache when accessing the flash
without having to do any physical-to-virtual address mapping in the CFI
driver.
The virtual memory mappings are defined at compile time through a sorted
array of virtual memory range objects. When a TLB miss exception
happens, the exception handler does a binary search through the array
until it finds a matching entry and loads it into the TLB. The u-boot
image itself is covered by a fixed TLB entry which is never replaced.
This makes the 'saveenv' command work again on ATNGW100 and other boards
using the CFI driver, hopefully without breaking any rules.
avr32: Use uncached() macro to get an address for SDRAM init
The paging system which is required to set up caching properties has not
yet been initialized when the SDRAM is initialized. So when the
map_physmem() function is converted to return the physical address
unchanged, the SDRAM initialization will break on some boards.
The avr32-specific uncached() macro will return an address which will
always cause uncached accessed to be made. Since this happens in the
board code, using avr32-specific features should be ok, and will allow
the SDRAM initialization to keep working.
In addition to the real PC value, also print the value of PC after
subtracting the relocation offset. This value will match the address in
the ELF file so it's much easier to figure out where things went wrong.
Stefan Roese [Thu, 26 Aug 2010 15:14:51 +0000 (17:14 +0200)]
ppc4xx: Fix 440EPx bug in reconfigure_pll()
This patch fixes a bug in reconfigure_pll(), where the detection of
the current bootstrap option is wrong. The ICS bits where incorrectly
shifted. This bug was found on the lwmon5 board, which uses bootstrap
option H (I2C bootstrap EEPROM).
Additionally a bit of code was moved into the if statement, since its
only used after later on. No need to run this code all the time.
Also, a few empty lines are added to make the code better readable.
Signed-off-by: Stefan Roese <sr@denx.de> Cc: Rupjyoti Sarmah <rsarmah@amcc.com> Cc: Victor Gallardo <vgallardo@appliedmicro.com>