bad_cache_range actually returned true if the range was OK, but it was used
according to its name, which resulted in all valid dcache range invalidate/flush
operations being dropped. Hence, most DMA transfers resulted in garbage data.
This patch renames this function according to what it does, and it fixes the
interpretation of its return value by other functions. The chosen naming is the
same as for ARM926EJ-S in order to be consistent.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Acked-by: Stefano Babic <sbabic@denx.de>
Use the same name as defined in the datasheet.
DSP_CLK -> DISP_CLK
Signed-off-by: Ashok Kumar Reddy Kourla <ashokkourla2000@gmail.com>
Acked-by: Marek Vasut <marex@denx.de>
The documented link to elftosb package tarball is not accessible,
change to another working link.
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Marek Vasut <marek.vasut@gmail.com>
Cc: Fabio Estevam <festevam@gmail.com>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Acked-by: Detlev Zundel <dzu@denx.de>
Acked-by: Marek Vasut <marex@denx.de>
commit 419ea2d84 (mx28: Split the README into a common part and a m28 specific part)
missed an entry in the Contents section that was part of the original file.
Add the missing entry for completeness.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Create a weak-aliased arch_cpu_init, so that we can get rid of CONFIG_ARCH_CPU_INIT
and always call arch_cpu_init.
This way we do not need to define CONFIG_ARCH_CPU_INIT in every board file, since
arch_cpu_init() is supposed to handle common CPU level code.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
In commit 1a89a217f5 we moved most of the
required pads and mux data for USB to the essential list so that later
on we could NOT enable anything that wasn't essential unless otherwise
configured. This was however missing a few pandaboard-specific parts
which left for example USB ethernet non-functional.
Tested this on OMAP4430 ES2.2, OMAP4460 ES1.1 PANDA boards.
(Reworded by Tom Rini to be more precise about what the problem was)
Signed-off-by: R Sricharan <r.sricharan@ti.com>
Tested-by: Gary Thomas <gary@mlbassoc.com>
Tested-by: Tom Rini <trini@ti.com>
Recent compiler generates unaligned memory access in armv7 default.
But current U-Boot does not allow unaligned memory access, so it causes
data abort exception.
This patch add compile option "-mno-unaligned-access" if it is available.
Signed-off-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
Tested-by: Gary Thomas <gary@mlbassoc.com>
In currently, when __bss_start is equal to __bss_end__,
The bss loop will clear all the things in memory space.
But just only when __bss_end__ greater than __bss_start__,
we do the clear bss section operation.
Signed-off-by: Zhong Hongbo <bocui107@gmail.com>
Acked-by: Andreas Bießmann <andreas.devel@googlemail.com>
* 'master' of git://git.denx.de/u-boot-usb:
usb_storage: fix ehci driver max transfer size
smsc95xx: align buffers to cache line size
ehci-hcd: change debug() to printf() in case of errors
usb: check return value of submit_{control, bulk}_msg
usb: pass cache-aligned buffer to usb_get_descriptor()
ehci-hcd: fix external buffer cache handling
ehci-hcd.c, musb_core, usb.h: Add USB_DMA_MINALIGN define for cache alignment
ehci-hcd: program asynclistaddr before every transfer
common.h: Introduce DEFINE_CACHE_ALIGN_BUFFER
ehci-omap: Do not call dcache_off from omap_ehci_hcd_init
Signed-off-by: Wolfgang Denk <wd@denx.de>
* 'sf' of git://git.denx.de/u-boot-blackfin:
sf: spansion: inline useless id defines
sf: drop unused/duplicate command defines
Signed-off-by: Wolfgang Denk <wd@denx.de>
We have a header file specifically for mingw cruft, so keep it there
to avoid crap spreading into the main tools. This lets our devs just
worry about *nix systems.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
In an effort to unify the spi flash drivers further, drop all the
unused and/or duplicate command defines.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Fix NetSetState function name used with CONFIG_BOOTP_MAY_FAIL.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau@advansee.com>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Now that common code doesn't declare these as volatile, we don't need to
either anymore. This fixes the build warning:
bfin_mac.c: In function 'bfin_EMAC_recv':
bfin_mac.c:193:23: warning: assignment discards qualifiers from pointer target type
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The commit 5dd95cf93d 'usb_storage:
Fix EHCI "out of buffer pointers" with CD-ROM' introduce a bug in
usb_storage as it wrongly assumes that every transfer can use
4096 bytes per qt_buffer. This is wrong if the start address of
the data is not page aligned to 4096 bytes and leads to 'EHCI
timed out on TD' messages because of 'out of buffer pointers'
in ehci_td_buffer function.
The bug appears during load of a fragmented file and
read from or write to an unaligned memory address.
Cc: Marek Vasut <marex@denx.de>
Signed-off-by: Stefan Herbrechtsmeier <stefan@herbrechtsmeier.net>
Return values of submit_{control,bulk}_msg() functions
should be checked to detect possible error.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
usb_get_descriptor passes it's buffer argument directly to
usb_control_msg() so it has to be properly aligned/padded.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
Buffer coming from upper layers should be cacheline aligned/padded
to perform safe cache operations. For now we don't do bounce
buffering so getting unaligned buffer is an upper layer error.
We can't check if the buffer is properly padded with current
interface so just assume it is (consider changing with in the
future). The following changes are done:
1. Remove useless length alignment check. We get actual transfer
length not the size of the underlying buffer so it's perfectly
valid for it to be unaligned.
2. Move flush_dcache_range() out of while loop or it will
flush too much.
3. Don't try to fix buffer address before calling invalidate:
if it's unaligned it's an error anyway so let cache subsystem
cry about that.
4. Fix end buffer address to be cacheline aligned assuming upper
layer reserved enough space. This is potentially dangerous
operation so upper layers should be careful about that.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
The USB spec says that 32 bytes is the minimum required alignment.
However on some platforms we have a larger minimum requirement for cache
coherency. In those cases, use that value rather than the USB spec
minimum. We add a cpp check to <usb.h> to define USB_DMA_MINALIGN and
make use of it in ehci-hcd.c and musb_core.h. We cannot use MAX() here
as we are not allowed to have tests inside of align(...).
Signed-off-by: Tom Rini <trini@ti.com>
[marek.vasut]: introduce some crazy macro voodoo
Signed-off-by: Marek Vasut <marex@denx.de>
[ilya.yanok]: moved external buffer fixes to separate patch,
we use {ALLOC,DEFINE}_ALIGN_BUFFER macros with alignment of USB_DMA_MINALIGN
for qh_list, qh and qtd structures to make sure they are proper aligned
for both controller and cache operations.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
Move or_asynclistaddr programming to ehci_submit_async()
function to make sure queue head is properly programmed
before every transfer. This solves the problem with changing
qh address.
Also remove unneeded qh_list->qh_link reprogramming at the
end of transfer.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
This is the out-of-function-scope counterpart of
ALLOC_CACHE_ALIGN_BUFFER.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Tom Rini <trini@ti.com>
[ilya.yanok]: added missing <linux/compiler.h> include and
{DEFINE,ALLOC}_ALIGN_BUFFER macros allowing explicit alignment
specification.
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
This has never been completely sufficient and now happens too late to
paper over the cache coherency problems with the current USB stack.
Cc: Marek Vasut <marex@denx.de>
Signed-off-by: Tom Rini <trini@ti.com>
Signed-off-by: Ilya Yanok <ilya.yanok@cogentembedded.com>
* 'next' of git://git.denx.de/u-boot-video:
ipu_common: Add ldb_clk for use in parenting the pixel clock
ipu_common: Do not hardcode the ipu_clk frequency
ipu_common: Rename MXC_CCM_BASE
ipu_common: Let clk_ipu_enable/disable only run on MX51 and MX53
ipu_common: Only apply the erratum to MX51
video: Rename CONFIG_VIDEO_MX5
mx6: Allow mx6 to access the IPUv3 registers
common lcd: minor coding style changes
Signed-off-by: Wolfgang Denk <wd@denx.de>
As suggested by Graeme Russ, move gd and bd data structrures
to BSS instead of calculating the locations around the stack
and heap.
Signed-off-by: Thomas Chou <thomas@wytron.com.tw>
Acked-by: Mike Frysinger <vapier@gentoo.org>
For FSL low-end processors (VVN2.2), in order to detect the SD card,
we should enable PEREN, HCKEN and IPGEN to enable the clock.
Otherwise, after booting the u-boot, and then inserting the SD card,
the SD card can't be detected.
For SDHC VVN2.3 IP, these bits are reserved, and SDCLKEN is used.
And when accessing to these reserved bit, no any impact happened.
Signed-off-by: Jerry Huang <Chang-Ming.Huang@freescale.com>
CC: Andy Fleming <afleming@gmail.com>
CC: Marek Vasut <marex@denx.de>
Signed-off-by: Andy Fleming <afleming@freescale.com>
Use the function 'mmc_send_status' to check the card status.
only when the card is ready, driver can send the next erase command
to the card, otherwise, the erase will failed:
=> mmc erase 0 1
MMC erase: dev # 0, block # 0, count 1 ... 1 blocks erase: OK
=> mmc erase 0 2
MMC erase: dev # 0, block # 0, count 2 ... mmc erase failed
1 blocks erase: ERROR
=> mmc erase 0 4
MMC erase: dev # 0, block # 0, count 4 ... mmc erase failed
1 blocks erase: ERROR
Signed-off-by: Jerry Huang <Chang-Ming.Huang@freescale.com>
CC: Andy Fleming <afleming@gmail.com>
CC: Marek Vasut <marex@denx.de>
Signed-off-by: Andy Fleming <afleming@freescale.com>
This code adds call to mmc_init(), for partition related commands (e.g.
fatls, fatinfo etc.).
It is safe to call mmc_init() multiple times since mmc->has_init flag
prevents from multiple initialization.
The FAT related code calls get_dev high level method and then uses
elements from mmc->block_dev, which is uninitialized until the mmc_init
(and thereof mmc_startup) is called.
This problem appears on boards, which don't use mmc as the default
place for envs
Signed-off-by: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Andy Fleming <afleming@gmail.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
mmc_set_clock is set to the hard-coding.
But i think good that use the tran_speed value.
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
* 'master' of git://git.denx.de/u-boot-i2c:
mx28evk: Add I2C support
mxs-i2c: Fix internal address byte order
mxc_i2c: remove setting speed at each start
mx6qsabrelite: add i2c support
mxc_i2c: specify i2c base address in config file
Signed-off-by: Wolfgang Denk <wd@denx.de>
Add I2C support.
Tested by placing a 24LC16 EEPROM into the U50 slot which comes empty from factory.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Large EEPROMs, e.g. 24lc32, need 2 byte to address the internal memory.
These devices require that the high byte of the internal address has to be
written first.
The mxs_i2c driver currently writes the address' low byte first.
The following patch fixes the byte order of the internal address that should
be written to the I2C device.
Signed-off-by: Torsten Fleischer <to-fleischer@t-online.de>
CC: Marek Vasut <marex@denx.de>
CC: Stefano Babic <sbabic@denx.de>
CC: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Marek Vasut <marex@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
Other then being very weird, this code was also wrong.
For example, say I set speed to 100K. I'll read back the speed
as 85937. But the speed is really 85937.5, so we I reset
the speed to 85937, I'll get 73660.7. After a couple of transactions
my speed is now exactly 68750 so it will remain there.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>
Acked-by: Marek Vasut <marex@denx.de>
Acked-by: Stefano Babic <sbabic@denx.de>
The following platforms had their config files changed
flea3, imx31_phycore, mx35pdk, mx53ard, mx53evk, mx53smd
and mx53loco.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>
Acked-by: Stefano Babic <sbabic@denx.de>
The clean up patch missed an &, so we end up passing an int rather than
a pointer to the sprintf function.
arp.c: In function 'ArpReceive':
arp.c:197: warning: format '%p' expects type 'void *', but argument 3 has type 'int'
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
When support sh7734 of sh-ether, ECSIPR_BRCRXIP and other were removed.
Therefore SH7757 and SH7724 can not build. This revise this probelem.
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Tested-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Add ldb_clk for use in parenting the pixel clock.
Signed-off-by: Eric Nelson <eric.nelson@boundarydevices.com>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
The registers accessed inside clk_ipu_enable/disable are not present on MX6,
so make sure they only run on MX51 and MX53.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
The following erratum :
"ENGcm08316
IPU: Clarification regarding the bypass mode registers setup for
display and camera interfaces"
only applies to mx51, so restrict its usage for this SoC only.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
