- 04 Oct, 2010 1 commit
-
-
Eric Bénard authored
- remove functions which are no more necessary for tsc2007 - indent platform_data for better readability Signed-off-by:
Eric Bénard <eric@eukrea.com> Signed-off-by:
Sascha Hauer <s.hauer@pengutronix.de>
-
- 01 Oct, 2010 30 commits
-
-
Uwe Kleine-König authored
The first argument to I2C_BOARD_INFO is used to assign .type, so it should not be specified a second time. For the rtc-pcf8563/pcf8563 entry gcc preferred pcf8563, so did I. Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by:
-
Uwe Kleine-König authored
This fixes the following sparse warning: arch/arm/mach-imx/pcm970-baseboard.c:224:13: warning: symbol 'pcm970_baseboard_init' was not declared. Should it be static? Signed-off-by:
-
Uwe Kleine-König authored
Making the lists signed fixes sparse warnings like: arch/arm/mach-imx/mach-mx27_3ds.c:94:31: warning: incorrect type in argument 1 (different signedness) arch/arm/mach-imx/mach-mx27_3ds.c:94:31: expected int const *pin_list arch/arm/mach-imx/mach-mx27_3ds.c:94:31: got unsigned int static [toplevel] *<noident> Signed-off-by:
-
Uwe Kleine-König authored
This fixes the following sparse warning: arch/arm/mach-imx/pcm970-baseboard.c:203:30: warning: symbol 'pcm970_sja1000_platform_data' was not declared. Should it be static? Signed-off-by:
-
Uwe Kleine-König authored
This fixes many sparse warnings like: arch/arm/mach-imx/clock-imx27.c:597:1: warning: Using plain integer as NULL pointer Signed-off-by:
-
Sascha Hauer authored
Currently the uarts and timer only work because they are turned on by reset default. Make them secondary clocks of their corresponding peripheral clocks to make sure they are turned on when necessary. Also, register some clocks to get rid of compiler warnings Signed-off-by:
-
Sascha Hauer authored
Signed-off-by:
-
Uwe Kleine-König authored
Addiontionally make the interrupt #defines match the base address defines MX.._NFC_BASE_ADDR. Signed-off-by:
-
Sascha Hauer authored
Signed-off-by:
-
Sascha Hauer authored
Signed-off-by:
-
Uwe Kleine-König authored
Signed-off-by:
-
Sascha Hauer authored
Signed-off-by:
-
Sascha Hauer authored
Signed-off-by:
-
Sascha Hauer authored
Signed-off-by:
-
Sascha Hauer authored
Signed-off-by:
-
Jason Wang authored
A 2M bytes SPI NOR flash(sst25vf016b) is soldered on the mx51_3ds board. So add the corresponding device for it. Signed-off-by:
Jason Wang <jason77.wang@gmail.com> Acked-by:
Grant Likely <grant.likely@secretlab.ca> Signed-off-by:
-
Jason Wang authored
Signed-off-by:
-
Jason Wang authored
On the imx51_3ds board, eCSPI2 is connected to a SPI NOR flash, now add iomux definitions for those used pins. Signed-off-by:
-
Jason Wang authored
Signed-off-by:
-
Uwe Kleine-König authored
Acked-by:
-
Uwe Kleine-König authored
Acked-by:
-
Uwe Kleine-König authored
Acked-by:
-
Uwe Kleine-König authored
The driver recently learned to handle platform ids. Make use of this new feature. The up side is that the driver needs less knowledge about the spi interfaces used on different SoCs. Acked-by:
-
Uwe Kleine-König authored
i.MX51 comes with two eCSPI interfaces (that are quite different from what was known before---the tried and tested Freescale way) and a CSPI interface that is identical to the devices found on i.MX25 and i.MX35. This patch is a merge of two very similar patches (by Jason Wang and Sascha Hauer resp.) plus a (now hopefully correct) reimplementation of the clock calculation. Acked-by:
-
Uwe Kleine-König authored
This prepares adding support for imx51's eCSPI. This IP has seperate control and config bits for all four supported chip selects, so the config routine needs to know which chip select is being used even if the chipselect is realized by a gpio. Acked-by:
-
Uwe Kleine-König authored
Nearly everything that is needed is provided by the version of the SPI IP. Now the only checks left using cpu_is_... are clk divider tuning on mx21/mx27 and autodetection (which will die soon). Acked-by:
-
Uwe Kleine-König authored
This has the advantage not to need to much cpu_is_... macros. Still more when imx51 support is added which has two different spi interfaces which would introduce additional checks on the device id. With this setup it's not possible for the compiler anymore to detect the unused functions, so four additional kconfig symbols are introduced to ifdef out the unneeded functions in the callback array and all these functions are marked with __maybe_unused to suppress the corresponding gcc warnings. Comparing the driver footprint with and without the patch for a mx27 kernel yields: add/remove: 2/0 grow/shrink: 2/0 up/down: 280/0 (280) function old new delta spi_imx_devtype - 192 +192 spi_imx_probe 980 1032 +52 spi_imx_devtype_data - 32 +32 spi_imx_setupxfer 276 280 +4 Later when the platform code is updated to use the platform ids, the autodetection can be removed which will make the driver a bit smaller again. (~60 Bytes in my test.) Acked-by:
-
Uwe Kleine-König authored
Acked-by:
-
Uwe Kleine-König authored
This is a follow up to c6987159 (ARM: imx: dynamically register imx-i2c devices (imx27)) 2b92084f (ARM: imx: dynamically register imx-i2c devices (imx21)) 6348e6b5 (ARM: imx: dynamically register imx-i2c devices (imx1)) Signed-off-by:
-
Uwe Kleine-König authored
Signed-off-by:
-
- 27 Sep, 2010 9 commits
-
-
Uwe Kleine-König authored
Signed-off-by: -
Uwe Kleine-König authored
Signed-off-by: -
Uwe Kleine-König authored
For mx1_defconfig this yields: add/remove: 1/0 grow/shrink: 1/4 up/down: 49/-108 (-59) function old new delta imx1_imx_uart_data - 48 +48 kernel_config_data 7277 7278 +1 imx_add_imx_uart_1irq 132 128 -4 imx_add_imx_uart_3irq 164 156 -8 scb9328_init 96 64 -32 mx1ads_init 220 156 -64 for mx21_defconfig this yields: add/remove: 1/0 grow/shrink: 0/3 up/down: 64/-52 (12) function old new delta imx21_imx_uart_data - 64 +64 imx_add_imx_uart_3irq 160 156 -4 imx_add_imx_uart_1irq 140 136 -4 mx21ads_board_init 220 176 -44 for a random mx25 config this yields: add/remove: 1/0 grow/shrink: 0/5 up/down: 80/-56 (24) function old new delta imx25_imx_uart_data - 80 +80 imx_add_imx_uart_3irq 160 156 -4 imx_add_imx_uart_1irq 140 136 -4 mx25pdk_init 288 272 -16 eukrea_mbimxsd_baseboard_init 272 256 -16 eukrea_cpuimx25_init 252 236 -16 for mx27_defconfig this yields: add/remove: 1/0 grow/shrink: 0/10 up/down: 96/-280 (-184) function old new delta imx27_imx_uart_data - 96 +96 imx_add_imx_uart_3irq 160 156 -4 imx_add_imx_uart_1irq 140 136 -4 pca100_init 560 544 -16 mx27pdk_init 112 96 -16 mx27lite_init 92 76 -16 eukrea_cpuimx27_init 332 316 -16 pcm038_init 388 348 -40 mxt_td60_board_init 320 280 -40 eukrea_mbimx27_baseboard_init 476 436 -40 mx27ads_board_init 368 280 -88 and finally for mx3_defconfig: add/remove: 2/0 grow/shrink: 0/9 up/down: 128/-344 (-216) function old new delta imx31_imx_uart_data - 80 +80 imx35_imx_uart_data - 48 +48 imx_add_imx_uart_1irq 132 128 -4 imx_add_imx_uart_3irq 164 152 -12 mx31moboard_devboard_init 360 344 -16 mx31lite_db_init 176 160 -16 mx31moboard_smartbot_init 384 360 -24 kzm_board_init 232 208 -24 armadillo5x0_init 392 364 -28 mx31lilly_db_init 248 208 -40 mxc_board_init 3760 3580 -180 Signed-off-by: -
Uwe Kleine-König authored
Group soc specific data in a global struct instead of repeating it for each call to imxXX_add_spi_imxX. The structs holding the actual data are placed in .init.constdata and so don't do much harm. Compared to the previous approach this reduces code size to call imx_add_spi_imx. Acked-by:
-
Uwe Kleine-König authored
Signed-off-by: -
Uwe Kleine-König authored
The names used now match the processor's reference manual. Also remove MXC from the interrupt defines to match the other imx platforms. Acked-by:
Wolfram Sang <w.sang@pengutronix.de> Acked-by:
-
Uwe Kleine-König authored
This makes the header more look like the other ones, i.e. - sort #defines by value - use lowercase hex constants - use a consistently named header guard Acked-by:
-
Uwe Kleine-König authored
Further remove FIXED_PHY as it breaks the ethernet. VGA_CONSOLE isn't selectable anymore since fb78b51c. EXT3_DEFAULTS_TO_ORDERED defaults to y since aa32a796. INOTIFY is gone since 2dfc1cae. CONFIG_DETECT_SOFTLOCKUP is gone since e16bb1d7. Enable TMPFS for udev. KEYS is selected by NFS_USE_KERNEL_DNS Signed-off-by:
-
Uwe Kleine-König authored
This machine is enabled in mx3_defconfig and so mx31pdk_defconfig isn't really useful. Signed-off-by:
-
