Re: [U-Boot] [PATCH 00/14] Support for HTKW mcx board
Hi Tom, On 11.11.2011 06:52, Tom Rini wrote: Added no-op cache functions for arm926ejs. However many DaVinci boards are still broken as suprisingly they have D-Cache enabled (I wonder if EMAC works on these boards). Given how quick builds are now, what's MAKEALL -s davinci look like, before/after? Or did you mean EMAC functions not just builds? That's MAKEALL logs are attached. Things are significantly worse after merging the series. That's because of DaVinci boards using EMAC driver and D-Cache enabled (I think this comes from times when D-Cache was disabled by default on ARM9). With this http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/116671 patch applied the results are almost the same as with master branch (see after+patch.log). another question which I've had a few conversations with people about (and I've got a TODO list item to poke the kernel about how it does cache flushing/etc here, but it also probably relies on page stuff we don't do..). Thanks! Well, kernel uses coherent memory for descriptors and do dma_map/dma_unmap on buffers. We don't have coherent memory support in U-Boot so I just perform flush/invalidate for both descriptors and buffers. Regards, Ilya. Configuring for da830evm board... cache.c:27:2: warning: #warning Please provide some real implementation for cache operations da830evm.c: In function 'board_init': da830evm.c:222:2: error: 'DAVINCI_SYSCFG_SUSPSRC_UART2' undeclared (first use in this function) da830evm.c:222:2: note: each undeclared identifier is reported only once for each function it appears in make[1]: *** [da830evm.o] Error 1 make: *** [board/davinci/da8xxevm/libda8xxevm.o] Error 2 Configuring for da850evm board... da850evm.c: In function 'board_init': da850evm.c:328:2: error: 'DAVINCI_SYSCFG_SUSPSRC_UART2' undeclared (first use in this function) da850evm.c:328:2: note: each undeclared identifier is reported only once for each function it appears in make[1]: *** [da850evm.o] Error 1 make: *** [board/davinci/da8xxevm/libda8xxevm.o] Error 2 Configuring for cam_enc_4xx board... text data bss dec hex filename 321180 7476 298416 627072 99180 ./u-boot Configuring for hawkboard board... cache.c:27:2: warning: #warning Please provide some real implementation for cache operations drivers/net/libnet.o: In function `davinci_invalidate_rx_descs': /work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:132: undefined reference to `invalidate_dcache_range' drivers/net/libnet.o: In function `davinci_eth_rcv_packet': /work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:707: undefined reference to `invalidate_dcache_range' drivers/net/libnet.o: In function `davinci_flush_desc': /work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:138: undefined reference to `flush_dcache_range' /work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:138: undefined reference to `flush_dcache_range' /work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:138: undefined reference to `flush_dcache_range' drivers/net/libnet.o: In function `davinci_eth_send_packet': /work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:660: undefined reference to `flush_dcache_range' drivers/net/libnet.o: In function `davinci_flush_desc': /work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:138: undefined reference to `flush_dcache_range' drivers/net/libnet.o:/work/yanok/am3517/u-boot/drivers/net/davinci_emac.c:125: more undefined references to `flush_dcache_range' follow arm-linux-gnueabi-ld: BFD (GNU Binutils) 2.21 assertion fail /opt/poky/build/eldk-2011-05-20-5cde06e-armv5te/tmp/work/i686-nativesdk-oesdk-linux/binutils-cross-canadian-arm-2.21-r0/binutils-2.21/bfd/elf32-arm.c:12235 arm-linux-gnueabi-ld: BFD (GNU Binutils) 2.21 assertion fail /opt/poky/build/eldk-2011-05-20-5cde06e-armv5te/tmp/work/i686-nativesdk-oesdk-linux/binutils-cross-canadian-arm-2.21-r0/binutils-2.21/bfd/elf32-arm.c:12235 arm-linux-gnueabi-ld: BFD (GNU Binutils) 2.21 assertion fail /opt/poky/build/eldk-2011-05-20-5cde06e-armv5te/tmp/work/i686-nativesdk-oesdk-linux/binutils-cross-canadian-arm-2.21-r0/binutils-2.21/bfd/elf32-arm.c:12472 /bin/sh: line 1: 6140 Segmentation fault (core dumped) arm-linux-gnueabi-ld --sysroot=/opt/eldk-5.0/armv5te/sysroots/arm-linux-gnueabi -pie -T u-boot.lds -Bstatic -Ttext 0xc118 $UNDEF_SYM arch/arm/cpu/arm926ejs/start.o --start-group api/libapi.o arch/arm/cpu/arm926ejs/davinci/libdavinci.o arch/arm/cpu/arm926ejs/libarm926ejs.o arch/arm/lib/libarm.o board/davinci/common/libdavinci.o common/libcommon.o disk/libdisk.o drivers/bios_emulator/libatibiosemu.o drivers/block/libblock.o drivers/dma/libdma.o drivers/fpga/libfpga.o drivers/gpio/libgpio.o drivers/hwmon/libhwmon.o drivers/i2c/libi2c.o drivers/input/libinput.o drivers/misc/libmisc.o drivers/mmc/libmmc.o drivers/mtd/libmtd.o drivers/mtd/nand/libnand.o drivers/mtd/onenand/libonenand.o drivers/mtd/spi/libspi_flash.o drivers/mtd/ubi/libubi.o drivers/net/libnet.o drivers/net/phy/libphy.o drivers/pci/libpci.o
Re: [U-Boot] [PATCH 00/14] Support for HTKW mcx board
Hi Tom, I've just posted the updated patches, please consider merging them. ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/116535 ) On 07.11.2011 04:14, Tom Rini wrote: I'm not sure what we should do about this series right now. Aside from some easily fixable things like all of the PHY files needing an ugly #include ../../../../../drivers/net/davinci_emac.h to compile I've fixed PHY files to include this. again and a CONFIG_SYS_CACHELINE_SIZE 32 in all of the platforms, the s/CONFIG_SYS_CACHELINE_SIZE/ARCH_DMA_MINALIGN/ should solve this problem. I'll fix this. Thanks. Fixed. lack of flush functions means that MAKEALL -s davinci fails for almost every board now. I'm not sure what we should do, at this point... Argh.. Actually my initial version had a workaround for this, only using {flush,invalidate}_dcache_range under #ifdef DAVINCI_EMAC_DCACHE clause... But Mike told me that I should remove this... Maybe we can just add empty functions for CONFIG_DCACHE_OFF case to fix the compilation? Added no-op cache functions for arm926ejs. However many DaVinci boards are still broken as suprisingly they have D-Cache enabled (I wonder if EMAC works on these boards). Regards, Ilya. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 00/14] Support for HTKW mcx board
On Thu, Nov 10, 2011 at 5:49 PM, Ilya Yanok ya...@emcraft.com wrote: Hi Tom, I've just posted the updated patches, please consider merging them. ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/116535 ) On 07.11.2011 04:14, Tom Rini wrote: I'm not sure what we should do about this series right now. Aside from some easily fixable things like all of the PHY files needing an ugly #include ../../../../../drivers/net/davinci_emac.h to compile I've fixed PHY files to include this. again and a CONFIG_SYS_CACHELINE_SIZE 32 in all of the platforms, the s/CONFIG_SYS_CACHELINE_SIZE/ARCH_DMA_MINALIGN/ should solve this problem. I'll fix this. Thanks. Fixed. lack of flush functions means that MAKEALL -s davinci fails for almost every board now. I'm not sure what we should do, at this point... Argh.. Actually my initial version had a workaround for this, only using {flush,invalidate}_dcache_range under #ifdef DAVINCI_EMAC_DCACHE clause... But Mike told me that I should remove this... Maybe we can just add empty functions for CONFIG_DCACHE_OFF case to fix the compilation? Added no-op cache functions for arm926ejs. However many DaVinci boards are still broken as suprisingly they have D-Cache enabled (I wonder if EMAC works on these boards). Given how quick builds are now, what's MAKEALL -s davinci look like, before/after? Or did you mean EMAC functions not just builds? That's another question which I've had a few conversations with people about (and I've got a TODO list item to poke the kernel about how it does cache flushing/etc here, but it also probably relies on page stuff we don't do..). Thanks! -- Tom ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 00/14] Support for HTKW mcx board
Hi Tom, 03.11.2011 4:33, Tom Rini wrote: these patches introduce support for HTKW mcx board (AM3517-based). Previously posted DaVinci EMAC patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112309 ) and NAND SPL patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112890 ) are also included as long as EHCI host support (moved from Beagle board code). Signed-off-by: Ilya Yanokya...@emcraft.com I'm not sure what we should do about this series right now. Aside from some easily fixable things like all of the PHY files needing an ugly #include ../../../../../drivers/net/davinci_emac.h to compile Shouldn't we use generic PHY layer and move all PHY files to drivers/net/phy? again and a CONFIG_SYS_CACHELINE_SIZE 32 in all of the platforms, the s/CONFIG_SYS_CACHELINE_SIZE/ARCH_DMA_MINALIGN/ should solve this problem. I'll fix this. lack of flush functions means that MAKEALL -s davinci fails for almost every board now. I'm not sure what we should do, at this point... Argh.. Actually my initial version had a workaround for this, only using {flush,invalidate}_dcache_range under #ifdef DAVINCI_EMAC_DCACHE clause... But Mike told me that I should remove this... Maybe we can just add empty functions for CONFIG_DCACHE_OFF case to fix the compilation? Regards, Ilya. ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 00/14] Support for HTKW mcx board
On Sunday, November 6, 2011, Ilya Yanok ya...@emcraft.com wrote: Hi Tom, 03.11.2011 4:33, Tom Rini wrote: these patches introduce support for HTKW mcx board (AM3517-based). Previously posted DaVinci EMAC patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112309 ) and NAND SPL patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112890 ) are also included as long as EHCI host support (moved from Beagle board code). Signed-off-by: Ilya Yanokya...@emcraft.com I'm not sure what we should do about this series right now. Aside from some easily fixable things like all of the PHY files needing an ugly #include ../../../../../drivers/net/davinci_emac.h to compile Shouldn't we use generic PHY layer and move all PHY files to drivers/net/phy? I'd love to see a switch to phylib (which the kernel version uses) but don't see that as a blocker to merging the series. again and a CONFIG_SYS_CACHELINE_SIZE 32 in all of the platforms, the s/CONFIG_SYS_CACHELINE_SIZE/ARCH_DMA_MINALIGN/ should solve this problem. I'll fix this. Thanks. lack of flush functions means that MAKEALL -s davinci fails for almost every board now. I'm not sure what we should do, at this point... Argh.. Actually my initial version had a workaround for this, only using {flush,invalidate}_dcache_range under #ifdef DAVINCI_EMAC_DCACHE clause... But Mike told me that I should remove this... Maybe we can just add empty functions for CONFIG_DCACHE_OFF case to fix the compilation? Let me think about this more... -- Tom -- Tom ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 00/14] Support for HTKW mcx board
On Tue, Nov 1, 2011 at 4:15 PM, Ilya Yanok ya...@emcraft.com wrote: Hello All, these patches introduce support for HTKW mcx board (AM3517-based). Previously posted DaVinci EMAC patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112309 ) and NAND SPL patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112890 ) are also included as long as EHCI host support (moved from Beagle board code). Signed-off-by: Ilya Yanok ya...@emcraft.com I'm not sure what we should do about this series right now. Aside from some easily fixable things like all of the PHY files needing an ugly #include ../../../../../drivers/net/davinci_emac.h to compile again and a CONFIG_SYS_CACHELINE_SIZE 32 in all of the platforms, the lack of flush functions means that MAKEALL -s davinci fails for almost every board now. I'm not sure what we should do, at this point... -- Tom ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
Re: [U-Boot] [PATCH 00/14] Support for HTKW mcx board
On Tue, Nov 1, 2011 at 4:15 PM, Ilya Yanok ya...@emcraft.com wrote: Hello All, these patches introduce support for HTKW mcx board (AM3517-based). Previously posted DaVinci EMAC patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112309 ) and NAND SPL patches ( http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/112890 ) are also included as long as EHCI host support (moved from Beagle board code). Signed-off-by: Ilya Yanok ya...@emcraft.com As a note, these patches, along with the board specific enablement work on am3517evm and I'll be posting those patches soon. -- Tom ___ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot