June 2016 - U-Boot

[U-Boot] [PATCH] mmc: msm_sdhci: Set mmc->dev pointer in msm_sdc_probe()
by Mateusz Kulikowski 08 Jul '16

08 Jul '16

MMC core expects (now) valid mmc->dev pointer. During conversion in commit cffe5d86 not every driver was updated. This patch fixes crash while accessing MMC on boards using Qualcomm SDHCI controller. Signed-off-by: Mateusz Kulikowski <mateusz.kulikowski(a)gmail.com> --- This patch fixes MMC support on dragonboard. Without it attempts to access MMC caused null-pointer derefernece. drivers/mmc/msm_sdhci.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/mmc/msm_sdhci.c b/drivers/mmc/msm_sdhci.c index 64bbf0c..96dcdbe 100644 --- a/drivers/mmc/msm_sdhci.c +++ b/drivers/mmc/msm_sdhci.c @@ -136,7 +136,12 @@ static int msm_sdc_probe(struct udevice *dev) host->version = sdhci_readw(host, SDHCI_HOST_VERSION); /* automatically detect max and min speed */ - return add_sdhci(host, 0, 0); + ret = add_sdhci(host, 0, 0); + if (ret) + return ret; + host->mmc->dev = dev; + + return 0; } static int msm_sdc_remove(struct udevice *dev) -- 2.7.4

3 3

Re: [U-Boot] Problem loading boot script from ext4 mmc on sunxi H3
by Michal Suchanek 08 Jul '16

08 Jul '16

Hello, On 29 June 2016 at 10:45, Karsten Merker <merker(a)debian.org> wrote: > On Wed, Jun 29, 2016 at 09:30:33AM +0200, Michal Suchanek wrote: > >> I tried loading system with recent u-boot but it does not work. >> >> Trying some ext4ls shows there is some problem accessing the filesystem. >> >> Not sure if this should be attributed to the ext driver or to the mmc >> driver or something else entirely. > > I haven't tested that myself, but I have seen reports on other > platforms that the u-boot ext4 driver has problems with ext4 > filesystems which use the "flex_bg" filesystem feature, as that > causes certain changes to the on-disk layout compared to ext4 > without this feature. From what people have reported, it appears > that "flex_bg" is enabled by default on ext4 filesystems created > with a recent mkfs.ext4 version. > Reformatting without flex_bg does not seem to have any effect. However, ls seems to give some random stuff for directories other than / so some similar remapping feature could cause this. Thanks Michal

2 2

[U-Boot] [PATCH 0/2] dm: mmc: dwmmc: fix the minor things
by Jaehoon Chung 07 Jul '16

07 Jul '16

It's working on converting from legacy to DM. It's also in progessing. There are some minor wrongs. These patches are fixed wrong explanations and removing unnecessary defined. These patchese are based on u-boot-dm. (blk-working branch) Jaehoon Chung (2): dm: mmc: dwmmc: fix the wrong explanation for clock values dm: mmc: dwmmc: use the callback functions as static drivers/mmc/dw_mmc.c | 4 ++-- include/dwmmc.h | 11 ++++------- 2 files changed, 6 insertions(+), 9 deletions(-) -- 1.9.1

3 8

[U-Boot] [PATCH 0/3] ARM: imx: enhance support for the cm-fx6 module
by Christopher Spinrath 07 Jul '16

07 Jul '16

Hi, this series aims at enhancing support for the cm-fx6 module. In particular, with respect to the upstream linux kernel. The first patch improves the default environment. It is non-functional but makes it more convenient to adapt certain settings. The later two patches add mtd partition support for the on-board spi flash chip. They pick up the discussion about specifying a default partitioning in the device tree from here [1]. In short: adding the default partitioning to the device tree was rejected by the linux/ device tree community during mainlining large parts of the device tree. It was proposed to implement the partition/mtd handling in u-boot. On the other hand, it was argued that the flash chip becomes some kind of "black-box" since there will be no partition labeling (in particular, with old u-boot versions). IMHO defining the mtd partitioning has the following (dis-)advantages: Advantages: - It is easier for the user to change the partitioning (e.g. to use the unsued 1MB free space). - The flash ship is used entirely for u-boot. So it is quite natural that u-boot manages it. Also, moving the partition table to it allows us to change the layout in future versions of u-boot (almost independently of the kernel - there are still non-device tree kernels). - U-Boot becomes the single point of definition for all device tree kernels. Otherwise, each kernel (vendor vs. upstream + version) would ship its own partitioning. Moreover, u-boot has to know something about the partitioning, too, because it has to know where the environment is saved. Disadvantages: - Users of the upstream linux kernel have to use a recent u-boot version to avoid the "black box" effect. A concrete impact is that the update routine (described/proposed by CompuLab) does not work out of the box with older u-boot versions. - Updating u-boot is something users might not want or miss to do. However, I think nowadays it is ok to demand a recent u-boot in combination with the upstream kernel. The cm-fx6 wouldn't be the first board doing so. Hence, I propose these patches here. Cheers, Christopher [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-June/434562.html

3 10

[U-Boot] mkimage can't read .tmp
by Tom Van Deun 07 Jul '16

07 Jul '16

Hello all I have been trying to create a FIT image using mkimage using the command: mkimage -D "-I dts -O dtb -p 2000" -f name.its name.fit The output is: DTC: dts->dtb on file "name.its" mkimage: Can't read name.fit.tmp: Invalid argument mkimage Can't add hashes to FIT blob No matter what I try, which arguments I add or remove or which modifications I make to the its file, I keep getting this output. Unfortunately I can't find much on Google for this exact issue. I found someone who solved the issue by installing DTC, but that is not the issue in my case. Uboot development tools, DTC and mkimage are all at the latest version. Does anyone have any idea what other issues might be causing this behavior? Kind regards Tom

2 2

[U-Boot] [PATCH v4 0/2] armv8: Support loading 32-bit OS in AArch32 execution state
by Alison Wang 06 Jul '16

06 Jul '16

This series is to support loading a 32-bit OS, the execution state will change from AArch64 to AArch32 when jumping to kernel. The architecture information will be got through checking FIT image, then U-Boot will load 32-bit OS or 64-bit OS automatically. Spin-table method is used for secondary cores to load 32-bit OS. The architecture information will be got through checking FIT image and saved in the os_arch element of spin-table, then the secondary cores will check os_arch and jump to 32-bit OS or 64-bit OS automatically. --------------------------------------- Changes in v4: - Correct config ARM64_SUPPORT_AARCH32. - Omit arch and ftaddr arguments. - Rename "xreg5" to "tmp". - Use xxx_RES1 to combine all RES1 fields in xxx register. - Use an immediate cmp directly. - Use #ifdef for CONFIG_ARM64_SUPPORT_AARCH32. Changes in v3: - Comments the functions and the arguments. - Rename the real parameters. - Use the macros instead of the magic values. - Remove the redundant codes. - Clean up all of the mess in boot_jump_linux(). - Add CONFIG_ARM64_SUPPORT_AARCH32 to detect for some ARM64 system doesn't support AArch32 state. - Adjust the arguments for armv8_switch_to_el2_m and armv8_switch_to_el1_m. Changes in v2: - armv8_switch_to_el2_aarch32() is removed. armv8_switch_to_el2_m is used to switch to AArch64 EL2 or AArch32 Hyp. - armv8_switch_to_el1_aarch32() is removed. armv8_switch_to_el1_m is used to switch to AArch64 EL1 or AArch32 SVC. - Support to call armv8_switch_to_el2_m and armv8_switch_to_el1_m. Alison Wang (2): armv8: Support loading 32-bit OS in AArch32 execution state armv8: fsl-layerscape: SMP support for loading 32-bit OS arch/arm/Kconfig | 6 ++++ arch/arm/cpu/armv8/fsl-layerscape/lowlevel.S | 35 +++++++++++++++++--- arch/arm/cpu/armv8/fsl-layerscape/mp.c | 10 ++++++ arch/arm/cpu/armv8/start.S | 1 + arch/arm/cpu/armv8/transition.S | 8 ++--- arch/arm/include/asm/arch-fsl-layerscape/mp.h | 6 ++++ arch/arm/include/asm/macro.h | 172 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++----------------------- arch/arm/include/asm/system.h | 111 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-- arch/arm/lib/bootm.c | 24 +++++++++++--- common/image-fit.c | 19 ++++++++++- 10 files changed, 336 insertions(+), 56 deletions(-)

4 13

[U-Boot] [PATCH 00/15] ARM: PSCI: Add secure stack and data sections
by Chen-Yu Tsai 06 Jul '16

06 Jul '16

Hi everyone, This is ARM PSCI improvements part 2. This series cleans up PSCI stack allocation and target PC storage by introducing a secure stack section and a secure data section. The series got larger than I planned. Some patches are just moving or consolidating code, and a few are adding missing macros. It's based on v3 of my "sunxi: PSCI implementation rewrite in C" series from earlier. Patch 1 consolidates the stack setup code from all the PSCI-enabled platforms (which likely originated from the initial sunxi version) into a common function, and moves it into a seperate common function that is called before psci_arch_init. A weak stub psci_arch_init is added for platforms that no longer need a custom version. Patch 2 converts the remaining sunxi PSCI assembly code into C. Patch 3 adds the missing CONFIG_ARMV7_PSCI_NR_CPUS=2 for sun7i/A20. Patch 4 makes the linker page align PSCI text only when it is directly executed, i.e. not copied to CONFIG_ARMV7_SECURE_BASE. Patch 5 adds a fallback value for CONFIG_ARMV7_PSCI_NR_CPUS. The value is 4, which was the fixed maximum number of stacks allocated. Patch 6 adds an empty stack section for the secure monitor. Patch 7 has PSCI allocate its stack in the new secure stack section. Patch 8 removes the now unused psci_text_end symbol. This was previously used to find where to allocate the stack. Patch 9 adds a config variable for specifying the maximum size of the secure section (text, data, stack, etc.) and checks if the binary exceeds it. Patch 10 defines CONFIG_ARMV7_SECURE_MAX_SIZE for sun6i/sun7i (A31/A20), Allwinner SoCs that have a block of secure SRAM. Patch 11 moves the __secure macro to asm/secure.h. Previously sunxi and i.MX7 were defining it themselves. Patch 12 adds a secure data section, and a __secure_data macro. Patch 13 adds helper functions to save and get per-CPU target PC addresses for use in PSCI. Patch 14 converts all PSCI-enabled platforms to the new helper functions to save/get target PC addresses. Patch 15 makes the psci_get_cpu_stack_top function local/static. This function should only be used by the stack setup routine. Regards ChenYu Chen-Yu Tsai (15): ARM: PSCI: Split out common stack setup code from psci_arch_init sunxi: Move remaining PSCI assembly code to C sunxi: Add missing CONFIG_ARMV7_PSCI_NR_CPUS for sun7i ARM: Page align secure section only when it is executed in situ ARM: PSCI: Add fallback value for CONFIG_ARMV7_PSCI_NR_CPUS ARM: Add an empty secure stack section ARM: PSCI: Allocate PSCI stack in secure stack section ARM: PSCI: Remove unused psci_text_end symbol ARM: Add CONFIG_ARMV7_SECURE_MAX_SIZE and check size of secure section sunxi: Define CONFIG_ARMV7_SECURE_MAX_SIZE for sun6i/sun7i ARM: Move __secure definition to common asm/secure.h ARM: Add secure section for initialized data ARM: PSCI: Add helper functions to access per-CPU target PC storage ARM: PSCI: Switch to per-CPU target PC storage in secure data section ARM: PSCI: Make psci_get_cpu_stack_top local to armv7/psci.S arch/arm/cpu/armv7/Makefile | 1 + arch/arm/cpu/armv7/ls102xa/psci.S | 26 ++++---------- arch/arm/cpu/armv7/mx7/psci-mx7.c | 2 +- arch/arm/cpu/armv7/mx7/psci.S | 31 +++++------------ arch/arm/cpu/armv7/nonsec_virt.S | 7 +++- arch/arm/cpu/armv7/psci-common.c | 39 +++++++++++++++++++++ arch/arm/cpu/armv7/psci.S | 55 ++++++++++++++++++++++-------- arch/arm/cpu/armv7/sunxi/Makefile | 1 - arch/arm/cpu/armv7/sunxi/psci.c | 9 +++-- arch/arm/cpu/armv7/sunxi/psci_head.S | 66 ------------------------------------ arch/arm/cpu/u-boot.lds | 60 +++++++++++++++++++++++++------- arch/arm/include/asm/armv7.h | 2 ++ arch/arm/include/asm/config.h | 5 +++ arch/arm/include/asm/psci.h | 9 ++++- arch/arm/include/asm/secure.h | 3 ++ arch/arm/lib/sections.c | 2 ++ arch/arm/mach-tegra/psci.S | 16 ++++----- include/configs/sun6i.h | 1 + include/configs/sun7i.h | 2 ++ 19 files changed, 183 insertions(+), 154 deletions(-) create mode 100644 arch/arm/cpu/armv7/psci-common.c delete mode 100644 arch/arm/cpu/armv7/sunxi/psci_head.S -- 2.8.1

3 32

[U-Boot] [PATCH] RFC: rockchip: video: Lower hpd wait time
by Sjoerd Simons 05 Jul '16

05 Jul '16

Waiting 30 seconds for the hpd to go high seems a bit much, especially on headless boots. Lowering the timeout to 300ms. Sending as RFC because frankly i don't know what a sensible timeout is here, but 30 seconds is clearly not it :) Signed-off-by: Sjoerd Simons <sjoerd.simons(a)collabora.co.uk> --- drivers/video/rockchip/rk_hdmi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/video/rockchip/rk_hdmi.c b/drivers/video/rockchip/rk_hdmi.c index 5fcb61a..8805c77 100644 --- a/drivers/video/rockchip/rk_hdmi.c +++ b/drivers/video/rockchip/rk_hdmi.c @@ -666,7 +666,7 @@ static int hdmi_wait_for_hpd(struct rk3288_hdmi *regs) if (hdmi_get_plug_in_status(regs)) return 0; udelay(100); - } while (get_timer(start) < 30000); + } while (get_timer(start) < 300); return -1; } -- 2.7.0

2 2

[U-Boot] [PATCH] armv8: Enable CPUECTLR.SMPEN for data coherency
by Gong Qianyu 05 Jul '16

05 Jul '16

From: Mingkai Hu <mingkai.hu(a)nxp.com> Data coherency is enabled only when the CPUECTLR.SMPEN bit is set. The SMPEN bit should be set before enabling the data cache. If not enabled, the cache is not coherent with other cores and data corruption could occur. Signed-off-by: Mingkai Hu <mingkai.hu(a)nxp.com> Signed-off-by: Gong Qianyu <Qianyu.Gong(a)nxp.com> diff --git a/arch/arm/cpu/armv8/start.S b/arch/arm/cpu/armv8/start.S index 670e323..735dd67 100644 --- a/arch/arm/cpu/armv8/start.S +++ b/arch/arm/cpu/armv8/start.S @@ -81,6 +81,11 @@ reset: msr cpacr_el1, x0 /* Enable FP/SIMD */ 0: + /* Enalbe SMPEN bit */ + mrs x0, S3_1_c15_c2_1 /* cpuactlr_el1 */ + orr x0, x0, #0x40 + msr S3_1_c15_c2_1, x0 + /* Apply ARM core specific erratas */ bl apply_core_errata -- 2.1.0.27.g96db324

7 8

[U-Boot] [PATCH 0/4] rockchip: rk3288: add fastboot support
by Ziyuan Xu 04 Jul '16

04 Jul '16

Hi, This short series add the fastboot support for rk3288, and I have tested on firefly-rk3288 board. I encounter an odd issue while debugging usb-otg ep2out-bulk data transfer. The buffer was always zero which mapped to DMA buffer, even though driver call invalidate_dcache_range() to make sure the cpu read memroy data directly. IMHO, dcache was abnormal after something initialization code. Everything work fine without CONFIG_EFI_LOADER. Moreover, invalidate_icache_all cause the issue. I'm not able to figure out why and explain it. But I think there is no need to enable EFI applications on rk3288 platform. Ziyuan Xu (4): usb: phy: implement usb-otg phy control for rk3288 usb: dwc2-otg: redefine fifo-size for rk3288 rockchip: rk3288: add fastboot support rockchip: firefly-rk3288: undef CONFIG_EFI_LOADER arch/arm/mach-rockchip/board.c | 20 ++++++++++++++++++++ configs/firefly-rk3288_defconfig | 1 + drivers/usb/gadget/dwc2_udc_otg_regs.h | 6 ++++++ drivers/usb/phy/Makefile | 1 + drivers/usb/phy/rk3288_usb_phy.c | 29 +++++++++++++++++++++++++++++ include/configs/rk3288_common.h | 25 +++++++++++++++++++++++++ 6 files changed, 82 insertions(+) create mode 100644 drivers/usb/phy/rk3288_usb_phy.c -- 1.9.1

4 15