July 2017 - U-Boot

[U-Boot] [RFC PATCH 0/8] Allwinner A83T PSCI SMP support
by Icenowy Zheng 03 Jul '17

03 Jul '17

This is an experimental support for the PSCI SMP bringup of the first cluster of Allwinner A83T SoC. It's based on some code by Timothy Pearson for A83T (already quite old) and Chen-Yu Tsai (which are originally for A80). In order to ensure the Linux kernel to work properly in SMP environment, some GIC workarounds implemented in [1] is needed. [1] https://github.com/wens/linux/commit/c48654c1f737116e7a7660183c8c74fa919705… Chen-Yu Tsai (2): sunxi: Add CPUCFG register definitions for A80/A83T SoCs sunxi: Add basic PSCI implementation for multi-cluster SoCs Icenowy Zheng (4): sunxi: add configuration of secure SRAM for A83T sunxi: add SUNXI_R_CPUCFG_BASE for A83T SoC sunxi: add multi-cluster CPU PRCM register definition sunxi: enable PSCI for A83T SoC tpearson(a)raptorengineering.com (2): sun8i: Add TZPC setup for A83T sun8i: Add a macro to read the silicon revision arch/arm/cpu/armv7/sunxi/Makefile | 5 + arch/arm/cpu/armv7/sunxi/psci-mcpm.c | 258 +++++++++++++++++++++++++ arch/arm/cpu/armv7/sunxi/psci.c | 2 +- arch/arm/cpu/armv7/sunxi/tzpc.c | 10 + arch/arm/include/asm/arch-sunxi/cpu_sun4i.h | 9 + arch/arm/include/asm/arch-sunxi/cpucfg.h | 64 +----- arch/arm/include/asm/arch-sunxi/cpucfg_sun4i.h | 68 +++++++ arch/arm/include/asm/arch-sunxi/cpucfg_sun9i.h | 60 ++++++ arch/arm/include/asm/arch-sunxi/prcm.h | 8 +- arch/arm/include/asm/arch-sunxi/tzpc.h | 6 + arch/arm/mach-sunxi/Kconfig | 4 + arch/arm/mach-sunxi/board.c | 2 +- arch/arm/mach-sunxi/cpu_info.c | 2 +- include/configs/sun8i.h | 1 + 14 files changed, 434 insertions(+), 65 deletions(-) create mode 100644 arch/arm/cpu/armv7/sunxi/psci-mcpm.c create mode 100644 arch/arm/include/asm/arch-sunxi/cpucfg_sun4i.h create mode 100644 arch/arm/include/asm/arch-sunxi/cpucfg_sun9i.h -- 2.12.2

6 46

[U-Boot] [PATCH 0/1] Set VLD04 output to 2.8V in PMIC initialization.
by Gautam Bhat 03 Jul '17

03 Jul '17

This change sets the VLDO4 settings output to 2.8V in PMIC initialization. I accidentally noticed this when I was testing the DSI voltage input which was 3.3V. Not setting the proper voltage can harm some of the voltage sensitive peripheral. Gautam Bhat (1): Set VLD04 output to 2.8V in PMIC initialization. board/freescale/mx7dsabresd/mx7dsabresd.c | 10 ++++++++++ 1 file changed, 10 insertions(+) -- 2.11.0

4 7

[U-Boot] [PATCH 1/1] bootefi: allow return without EFI_BOOT_SERVICES.Exit
by Heinrich Schuchardt 03 Jul '17

03 Jul '17

The Unified Extensible Firmware Interface Specification, version 2.7, defines in chapter 2.1.2 - UEFI Application that an EFI application may either directly return or call EFI_BOOT_SERVICES.Exit(). Unfortunately U-Boot makes the incorrect assumption that EFI_BOOT_SERVICES.Exit() is always called. So the following application leads to a memory exception on the aarch64 architecture when returning: EFI_STATUS efi_main( EFI_HANDLE handle, EFI_SYSTEM_TABlE systable) { return EFI_SUCCESS; } With this patch the entry point is stored in the image handle. The new wrapper function do_enter is used to call the EFI entry point. Signed-off-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> --- cmd/bootefi.c | 18 ++++++++++++++++-- include/efi_api.h | 4 ++++ 2 files changed, 20 insertions(+), 2 deletions(-) diff --git a/cmd/bootefi.c b/cmd/bootefi.c index a0a5434967..fcb223e999 100644 --- a/cmd/bootefi.c +++ b/cmd/bootefi.c @@ -9,6 +9,7 @@ #include <common.h> #include <command.h> #include <dm.h> +#include <efi_api.h> #include <efi_loader.h> #include <errno.h> #include <libfdt.h> @@ -141,6 +142,18 @@ static void *copy_fdt(void *fdt) return new_fdt; } +static asmlinkage ulong do_enter(void *image_handle, + struct efi_system_table *st) +{ + struct efi_loaded_image *handle = image_handle; + efi_status_t ret = EFI_LOAD_ERROR; + + if (handle->entry) + ret = handle->entry(image_handle, st); + st->boottime->exit(image_handle, ret, 0 , NULL); + return ret; +} + #ifdef CONFIG_ARM64 static unsigned long efi_run_in_el2(ulong (*entry)(void *image_handle, struct efi_system_table *st), void *image_handle, @@ -149,7 +162,7 @@ static unsigned long efi_run_in_el2(ulong (*entry)(void *image_handle, /* Enable caches again */ dcache_enable(); - return entry(image_handle, st); + return do_enter(image_handle, st); } #endif @@ -237,6 +250,7 @@ static unsigned long do_bootefi_exec(void *efi, void *fdt) efi_status_t status = loaded_image_info.exit_status; return status == EFI_SUCCESS ? 0 : -EINVAL; } + loaded_image_info.entry = entry; #ifdef CONFIG_ARM64 /* On AArch64 we need to make sure we call our payload in < EL3 */ @@ -254,7 +268,7 @@ static unsigned long do_bootefi_exec(void *efi, void *fdt) } #endif - return entry(&loaded_image_info, &systab); + return do_enter(&loaded_image_info, &systab); } diff --git a/include/efi_api.h b/include/efi_api.h index 5c3836a51b..611c35c422 100644 --- a/include/efi_api.h +++ b/include/efi_api.h @@ -253,6 +253,10 @@ struct efi_loaded_image { efi_status_t exit_status; struct jmp_buf_data exit_jmp; #endif +#ifdef CONFIG_CMD_BOOTEFI + ulong (*entry)(void *image_handle, struct efi_system_table *st) + asmlinkage; +#endif }; #define DEVICE_PATH_GUID \ -- 2.11.0

2 3

[U-Boot] [PATCH v8 0/8] rk3328: add support of usb host and gadget
by Meng Dongyang 03 Jul '17

03 Jul '17

Add support of usb host and gadget function for rk3328. Changed in v8: - Convert CONFIG_USB_DWC2 to a Kconfig entry Changed in v7: - Romove unused variable ‘prop’ in dwc2 driver Changed in v6: - Rebase on the newest code of master branch - Make xhci-rockchip driver depend on DM_REGULATOR and DM_USB - Remove #ifdef DM_RERULATOR and DM_USB in xhci-rockchip driver - Use dev_read_boot() instead of fdt_getprop() in dwc2 driver Changes in v5: - Propagate return value and print error message when failed Changes in v4: - Splite patch [U-boot,v3,04/10] into two patches - Define set vbus as empty function if the macros aren't set - Prepare a mask and set capabilities in GUSBCFG register once Changes in v3: - Revert change of macro definition in dwc2 driver - Support host mode without HNP/SRP capability through DTS Changes in v2: - Add commit messages - Split patch [U-boot,7/8] into two patches - Use fixed regulator to control vbus instead of gpio Meng Dongyang (8): usb: Kconfig: config USB_XHCI_ROCKCHIP depends on DM_REGULATOR and DM_USB usb: host: xhci-rockchip: use fixed regulator to control vbus rockchip: dts: rk3328: add fixed regulator node for xhci rockchip: configs: rk3328: enable dwc2 driver and config fastboot usb: dwc2: use dev_read_bool() instead of fdt_getprop() rockchip: rk3328: board: add support of dwc2 gadget rockchip: dts: rk3328: support and enable dwc2 rockchip: dts: rk3399: control vbus of typec by fixed regulator arch/arm/dts/rk3328-evb.dts | 25 +++++++++++++++++++- arch/arm/dts/rk3328.dtsi | 10 ++++++++ arch/arm/dts/rk3399-evb.dts | 16 +++++++++++-- board/rockchip/evb_rk3328/evb-rk3328.c | 42 ++++++++++++++++++++++++++++++++++ configs/evb-rk3328_defconfig | 20 ++++++++++++++-- drivers/usb/host/Kconfig | 12 ++++++++++ drivers/usb/host/dwc2.c | 16 +++++-------- drivers/usb/host/xhci-rockchip.c | 36 ++++++++++++++--------------- 8 files changed, 144 insertions(+), 33 deletions(-) -- 1.9.1

4 19

[U-Boot] [PATCH] usb: dwc2: make truncation of ep->dma_buf explicit for setdma_rx/tx
by Philipp Tomsich 03 Jul '17

03 Jul '17

There's an implicit truncation in setdma_rx/tx of the dma_buf address to 32bit that triggers the following warning with recent GCC versions: writel((unsigned int) ep->dma_buf, &reg->out_endp[ep_num].doepdma); ^ drivers/usb/gadget/dwc2_udc_otg_xfer_dma.c: In function 'setdma_rx': drivers/usb/gadget/dwc2_udc_otg_xfer_dma.c:116:9: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast] This change makes the truncation explicit by casting to u32 via uintptr_t. Signed-off-by: Philipp Tomsich <philipp.tomsich(a)theobroma-systems.com> --- drivers/usb/gadget/dwc2_udc_otg_xfer_dma.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/usb/gadget/dwc2_udc_otg_xfer_dma.c b/drivers/usb/gadget/dwc2_udc_otg_xfer_dma.c index 0d6d2fb..de2c2b7 100644 --- a/drivers/usb/gadget/dwc2_udc_otg_xfer_dma.c +++ b/drivers/usb/gadget/dwc2_udc_otg_xfer_dma.c @@ -113,7 +113,7 @@ static int setdma_rx(struct dwc2_ep *ep, struct dwc2_request *req) invalidate_dcache_range((unsigned long) ep->dma_buf, (unsigned long) ep->dma_buf + ep->len); - writel((unsigned int) ep->dma_buf, &reg->out_endp[ep_num].doepdma); + writel((u32)(uintptr_t) ep->dma_buf, &reg->out_endp[ep_num].doepdma); writel(DOEPT_SIZ_PKT_CNT(pktcnt) | DOEPT_SIZ_XFER_SIZE(length), &reg->out_endp[ep_num].doeptsiz); writel(DEPCTL_EPENA|DEPCTL_CNAK|ctrl, &reg->out_endp[ep_num].doepctl); @@ -161,7 +161,7 @@ static int setdma_tx(struct dwc2_ep *ep, struct dwc2_request *req) while (readl(&reg->grstctl) & TX_FIFO_FLUSH) ; - writel((unsigned long) ep->dma_buf, &reg->in_endp[ep_num].diepdma); + writel((u32)(uintptr_t) ep->dma_buf, &reg->in_endp[ep_num].diepdma); writel(DIEPT_SIZ_PKT_CNT(pktcnt) | DIEPT_SIZ_XFER_SIZE(length), &reg->in_endp[ep_num].dieptsiz); -- 2.1.4

1 0

[U-Boot] [PATCH] efi_loader: Add efi-next git tree to MAINTAINERS file
by Alexander Graf 03 Jul '17

03 Jul '17

The efi-next tree lives on github, not the usual denx git. Reflect this in the MAINTAINERS file so that people can find it. Signed-off-by: Alexander Graf <agraf(a)suse.de> --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 957e27c..85dfa14 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -264,6 +264,7 @@ F: test/dm/ EFI PAYLOAD M: Alexander Graf <agraf(a)suse.de> S: Maintained +T: git git://github.com/agraf/u-boot.git F: include/efi_loader.h F: lib/efi_loader/ F: cmd/bootefi.c -- 1.8.5.6

1 0

[U-Boot] [PATCH 1/2] efi_loader: run CreateEvent() notify function based on flags
by Jonathan Gray 03 Jul '17

03 Jul '17

The UEFI specification states that the tpl, function and context arguments are to be ignored if neither EVT_NOTIFY_WAIT or EVT_NOTIFY_SIGNAL are specified. This matches observed behaviour with an AMI EDK2 based UEFI implementation. Skip calling the notify function if neither flag is present. Signed-off-by: Jonathan Gray <jsg(a)jsg.id.au> --- include/efi_api.h | 3 +++ lib/efi_loader/efi_boottime.c | 4 +++- 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/include/efi_api.h b/include/efi_api.h index 5c3836a51b..f071b36b53 100644 --- a/include/efi_api.h +++ b/include/efi_api.h @@ -28,6 +28,9 @@ enum efi_event_type { EFI_TIMER_RELATIVE = 2 }; +#define EVT_NOTIFY_WAIT 0x00000100 +#define EVT_NOTIFY_SIGNAL 0x00000200 + /* EFI Boot Services table */ struct efi_boot_services { struct efi_table_hdr hdr; diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c index 51080cbeed..eb5946a959 100644 --- a/lib/efi_loader/efi_boottime.c +++ b/lib/efi_loader/efi_boottime.c @@ -210,7 +210,9 @@ void efi_timer_check(void) /* Triggering! */ if (efi_event.trigger_type == EFI_TIMER_PERIODIC) efi_event.trigger_next += efi_event.trigger_time / 10; - efi_event.notify_function(&efi_event, efi_event.notify_context); + if (efi_event.type & (EVT_NOTIFY_WAIT | EVT_NOTIFY_SIGNAL)) + efi_event.notify_function(&efi_event, + efi_event.notify_context); } WATCHDOG_RESET(); -- 2.12.0

3 7

[U-Boot] [PATCH] efi_loader: Add check for fallback fdt memory reservation
by Alexander Graf 03 Jul '17

03 Jul '17

When running bootefi, we allocate new space but never check whether the allocation succeeded. This patch adds a check so that in case things go wrong, we at least know they did. Signed-off-by: Alexander Graf <agraf(a)suse.de> --- cmd/bootefi.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/cmd/bootefi.c b/cmd/bootefi.c index a0a5434..771300e 100644 --- a/cmd/bootefi.c +++ b/cmd/bootefi.c @@ -133,7 +133,13 @@ static void *copy_fdt(void *fdt) &new_fdt_addr) != EFI_SUCCESS) { /* If we can't put it there, put it somewhere */ new_fdt_addr = (ulong)memalign(4096, fdt_size); + if (efi_allocate_pages(1, EFI_BOOT_SERVICES_DATA, fdt_pages, + &new_fdt_addr) != EFI_SUCCESS) { + printf("ERROR: Failed to reserve space for FDT\n"); + return NULL; + } } + new_fdt = (void*)(ulong)new_fdt_addr; memcpy(new_fdt, fdt, fdt_totalsize(fdt)); fdt_set_totalsize(new_fdt, fdt_size); -- 1.8.5.6

1 1

[U-Boot] [PATCH] efi_loader: add static to local functions
by Masahiro Yamada 03 Jul '17

03 Jul '17

These are locally used in lib/efi_loader/efi_boottime.c Signed-off-by: Masahiro Yamada <yamada.masahiro(a)socionext.com> --- lib/efi_loader/efi_boottime.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c index 51080cbeed2f..ba4d30540b7b 100644 --- a/lib/efi_loader/efi_boottime.c +++ b/lib/efi_loader/efi_boottime.c @@ -104,9 +104,9 @@ static void EFIAPI efi_restore_tpl(unsigned long old_tpl) EFI_EXIT(efi_unsupported(__func__)); } -efi_status_t EFIAPI efi_allocate_pages_ext(int type, int memory_type, - unsigned long pages, - uint64_t *memory) +static efi_status_t EFIAPI efi_allocate_pages_ext(int type, int memory_type, + unsigned long pages, + uint64_t *memory) { efi_status_t r; @@ -115,7 +115,8 @@ efi_status_t EFIAPI efi_allocate_pages_ext(int type, int memory_type, return EFI_EXIT(r); } -efi_status_t EFIAPI efi_free_pages_ext(uint64_t memory, unsigned long pages) +static efi_status_t EFIAPI efi_free_pages_ext(uint64_t memory, + unsigned long pages) { efi_status_t r; @@ -124,11 +125,12 @@ efi_status_t EFIAPI efi_free_pages_ext(uint64_t memory, unsigned long pages) return EFI_EXIT(r); } -efi_status_t EFIAPI efi_get_memory_map_ext(unsigned long *memory_map_size, - struct efi_mem_desc *memory_map, - unsigned long *map_key, - unsigned long *descriptor_size, - uint32_t *descriptor_version) +static efi_status_t EFIAPI efi_get_memory_map_ext( + unsigned long *memory_map_size, + struct efi_mem_desc *memory_map, + unsigned long *map_key, + unsigned long *descriptor_size, + uint32_t *descriptor_version) { efi_status_t r; -- 2.7.4

3 2

[U-Boot] [PATCH] usb: kbd: don't fail with iomux
by Rob Clark 03 Jul '17

03 Jul '17

stdin might not be set, which would cause iomux_doenv() to fail therefore causing probe_usb_keyboard() to fail. Furthermore if we do have iomux enabled, the sensible thing (in terms of user experience) would be to simply add ourselves to the list of stdin devices. Signed-off-by: Rob Clark <robdclark(a)gmail.com> --- common/usb_kbd.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/common/usb_kbd.c b/common/usb_kbd.c index d2d29cc..72cf78a 100644 --- a/common/usb_kbd.c +++ b/common/usb_kbd.c @@ -517,7 +517,21 @@ static int probe_usb_keyboard(struct usb_device *dev) stdinname = getenv("stdin"); #if CONFIG_IS_ENABLED(CONSOLE_MUX) + char *devname = DEVNAME; + /* stdin might not be set yet.. either way, with console-mux the + * sensible thing to do is add ourselves to the list of stdio + * devices: + */ + if (stdinname) { + char *newstdin = malloc(strlen(stdinname) + strlen(","DEVNAME) + 1); + sprintf(newstdin, "%s,"DEVNAME, stdinname); + stdinname = newstdin; + } else { + stdinname = devname; + } error = iomux_doenv(stdin, stdinname); + if (stdinname != devname) + free(stdinname); if (error) return error; #else -- 2.9.4

1 0