November 2018 - U-Boot

[U-Boot] Good news on migration
by Simon Glass 25 Nov '18

25 Nov '18

Hi, I just wanted to mention that the CONFIG whitelist file started out two years ago at 8420 lines and is now only 4690 line, a reduction of close to half, in about 550 commits. So the Kconfig conversion has gone quite well. Good news! Regards, Simon

1 0

[U-Boot] [PATCH v15 0/4] efi_loader: Code refactoring and improvement
by Simon Glass 25 Nov '18

25 Nov '18

This collects the patches previously sent to break up the very large functions in efi_loader into smaller pieces. Now that the other sandbox stuff is applied, perhaps it is time to apply these patches. This also adds a few new patches to fix more recent breakages. Unfortunately we still cannot enable the efi loader tests since one of the tests fails. Thus we should expect additional failures to appear until that is resolved. Changes in v15: - Add a comment about a leaked device path - Add check for return values to bootefi_test_prepare() - Drop call to efi_save_gd() in bootefi_test_prepare() - Fix minor checkpatch nit with bracket Changes in v14: - Fix condition for invalid pointer - Go back to the horrible long variable names - Hopefully correct error paths in do_bootefi_exec() Changes in v13: - Drop 'efi_loader: Drop setup_ok' as we have an existing patch for that - Drop patches previously applied - Rebase to efi/efi-next - Update code to deal with the struct_table_address member Changes in v12: - Rename image to image_prot Changes in v11: - Drop patches previously applied - Fix the EFI code that has since been added and relies on broken behaviour Changes in v9: - Add comments to bootefi_test_prepare() about the memset()s Changes in v7: - Drop patch "efi: Init the 'rows' and 'cols' variables" - Drop patches previous applied Changes in v5: - Drop call to efi_init_obj_list() which is now done in do_bootefi() - Introduce load_options_path to specifyc U-Boot env var for load_options_path - Rebase to master Changes in v4: - Rebase to master Changes in v3: - Add new patch to rename bootefi_test_finish() to bootefi_run_finish() - Add new patch to split out test init/uninit into functions - Add patch to create a function to set up for running EFI code - Drop incorrect map_sysmem() in write_smbios_table() Simon Glass (4): sandbox: smbios: Update to support sandbox efi: Split out test init/uninit into functions efi: Create a function to set up for running EFI code efi: Rename bootefi_test_finish() to bootefi_run_finish() cmd/bootefi.c | 120 +++++++++++++++++++++++++++--------- lib/efi_loader/efi_smbios.c | 20 +++--- lib/smbios.c | 52 +++++++++++++--- 3 files changed, 147 insertions(+), 45 deletions(-) -- 2.19.1.1215.g8438c0b245-goog

3 12

[U-Boot] [PATCH 1/3] Makefile: Add a warning for boards that don't use CONFIG_BLK
by Simon Glass 24 Nov '18

24 Nov '18

The deadline for this has passed and we are starting to remove boards that do not use driver model for block devices. Add a noisy Makefile warning. Signed-off-by: Simon Glass <sjg(a)chromium.org> --- Makefile | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/Makefile b/Makefile index 552687db538..8bf422f6a85 100644 --- a/Makefile +++ b/Makefile @@ -917,6 +917,13 @@ ifeq ($(CONFIG_DM_I2C_COMPAT)$(CONFIG_SANDBOX),y) @echo "(possibly in a subsequent patch in your series)" @echo "before sending patches to the mailing list." @echo "====================================================" +endif +ifneq ($(CONFIG_BLK),y) + @echo "===================== WARNING ======================" + @echo "This board does not use CONFIG_BLK. Please update" + @echo "the board to use CONFIG_BLK before the end of 2018." + @echo "See doc/driver-model/MIGRATION.txt for more info." + @echo "====================================================" endif @# Check that this build does not use CONFIG options that we do not @# know about unless they are in Kconfig. All the existing CONFIG -- 2.19.1.1215.g8438c0b245-goog

3 4

Re: [U-Boot] [linux-sunxi] Re: [PATCH v3 0/9] Stop AXP from crashing when enabling LDO3
by Priit Laes 24 Nov '18

24 Nov '18

On Wed, Nov 21, 2018 at 03:58:18PM -0500, Stefan Monnier wrote: > > When powering up an AXP209, the default value for LDO3 output is > > enabled. This works fine. However if for whatever reason, LDO3 is > > disabled, for example by OS during reboot and u-boot enables LDO3 > > again, the PMIC shuts down (without setting an interrupt) causing the > > board to hang. This behavior has been seen from Linux as well, u-boot > > disables LDO3 as a default value, the kernel enables it per its DTS, > > the kernel hangs as the PMIC gets shut down. > > Hmm... I've never had any problems booting my Orange Pi mini, but it > (almost) always hangs on reboot (right after the final message from the > OS and before the first message from U-boot). > Any chance this is the same problem? Well, f you have the hardware, then there's only one way to find out: 1) Apply patchset 2) Add following into boards defconfig: CONFIG_AXP_ALDO3_INRUSH_QUIRK=y 3) If it still fails to reboot, add following CONFIG_AXP_ALDO3_VOLT_SLOPE_08=y > > > Stefan > > -- > You received this message because you are subscribed to the Google Groups "linux-sunxi" group. > To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscribe(a)googlegroups.com. > For more options, visit https://groups.google.com/d/optout.

2 1

[U-Boot] [PATCH] arm: socfpga: crc-protect SPL on warm boot
by Simon Goldschmidt 24 Nov '18

24 Nov '18

From: Simon Goldschmidt <sgoldschmidt(a)de.pepperl-fuchs.com> On socfpga gen5, a warm reboot from Linux currently triggers a warm reset via reset manager ctrl register. This currently leads to the boot rom just jumping to onchip ram executing the SPL that is supposed to still be there. This is because we tell the boot rom to do so by writing a magin value the warmramgrp_enable register in arch_early_init_r(). However, this can lead to lockups on reboot if this register still contains its magic value but the SPL is not intact any more (e.g. partly overwritten onchip ram). To fis this, store a crc calculated over SPL in sysmgr registers to let the boot rom check it on next warm boot. If the crc is still correct, SPL can be executd from onchip ram. If the crc fails, SPL is loaded from original boot source. The crc that is written to the warmramgrp_crc register is the crc found in the SPL sfp image but with one addiional u32 added. For this, we need to add a function to calculate the updated crc. This is done as a bitwise calculation to keep the code increase small. This whole patch added 96 bytes to .text for SPL for socfpga_socrates_defconfig. Signed-off-by: Simon Goldschmidt <simon.k.r.goldschmidt(a)gmail.com> --- arch/arm/mach-socfpga/misc_gen5.c | 9 ---- arch/arm/mach-socfpga/spl_gen5.c | 73 +++++++++++++++++++++++++++++++ 2 files changed, 73 insertions(+), 9 deletions(-) diff --git a/arch/arm/mach-socfpga/misc_gen5.c b/arch/arm/mach-socfpga/misc_gen5.c index 5fa40937c4..492a3082de 100644 --- a/arch/arm/mach-socfpga/misc_gen5.c +++ b/arch/arm/mach-socfpga/misc_gen5.c @@ -204,15 +204,6 @@ int arch_early_init_r(void) { int i; - /* - * Write magic value into magic register to unlock support for - * issuing warm reset. The ancient kernel code expects this - * value to be written into the register by the bootloader, so - * to support that old code, we write it here instead of in the - * reset_cpu() function just before resetting the CPU. - */ - writel(0xae9efebc, &sysmgr_regs->romcodegrp_warmramgrp_enable); - for (i = 0; i < 8; i++) /* Cache initial SW setting regs */ iswgrp_handoff[i] = readl(&sysmgr_regs->iswgrp_handoff[i]); diff --git a/arch/arm/mach-socfpga/spl_gen5.c b/arch/arm/mach-socfpga/spl_gen5.c index ccdc661d05..3416e19f79 100644 --- a/arch/arm/mach-socfpga/spl_gen5.c +++ b/arch/arm/mach-socfpga/spl_gen5.c @@ -63,6 +63,76 @@ u32 spl_boot_mode(const u32 boot_device) } #endif +/* This function calculates the CRC32 used by the Cyclone 5 SoC Boot Rom */ +static u32 socfpga_boot_crc(u32 crc, const unsigned char *ptr, u32 length) +{ + uint i; + u8 bit; + unsigned char data; + const u32 poly = 0x02608edb; + + for (; length > 0; length--, ptr++) { + data = *ptr; + for (i = 0; i < 8; i++) { + if (data & 0x80) + bit = 1; + else + bit = 0; + + data = data << 1; + if (crc & 0x80000000) + bit = 1 - bit; + + if (bit) { + crc ^= poly; + crc = crc << 1; + crc |= 1; + } else { + crc = crc << 1; + } + } + } + return crc; +} + +/* + * Write magic value into magic register to unlock support for the boot rom to + * execute spl from sram on warm reset. This may be required at least on some + * boards that start from qspi where the flash chip might be in a state that + * cannot be handled by the boot rom (e.g. 4 byte mode). + * + * To prevent just jumping to corrupted memory, a crc of the spl is calculated. + * This crc is loaded from the running image, but has to be extended by the + * modified contents of the "datastart" register (i.e. 0xffff0000). + */ +static void spl_init_reboot_config(void) +{ + u32 spl_crc, spl_length; + const u32 spl_start = (u32)__image_copy_start; + const u32 spl_start_16 = spl_start & 0xffff; + u32 spl_length_u32; + + /* load image length from sfp header (includes crc) */ + spl_length_u32 = *(const u16 *)(spl_start + 0x46); + /* subtract crc */ + spl_length_u32--; + /* get length in bytes */ + spl_length = spl_length_u32 * 4; + /* load crc */ + spl_crc = *(const u32 *)(spl_start + spl_length); + /* undo xor */ + spl_crc ^= 0xffffffff; + /* add contents of modified datastart register */ + spl_crc = socfpga_boot_crc(spl_crc, (const u8 *)&spl_start, 4); + /* finalize */ + spl_crc ^= 0xffffffff; + + writel(0xae9efebc, &sysmgr_regs->romcodegrp_warmramgrp_enable); + writel(spl_start_16, &sysmgr_regs->romcodegrp_warmramgrp_datastart); + writel(spl_length, &sysmgr_regs->romcodegrp_warmramgrp_length); + writel(spl_crc, &sysmgr_regs->romcodegrp_warmramgrp_crc); +} + void board_init_f(ulong dummy) { const struct cm_config *cm_default_cfg = cm_get_default_config(); @@ -82,6 +152,9 @@ void board_init_f(ulong dummy) writel(SYSMGR_ECC_OCRAM_DERR | SYSMGR_ECC_OCRAM_EN, &sysmgr_regs->eccgrp_ocram); + if (!socfpga_is_booting_from_fpga()) + spl_init_reboot_config(); + memset(__bss_start, 0, __bss_end - __bss_start); socfpga_sdram_remap_zero(); -- 2.17.1

3 14

[U-Boot] [PATCH] RFC: checkpatch: Add a check for tests needed for uclasses
by Simon Glass 24 Nov '18

24 Nov '18

This is an experimental check for adding new uclasses without a test. I am not sure of the best way to add U-Boot-specific tests, although in this case, it would not fire on Linux. Signed-off-by: Simon Glass <sjg(a)chromium.org> --- scripts/checkpatch.pl | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl index 373094e59ef..2f9edb429d5 100755 --- a/scripts/checkpatch.pl +++ b/scripts/checkpatch.pl @@ -3033,6 +3033,13 @@ sub process { "adding a line without newline at end of file\n" . $herecurr); } + # ask for a test if a new uclass ID is added + if ($realfile =~ /uclass-id.h/ && $line =~ /^\+/) { + WARN("NEW_UCLASS", + "Possible new uclass - make sure to add a test in test/dm/<name>.c\n" . $herecurr); + } + + # check we are in a valid source file C or perl if not then ignore this hunk next if ($realfile !~ /\.(h|c|pl|dtsi|dts)$/); -- 2.19.1.1215.g8438c0b245-goog

2 4

[U-Boot] [PATCH v14 0/4] efi_loader: Code refactoring and improvement
by Simon Glass 24 Nov '18

24 Nov '18

This collects the patches previously sent to break up the very large functions in efi_loader into smaller pieces. Now that the other sandbox stuff is applied, perhaps it is time to apply these patches. This also adds a few new patches to fix more recent breakages. Unfortunately we still cannot enable the efi loader tests since one of the tests fails. Thus we should expect additional failures to appear until that is resolved. Changes in v14: - Fix condition for invalid pointer - Go back to the horrible long variable names - Hopefully correct error paths in do_bootefi_exec() Changes in v13: - Drop 'efi_loader: Drop setup_ok' as we have an existing patch for that - Drop patches previously applied - Rebase to efi/efi-next - Update code to deal with the struct_table_address member Changes in v12: - Rename image to image_prot Changes in v11: - Drop patches previously applied - Fix the EFI code that has since been added and relies on broken behaviour Changes in v9: - Add comments to bootefi_test_prepare() about the memset()s Changes in v7: - Drop patch "efi: Init the 'rows' and 'cols' variables" - Drop patches previous applied Changes in v5: - Drop call to efi_init_obj_list() which is now done in do_bootefi() - Introduce load_options_path to specifyc U-Boot env var for load_options_path - Rebase to master Changes in v4: - Rebase to master Changes in v3: - Add new patch to rename bootefi_test_finish() to bootefi_run_finish() - Add new patch to split out test init/uninit into functions - Add patch to create a function to set up for running EFI code - Drop incorrect map_sysmem() in write_smbios_table() Simon Glass (4): sandbox: smbios: Update to support sandbox efi: Split out test init/uninit into functions efi: Create a function to set up for running EFI code efi: Rename bootefi_test_finish() to bootefi_run_finish() cmd/bootefi.c | 120 +++++++++++++++++++++++++++--------- lib/efi_loader/efi_smbios.c | 20 +++--- lib/smbios.c | 52 +++++++++++++--- 3 files changed, 147 insertions(+), 45 deletions(-) -- 2.19.1.1215.g8438c0b245-goog

2 14

[U-Boot] [PATCH v2] net: e1000: support 'write_hwaddr' in DM
by Hannes Schmelzer 24 Nov '18

24 Nov '18

This commit ports the existing (non-DM) function for writing the MAC- address into the shadow ram (and flash) for DM. Signed-off-by: Hannes Schmelzer <hannes.schmelzer(a)br-automation.com> --- Changes in v2: - fix build for non-DM board - rebase on current master drivers/net/e1000.c | 89 +++++++++++++++++++++++++++++------------------------ 1 file changed, 48 insertions(+), 41 deletions(-) diff --git a/drivers/net/e1000.c b/drivers/net/e1000.c index a34f697..3df9999 100644 --- a/drivers/net/e1000.c +++ b/drivers/net/e1000.c @@ -862,7 +862,6 @@ e1000_read_eeprom(struct e1000_hw *hw, uint16_t offset, return E1000_SUCCESS; } -#ifndef CONFIG_DM_ETH /****************************************************************************** * e1000_write_eeprom_srwr - Write to Shadow Ram using EEWR * @hw: pointer to the HW structure @@ -1028,7 +1027,6 @@ static int32_t e1000_update_eeprom_checksum_i210(struct e1000_hw *hw) out: return ret_val; } -#endif /****************************************************************************** * Verifies that the EEPROM has a valid checksum @@ -5488,6 +5486,53 @@ void e1000_get_bus_type(struct e1000_hw *hw) } #ifndef CONFIG_DM_ETH +static int e1000_write_hwaddr(struct eth_device *dev) +#else +static int e1000_write_hwaddr(struct udevice *dev) +#endif +{ +#ifndef CONFIG_E1000_NO_NVM +#ifndef CONFIG_DM_ETH + unsigned char *mac = dev->enetaddr; + struct e1000_hw *hw = dev->priv; +#else + struct eth_pdata *plat = dev_get_platdata(dev); + unsigned char *mac = plat->enetaddr; + struct e1000_hw *hw = dev_get_priv(dev); +#endif + unsigned char current_mac[6]; + u16 data[3]; + int ret_val, i; + + DEBUGOUT("%s: mac=%pM\n", __func__, mac); + + memset(current_mac, 0, 6); + + /* Read from EEPROM, not from registers, to make sure + * the address is persistently configured + */ + ret_val = e1000_read_mac_addr_from_eeprom(hw, current_mac); + DEBUGOUT("%s: current mac=%pM\n", __func__, current_mac); + + /* Only write to EEPROM if the given address is different or + * reading the current address failed + */ + if (!ret_val && memcmp(current_mac, mac, 6) == 0) + return 0; + + for (i = 0; i < 3; ++i) + data[i] = mac[i * 2 + 1] << 8 | mac[i * 2]; + + ret_val = e1000_write_eeprom_srwr(hw, 0x0, 3, data); + if (!ret_val) + ret_val = e1000_update_eeprom_checksum_i210(hw); + return ret_val; +#else /* CONFIG_E1000_NO_NVM */ + return 0; +#endif +} + +#ifndef CONFIG_DM_ETH /* A list of all registered e1000 devices */ static LIST_HEAD(e1000_hw_list); #endif @@ -5649,45 +5694,6 @@ e1000_poll(struct eth_device *nic) return len ? 1 : 0; } -static int e1000_write_hwaddr(struct eth_device *dev) -{ -#ifndef CONFIG_E1000_NO_NVM - unsigned char *mac = dev->enetaddr; - unsigned char current_mac[6]; - struct e1000_hw *hw = dev->priv; - uint16_t data[3]; - int ret_val, i; - - DEBUGOUT("%s: mac=%pM\n", __func__, mac); - - memset(current_mac, 0, 6); - - /* Read from EEPROM, not from registers, to make sure - * the address is persistently configured - */ - ret_val = e1000_read_mac_addr_from_eeprom(hw, current_mac); - DEBUGOUT("%s: current mac=%pM\n", __func__, current_mac); - - /* Only write to EEPROM if the given address is different or - * reading the current address failed - */ - if (!ret_val && memcmp(current_mac, mac, 6) == 0) - return 0; - - for (i = 0; i < 3; ++i) - data[i] = mac[i * 2 + 1] << 8 | mac[i * 2]; - - ret_val = e1000_write_eeprom_srwr(hw, 0x0, 3, data); - - if (!ret_val) - ret_val = e1000_update_eeprom_checksum_i210(hw); - - return ret_val; -#else - return 0; -#endif -} - /************************************************************************** PROBE - Look for an adapter, this routine's visible to the outside You should omit the last argument struct pci_device * for a non-PCI NIC @@ -5914,6 +5920,7 @@ static const struct eth_ops e1000_eth_ops = { .recv = e1000_eth_recv, .stop = e1000_eth_stop, .free_pkt = e1000_free_pkt, + .write_hwaddr = e1000_write_hwaddr, }; static const struct udevice_id e1000_eth_ids[] = { -- 2.7.4

4 7

[U-Boot] [PATCH] x86/bootm: fix compressed image boot
by Hannes Schmelzer 24 Nov '18

24 Nov '18

If we're booting some u-boot module with compressed payload, we have to use the pointer where the image really has been loaded (unzipped) to instead the pointer to the payload of the u-boot module. Signed-off-by: Hannes Schmelzer <hannes.schmelzer(a)br-automation.com> --- arch/x86/lib/bootm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/x86/lib/bootm.c b/arch/x86/lib/bootm.c index 54c22fe..4102bcb 100644 --- a/arch/x86/lib/bootm.c +++ b/arch/x86/lib/bootm.c @@ -94,8 +94,8 @@ static int boot_prep_linux(bootm_headers_t *images) len = os_len; } else { /* otherwise get image data */ - data = (void *)image_get_data(hdr); - len = image_get_data_size(hdr); + data = (void *)images->os.load; + len = 0; } is_zimage = 1; #if defined(CONFIG_FIT) -- 2.7.4

3 5

[U-Boot] [PATCH 1/1] efi_loader: remove efi_exit_caches()
by Heinrich Schuchardt 24 Nov '18

24 Nov '18

Since GRUB patch d0c070179d4d ("arm/efi: Switch to arm64 linux loader", 2018-07-09) we do not need a workaround for GRUB on 32bit ARM anymore. So let's eliminate function efi_exit_caches(). This will require Linux distributions to update grub-efi-arm to the GRUB git HEAD (a tag containing the aforementioned GRUB patch is not available yet). Signed-off-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> --- lib/efi_loader/efi_boottime.c | 28 ---------------------------- 1 file changed, 28 deletions(-) diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c index 2496608981..97eb19cd14 100644 --- a/lib/efi_loader/efi_boottime.c +++ b/lib/efi_loader/efi_boottime.c @@ -26,14 +26,6 @@ LIST_HEAD(efi_obj_list); /* List of all events */ LIST_HEAD(efi_events); -/* - * If we're running on nasty systems (32bit ARM booting into non-EFI Linux) - * we need to do trickery with caches. Since we don't want to break the EFI - * aware boot path, only apply hacks when loading exiting directly (breaking - * direct Linux EFI booting along the way - oh well). - */ -static bool efi_is_direct_boot = true; - #ifdef CONFIG_ARM /* * The "gd" pointer lives in a register on ARM and AArch64 that we declare @@ -1686,8 +1678,6 @@ static efi_status_t EFIAPI efi_start_image(efi_handle_t image_handle, EFI_ENTRY("%p, %p, %p", image_handle, exit_data_size, exit_data); - efi_is_direct_boot = false; - /* call the image! */ if (setjmp(&image_obj->exit_jmp)) { /* @@ -1795,21 +1785,6 @@ static efi_status_t EFIAPI efi_unload_image(efi_handle_t image_handle) return EFI_EXIT(EFI_SUCCESS); } -/** - * efi_exit_caches() - fix up caches for EFI payloads if necessary - */ -static void efi_exit_caches(void) -{ -#if defined(CONFIG_ARM) && !defined(CONFIG_ARM64) - /* - * Grub on 32bit ARM needs to have caches disabled before jumping into - * a zImage, but does not know of all cache layers. Give it a hand. - */ - if (efi_is_direct_boot) - cleanup_before_linux(); -#endif -} - /** * efi_exit_boot_services() - stop all boot services * @image_handle: handle of the loaded image @@ -1863,9 +1838,6 @@ static efi_status_t EFIAPI efi_exit_boot_services(efi_handle_t image_handle, board_quiesce_devices(); - /* Fix up caches for EFI payloads if necessary */ - efi_exit_caches(); - /* This stops all lingering devices */ bootm_disable_interrupts(); -- 2.19.0

4 3