August 2023 - U-Boot

Porting feature to ancient U-boot V.S porting ancient hardware code patches to new U-boot
by Mr Academia Euforia 09 Aug '23

09 Aug '23

Hi there, I'm working on certain feature to be added to ONIE that is running on Broadcom's Helix4 -based network switch. Part of the task is to add FIT signature signing to U-boot versions supported by ONIE. ONIE as part of the build process uses uboot snapshots, and applies platform patches on top of it. The list of U-boot snapshots to pick from is as follows: 2012.10, 2013.01.01, 2015.10, 2018.03 My platform by default, uses the oldest, 2012.10 and this version does not support FIT signatures. I was thinking how to address that, and came up with two solutions, neither being better (in my mind): 1) Create patch set based on the difference between more recent U-boot versions and 2012 that would be focused around fit image signatures, 2) Pick 2018.03 as code base. Problems I see: Ad 1) I can't find any repository that would contain commits from that far away history-wise. Did U-boot project even use any public repository in 2012 ? General problem here is how to prepare patchset, Ad 2) The problem is that platform code patches don't apply on more recent U-boot versions. Some of the files no longer exist, some code has been refactored, etc. Could you guys offer me any sort of advise on how to approach this ? Best regards

2 1

[u-boot-test-hooks 1/4] bin/flash.sdwire_common_mount: Switch to sourcing the next writer script
by Tom Rini 09 Aug '23

09 Aug '23

Rather than invoking the script that will write to the mounted directory as a binary, source it as a script so that it has access to more than just two parameters. This will allow us to have the same flexibility in our writers that other flash methods have. Signed-off-by: Tom Rini <trini(a)konsulko.com> --- bin/flash.sdwire_common_mount | 3 ++- bin/writer.rpi3_mount | 19 +++++++------------ bin/writer.zynq_mount | 15 +++++---------- 3 files changed, 14 insertions(+), 23 deletions(-) diff --git a/bin/flash.sdwire_common_mount b/bin/flash.sdwire_common_mount index 6c763e62f47d..b76add064fb0 100644 --- a/bin/flash.sdwire_common_mount +++ b/bin/flash.sdwire_common_mount @@ -52,7 +52,8 @@ if ! mountpoint -q ${mount_dir}; then exit 1 fi -writer.${flash_writer} ${mount_dir} ${U_BOOT_BUILD_DIR} +# Perform the write, pass along as much environment as possible +. writer.${flash_writer} complete=false for i in {0..9}; do diff --git a/bin/writer.rpi3_mount b/bin/writer.rpi3_mount index 97f24a5ac694..a63e7999e57b 100755 --- a/bin/writer.rpi3_mount +++ b/bin/writer.rpi3_mount @@ -20,23 +20,18 @@ # Writes rpi3_b to the board -# Args: -# $1: Mount point of the sdcard when board is off -# $2: U-Boot build directory - set -ex -mount=$1 -build=$2 +build=${U_BOOT_BUILD_DIR} -echo "Writing to ${mount} from build at ${build}" +echo "Writing to ${mount_dir} from build at ${build}" # First make a copy of the original files if we haven't already -if [[ ! -e ${mount}/config.orig ]]; then - cp ${mount}/config.txt ${mount}/config.orig +if [[ ! -e ${mount_dir}/config.orig ]]; then + cp ${mount_dir}/config.txt ${mount_dir}/config.orig fi -if [[ ! -e ${mount}/rpi3-u-boot.bin.orig ]]; then - cp ${mount}/rpi3-u-boot.bin ${mount}/rpi3-u-boot.bin.orig +if [[ ! -e ${mount_dir}/rpi3-u-boot.bin.orig ]]; then + cp ${mount_dir}/rpi3-u-boot.bin ${mount_dir}/rpi3-u-boot.bin.orig fi # Enable the UART and fix the GPU frequency so it works correctly @@ -46,4 +41,4 @@ if ! grep -q "^gpu_freq=250" /media/rpi3_b_boot/config.txt; then fi # Copy U-Boot over from the build directory -cp ${build}/u-boot.bin ${mount}/rpi3-u-boot.bin +cp ${build}/u-boot.bin ${mount_dir}/rpi3-u-boot.bin diff --git a/bin/writer.zynq_mount b/bin/writer.zynq_mount index 9d0958880422..c8395a40680e 100755 --- a/bin/writer.zynq_mount +++ b/bin/writer.zynq_mount @@ -20,22 +20,17 @@ # Writes zynq images to the board -# Args: -# $1: Mount point of the sdcard when board is off -# $2: U-Boot build directory - set -ex tmp=$(mktemp -d) -mount=$1 -build=$2 +build=${U_BOOT_BUILD_DIR} -echo "Writing to ${mount} from build at ${build}" +echo "Writing to ${mount_dir} from build at ${build}" # Copy U-Boot over from the build directory -cp ${build}/u-boot.bin ${mount}/rpi3-u-boot.bin +cp ${build}/u-boot.bin ${mount_dir}/rpi3-u-boot.bin zynq-boot-bin.py -o ${tmp}/boot.bin -u ${build}/spl/u-boot-spl-dtb.bin -cp ${tmp}/boot.bin ${mount}/BOOT.bin -cp ${build}/u-boot.img ${mount}/. +cp ${tmp}/boot.bin ${mount_dir}/BOOT.bin +cp ${build}/u-boot.img ${mount_dir}/. rm -rf ${tmp} -- 2.34.1

3 8

[PATCH] btrfs: fix some error checking for btrfs_decompress()
by Dan Carpenter 09 Aug '23

09 Aug '23

The btrfs_decompress() function mostly (u32)-1 on error but it can also return -EPERM or other kernel error codes from zstd_decompress(). The "ret" variable is an int, so we could just check for negatives. Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> --- fs/btrfs/inode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 38e285bf94b0..4691612eda33 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -390,7 +390,7 @@ int btrfs_read_extent_inline(struct btrfs_path *path, csize); ret = btrfs_decompress(btrfs_file_extent_compression(leaf, fi), cbuf, csize, dbuf, dsize); - if (ret == (u32)-1) { + if (ret < 0) { ret = -EIO; goto out; } @@ -500,7 +500,7 @@ int btrfs_read_extent_reg(struct btrfs_path *path, ret = btrfs_decompress(btrfs_file_extent_compression(leaf, fi), cbuf, csize, dbuf, dsize); - if (ret == (u32)-1) { + if (ret < 0) { ret = -EIO; goto out; } -- 2.39.2

3 2

[PATCH] test: fix a couple NULL vs IS_ERR() checks
by Dan Carpenter 09 Aug '23

09 Aug '23

The x509_cert_parse() and pkcs7_parse_message() functions return error pointers. They don't return NULL. Update the checks accordingly. Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> --- test/lib/asn1.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/test/lib/asn1.c b/test/lib/asn1.c index 8661fdd30687..a66cdd77df0a 100644 --- a/test/lib/asn1.c +++ b/test/lib/asn1.c @@ -120,7 +120,7 @@ static int lib_asn1_x509(struct unit_test_state *uts) cert = x509_cert_parse(cert_data, cert_data_len); - ut_assertf(cert != NULL, "decoding failed\n"); + ut_assertf(!IS_ERR(cert), "decoding failed\n"); ut_assertf(!strcmp(cert->subject, "Linaro: Tester"), "subject doesn't match\n"); ut_assertf(!strcmp(cert->issuer, "Linaro: Tester"), @@ -313,7 +313,7 @@ static int lib_asn1_pkcs7(struct unit_test_state *uts) pkcs7 = pkcs7_parse_message(image_pk7, image_pk7_len); - ut_assertf(pkcs7 != NULL, "decoding failed\n"); + ut_assertf(!IS_ERR(pkcs7), "decoding failed\n"); ut_assertf(pkcs7->data_len == 104, "signature size doesn't match\n"); ut_assertf(pkcs7->signed_infos != NULL, "sign-info doesn't exist\n"); ut_assertf(pkcs7->signed_infos->msgdigest_len == 32, -- 2.39.2

3 2

[PATCH] expo: allocate correct ammount of memory
by Dan Carpenter 09 Aug '23

09 Aug '23

This should be allocating the memory for "item" instead of "menu". The item struct is 48 bytes instead of 96 (assuming a 64bit system) so this saves a little memory. Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> --- boot/scene_menu.c | 2 +- 1 file changed, 1 insertions(+), 1 deletions(-) diff --git a/boot/scene_menu.c b/boot/scene_menu.c index 8a355f838cc8..57ffb523ff3f 100644 --- a/boot/scene_menu.c +++ b/boot/scene_menu.c @@ -416,7 +416,7 @@ int scene_menuitem(struct scene *scn, uint menu_id, const char *name, uint id, if (!scene_obj_find(scn, label_id, SCENEOBJT_TEXT)) return log_msg_ret("txt", -EINVAL); - item = calloc(1, sizeof(struct scene_obj_menu)); + item = calloc(1, sizeof(struct scene_menitem)); if (!item) return log_msg_ret("item", -ENOMEM); item->name = strdup(name); -- 2.39.2

3 2

[PATCH] cmd: improve string matching for hex
by Dan Carpenter 09 Aug '23

09 Aug '23

Match the "=0x" instead of just "=0". Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> --- We sometimes two character partial matching for commands so people can type "re" instead of "read". But here reading two characters doesn't seem correct. cmd/nvedit_efi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/cmd/nvedit_efi.c b/cmd/nvedit_efi.c index 24944ab81e23..7a30b5cc8f87 100644 --- a/cmd/nvedit_efi.c +++ b/cmd/nvedit_efi.c @@ -262,7 +262,7 @@ static int append_value(char **bufp, size_t *sizep, char *data) char *tmp_buf = NULL, *new_buf = NULL, *value; unsigned long len = 0; - if (!strncmp(data, "=0x", 2)) { /* hexadecimal number */ + if (!strncmp(data, "=0x", 3)) { /* hexadecimal number */ union { u8 u8; u16 u16; -- 2.39.2

5 4

[PATCH 1/1] pci: correct function name in message
by Heinrich Schuchardt 09 Aug '23

09 Aug '23

If an error message contains a function name, it should match the name of the function throwing the message. Fixes: 7739d93d8288 ("pci: Match region flags using a mask") Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt(a)canonical.com> --- drivers/pci/pci-uclass.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/pci/pci-uclass.c b/drivers/pci/pci-uclass.c index 632c1a63cf..7f3d6ddf91 100644 --- a/drivers/pci/pci-uclass.c +++ b/drivers/pci/pci-uclass.c @@ -1446,7 +1446,7 @@ phys_addr_t dm_pci_bus_to_phys(struct udevice *dev, pci_addr_t bus_addr, return res->phys_start + offset; } - puts("pci_hose_bus_to_phys: invalid physical address\n"); + puts("dm_pci_bus_to_phys: invalid physical address\n"); return 0; } @@ -1486,7 +1486,7 @@ pci_addr_t dm_pci_phys_to_bus(struct udevice *dev, phys_addr_t phys_addr, return res->bus_start + offset; } - puts("pci_hose_phys_to_bus: invalid physical address\n"); + puts("dm_pci_phys_to_bus: invalid physical address\n"); return 0; } -- 2.40.1

3 2

[PATCH] cramfs: clean up some error messages
by Dan Carpenter 09 Aug '23

09 Aug '23

This line break is not done correctly. We don't want to have all those tabs in the printed output. Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> --- fs/cramfs/cramfs.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/cramfs/cramfs.c b/fs/cramfs/cramfs.c index 6c017cebc50f..abb2de34eb05 100644 --- a/fs/cramfs/cramfs.c +++ b/fs/cramfs/cramfs.c @@ -166,8 +166,7 @@ static unsigned long cramfs_resolve (unsigned long begin, unsigned long offset, unsigned long ret; char *link; if (p && strlen(p)) { - printf ("unsupported symlink to \ - non-terminal path\n"); + printf ("unsupported symlink to non-terminal path\n"); return 0; } link = cramfs_uncompress_link (begin, @@ -177,8 +176,7 @@ static unsigned long cramfs_resolve (unsigned long begin, unsigned long offset, namelen, namelen, name); return 0; } else if (link[0] == '/') { - printf ("unsupported symlink to \ - absolute path\n"); + printf ("unsupported symlink to absolute path\n"); free (link); return 0; } -- 2.39.2

3 2

[PATCH] test: unicode: fix a sizeof() vs ARRAY_SIZE() bug
by Dan Carpenter 09 Aug '23

09 Aug '23

The u16_strlcat() is in units of u16 not bytes. So the limit needs to be ARRAY_SIZE() instead of sizeof(). Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> --- test/unicode_ut.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/test/unicode_ut.c b/test/unicode_ut.c index b27d7116b9ee..62ff5d10bf94 100644 --- a/test/unicode_ut.c +++ b/test/unicode_ut.c @@ -807,12 +807,12 @@ static int unicode_test_u16_strlcat(struct unit_test_state *uts) /* dest and src are empty string */ memset(buf, 0, sizeof(buf)); - ret = u16_strlcat(buf, &null_src, sizeof(buf)); + ret = u16_strlcat(buf, &null_src, ARRAY_SIZE(buf)); ut_asserteq(1, ret); /* dest is empty string */ memset(buf, 0, sizeof(buf)); - ret = u16_strlcat(buf, src, sizeof(buf)); + ret = u16_strlcat(buf, src, ARRAY_SIZE(buf)); ut_asserteq(5, ret); ut_assert(!unicode_test_u16_strcmp(buf, src, 40)); @@ -820,7 +820,7 @@ static int unicode_test_u16_strlcat(struct unit_test_state *uts) memset(buf, 0xCD, (sizeof(buf) - sizeof(u16))); buf[39] = 0; memcpy(buf, dest, sizeof(dest)); - ret = u16_strlcat(buf, &null_src, sizeof(buf)); + ret = u16_strlcat(buf, &null_src, ARRAY_SIZE(buf)); ut_asserteq(6, ret); ut_assert(!unicode_test_u16_strcmp(buf, dest, 40)); -- 2.39.2

3 2

[PATCH] cmd: pxe_utils: add some missing tabs
by Dan Carpenter 09 Aug '23

09 Aug '23

These lines are supposed to be indented one more tab. Otherwise it's confusing to read. Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> --- boot/pxe_utils.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/boot/pxe_utils.c b/boot/pxe_utils.c index d13c47dd9429..ac1414a5f26d 100644 --- a/boot/pxe_utils.c +++ b/boot/pxe_utils.c @@ -702,8 +702,8 @@ static int label_boot(struct pxe_context *ctx, struct pxe_label *label) } } - if (label->kaslrseed) - label_boot_kaslrseed(); + if (label->kaslrseed) + label_boot_kaslrseed(); #ifdef CONFIG_OF_LIBFDT_OVERLAY if (label->fdtoverlays) -- 2.39.2

4 3