July 2017 - U-Boot

[U-Boot] [PATCH v10 00/10] usb: Extend ehci and ohci generic driver
by patrice.chotard＠st.com 07 Jul '17

07 Jul '17

From: Patrice Chotard <patrice.chotard(a)st.com> This series improves generic ehci and ohci drivers by addition of : _ error path during probe (clocks, resets and phy release) _ .remove callback _ add generic PHY framework for both generic ehci and ohci drivers _ add RESET and CLOCK framework for generic ohci driver To implement these features, some new methods are needed in reset, clock and in dm/core framework: _ add reset_request() and reset_assert_all() methods in RESET framework _ add clk_count() and clk_disable_all() methods in CLOCK framework _ add ofnode_count_phandle_with_args() and dev_count_phandle_with_args() in dm/core v10: _ add dev_count_phandle_with_args() requested by Simon Glass _ fix some mirno remarks v9: _ rename reset_assert_all() in reset_release_all() as this function not only assert all resets but also free all of them _ rename clk_disable_all() in clk_release_all() as this function not only disable all clocks but also free all of them _ add a check in reset_release_all()/clk_disable_all() to verify if reset/clock has been previously requested before asserting/disabling and freeing it. v8: _ rework error path by propagating the initial error code until the end of probe() _ replace devm_kmalloc() with devm_kcalloc() _ fix cosmetics remarks v7: _ replace clk_count() and reset_count() methods by ofnode_count_phandle_with_args() in patches 3, 4 and 5 v6: _ replace clk_get_by_index() by dev_read_phandle_with_args() in clk_count() in patch 4 _ add Reviewed-by Simon Glass for patch 2 and 5 v5: _ rebase on top of dm/master requested by Simon Glass in order to use livetree update _ replace fdtdec_parse_phandle_with_args() by dev_read_phandle_with_args() in patch 2 v4: _ add clk_disable_all() and reset_assert_all() methods into CLOCK and RESET framework as suggested by Simon Glass and Marek Vasut _ add reset_count() and clk_count() methods which returns respectively the number of resets and clocks declared into "resets" and "clocks" DT properties. This allows to allocate the right amount of memory to keep resets and clocks reference _ update the memory allocation for deasserted resets and enabled clocks reference list. Replace lists by arrays. v3: _ keep enabled clocks and deasserted resets reference in list in order to disable clock or assert resets in error path or in .remove callback _ add missing commit message _ use struct generic_ehci * instead of struct udevice * as parameter for ehci_release_resets() and ehci_release_clocks() _ test return value on generic_phy_get_by_index() and generic_phy_init() _ split previous patch 5 in 3 independant patch for CLOCK, RESET and PHY support v2: _ add needed reset_request() in RESET framework _ add error path in ehci/ohci-generic to disable clocks and to assert resets _ add .remove callback with clocks, resets and phy release _ split the replacement of printf() by error() in an independant patch Patrice Chotard (10): reset: add reset_request() reset: add reset_release_all() clk: add clk_release_all() dm: core: add ofnode_count_phandle_with_args() usb: host: ehci-generic: replace printf() by error() usb: host: ehci-generic: add error path and .remove callback usb: host: ehci-generic: add generic PHY support usb: host: ohci-generic: add CLOCK support usb: host: ohci-generic: add RESET support usb: host: ohci-generic: add generic PHY support drivers/clk/clk-uclass.c | 26 +++++++ drivers/core/of_access.c | 7 ++ drivers/core/ofnode.c | 12 ++++ drivers/reset/reset-uclass.c | 34 ++++++++++ drivers/usb/host/ehci-generic.c | 147 +++++++++++++++++++++++++++++++++------- drivers/usb/host/ohci-generic.c | 122 ++++++++++++++++++++++++++++++++- include/clk.h | 14 ++++ include/dm/of_access.h | 18 +++++ include/dm/ofnode.h | 17 +++++ include/dm/read.h | 25 +++++++ include/reset.h | 27 ++++++++ 11 files changed, 424 insertions(+), 25 deletions(-) -- 1.9.1

2 14

[U-Boot] [PATCH] dm: include <dm/util.h> from drivers/core/util.c
by Masahiro Yamada 07 Jul '17

07 Jul '17

Fix sparse warnings "... was not declared. Should it be static?" Also, fix redefinition of dm_warn/dm_dbg. Signed-off-by: Masahiro Yamada <yamada.masahiro(a)socionext.com> --- drivers/core/util.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/core/util.c b/drivers/core/util.c index 5ceac8bbb15b..2e232d57a14f 100644 --- a/drivers/core/util.c +++ b/drivers/core/util.c @@ -5,9 +5,11 @@ */ #include <common.h> +#include <dm/util.h> #include <libfdt.h> #include <vsprintf.h> +#ifdef CONFIG_DM_WARN void dm_warn(const char *fmt, ...) { va_list args; @@ -16,7 +18,9 @@ void dm_warn(const char *fmt, ...) vprintf(fmt, args); va_end(args); } +#endif +#ifdef DEBUG void dm_dbg(const char *fmt, ...) { va_list args; @@ -25,6 +29,7 @@ void dm_dbg(const char *fmt, ...) vprintf(fmt, args); va_end(args); } +#endif int list_count_items(struct list_head *head) { -- 2.7.4

4 5

[U-Boot] [PATCH] fdt: Check for NULL return from fdt_getprop in 'fdt set'
by Tom Rini 06 Jul '17

06 Jul '17

While the previous pass through fixed one place where we knew that fdt_getprop would be given a positive len, in the case of 'fdt set' we do not, so check that we did no get NULL from fdt_getprop(). Cc: Simon Glass <sjg(a)chromium.org> Reported-by: Coverity (CID: 163249) Fixes 72c98ed1ab48 ("fdt: Add a check to do_fdt() for coverity") Signed-off-by: Tom Rini <trini(a)konsulko.com> --- cmd/fdt.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/cmd/fdt.c b/cmd/fdt.c index 31a536198c8e..05e19f8a72bf 100644 --- a/cmd/fdt.c +++ b/cmd/fdt.c @@ -284,6 +284,10 @@ static int do_fdt(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) len = 0; } else { ptmp = fdt_getprop(working_fdt, nodeoffset, prop, &len); + if (!ptmp) { + printf("prop (%s) not found!\n", prop); + return 1; + } if (len > SCRATCHPAD) { printf("prop (%d) doesn't fit in scratchpad!\n", len); -- 1.9.1

3 2

[U-Boot] [PATCH] dm: include <dm/util.h> from driver/core/dump.c
by Masahiro Yamada 06 Jul '17

06 Jul '17

Include <dm/util.h> to fix sparse warnings: symbol 'dm_dump_all' was not declared. Should it be static? symbol 'dm_dump_uclass' was not declared. Should it be static? Signed-off-by: Masahiro Yamada <yamada.masahiro(a)socionext.com> --- drivers/core/dump.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/core/dump.c b/drivers/core/dump.c index fd4596ee68b7..c3e109e7ed11 100644 --- a/drivers/core/dump.c +++ b/drivers/core/dump.c @@ -8,6 +8,7 @@ #include <dm.h> #include <mapmem.h> #include <dm/root.h> +#include <dm/util.h> static void show_devices(struct udevice *dev, int depth, int last_flag) { -- 2.7.4

4 4

[U-Boot] [PATCH] dm: ofnode: use fdt32_t for DT property value to fix sparse warning
by Masahiro Yamada 06 Jul '17

06 Jul '17

DTB is encoded in big endian. When we retrieve property values, we need to use fdt32_to_cpu (aka be32_to_cpu) for endian conversion. This is a bit error-prone, but sparse is useful to detect endian mismatch. We need to use (fdt32_t *) instead of (u32 *) for a pointer of a property value. Otherwise sparse warns "cast to restricted __be32". Signed-off-by: Masahiro Yamada <yamada.masahiro(a)socionext.com> --- drivers/core/ofnode.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/core/ofnode.c b/drivers/core/ofnode.c index 5a3e1d452fbb..0ad0add7cb18 100644 --- a/drivers/core/ofnode.c +++ b/drivers/core/ofnode.c @@ -23,7 +23,7 @@ int ofnode_read_u32(ofnode node, const char *propname, u32 *outp) if (ofnode_is_np(node)) { return of_read_u32(ofnode_to_np(node), propname, outp); } else { - const int *cell; + const fdt32_t *cell; int len; cell = fdt_getprop(gd->fdt_blob, ofnode_to_offset(node), @@ -477,7 +477,7 @@ const uint8_t *ofnode_read_u8_array_ptr(ofnode node, const char *propname, int ofnode_read_pci_addr(ofnode node, enum fdt_pci_space type, const char *propname, struct fdt_pci_addr *addr) { - const u32 *cell; + const fdt32_t *cell; int len; int ret = -ENOENT; -- 2.7.4

3 2

[U-Boot] [PATCH v3 1/6] display_options: Refactor to allow obtaining the banner
by Simon Glass 06 Jul '17

06 Jul '17

Move the display options code into a separate function so that the U-Boot banner can be obtained from other code. Adjust the 'version' command to use it. Signed-off-by: Simon Glass <sjg(a)chromium.org> --- Changes in v3: None Changes in v2: - Update the code to be more testable - Avoid a warning on spring, etc. cmd/version.c | 4 +++- include/display_options.h | 19 +++++++++++++++++++ lib/display_options.c | 36 +++++++++++++++++++++++++++++++----- 3 files changed, 53 insertions(+), 6 deletions(-) diff --git a/cmd/version.c b/cmd/version.c index 1be0667f09..15aab5dc18 100644 --- a/cmd/version.c +++ b/cmd/version.c @@ -17,7 +17,9 @@ const char __weak version_string[] = U_BOOT_VERSION_STRING; static int do_version(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) { - printf("\n%s\n", version_string); + char buf[DISPLAY_OPTIONS_BANNER_LENGTH]; + + printf(display_options_get_banner(false, buf, sizeof(buf))); #ifdef CC_VERSION_STRING puts(CC_VERSION_STRING "\n"); #endif diff --git a/include/display_options.h b/include/display_options.h index ac44c459b3..ad707a344c 100644 --- a/include/display_options.h +++ b/include/display_options.h @@ -56,4 +56,23 @@ int print_buffer(ulong addr, const void *data, uint width, uint count, */ int display_options(void); +/* Suggested length of the buffer to pass to display_options_get_banner() */ +#define DISPLAY_OPTIONS_BANNER_LENGTH 200 + +/** + * display_options_get_banner() - Get the U-Boot banner as a string + * + * This returns the U-Boot banner string + * + * @newlines: true to include two newlines at the start + * @buf: place to put string + * @size: Size of buf (string is truncated to fit) + * @return buf + */ +char *display_options_get_banner(bool newlines, char *buf, int size); + +/** This function is used for testing only */ +char *display_options_get_banner_priv(bool newlines, const char *build_tag, + char *buf, int size); + #endif diff --git a/lib/display_options.c b/lib/display_options.c index 29343fc00e..4ea27ca99d 100644 --- a/lib/display_options.c +++ b/lib/display_options.c @@ -13,13 +13,39 @@ #include <linux/ctype.h> #include <asm/io.h> -int display_options (void) +char *display_options_get_banner_priv(bool newlines, const char *build_tag, + char *buf, int size) { -#if defined(BUILD_TAG) - printf ("\n\n%s, Build: %s\n\n", version_string, BUILD_TAG); -#else - printf ("\n\n%s\n\n", version_string); + int len; + + len = snprintf(buf, size, "%s%s", newlines ? "\n\n" : "", + version_string); + if (build_tag && len < size) + len += snprintf(buf + len, size - len, ", Build: %s", + build_tag); + if (len > size - 3) + len = size - 3; + strcpy(buf + len, "\n\n"); + + return buf; +} + +#ifndef BUILD_TAG +#define BUILD_TAG NULL #endif + +char *display_options_get_banner(bool newlines, char *buf, int size) +{ + return display_options_get_banner_priv(newlines, BUILD_TAG, buf, size); +} + +int display_options(void) +{ + char buf[DISPLAY_OPTIONS_BANNER_LENGTH]; + + display_options_get_banner(true, buf, sizeof(buf)); + printf("%s", buf); + return 0; } -- 2.13.1.518.g3df882009-goog

3 16

[U-Boot] [PATCH 0/3] moveconfig: Allow automatic adding of 'imply' to options
by Simon Glass 06 Jul '17

06 Jul '17

A common requirement when converting options to Kconfig is to minimise the number of defconfig files that must be changed. For example, if the pre-Kconfig CONFIG_SCSI option is defined in zynq-common.h, then hopefully it can be added as an 'imply SCSI' to the Kconfig under ARCH_ZYNQ. This allows a single Kconfig change and avoids changing 18 zynq defconfig files. Unfortunately it is extremely tedious to find which 'implying' configs are best for each option. The moveconfig database helps with this since it is possible to obtain a list of possible implying configs. To further help with the pain this series automates this work a bit. It uses a Kconfig parser to find the location of each option in the Kconfig tree. Using the information from the database it is then able to automatically add an 'imply' option into the right place if requested by the user. If an 'imply' already exists we can show its location as confirmation that it exists. Simon Glass (3): moveconfig: Allow piping in 'git show --stat' output moveconfig: Allow control of which implying configs are shown moveconfig: Allow automatic location and adding of 'imply' tools/moveconfig.py | 256 ++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 246 insertions(+), 10 deletions(-) -- 2.13.1.518.g3df882009-goog

2 6

[U-Boot] [PATCH v4 1/6] display_options: Refactor to allow obtaining the banner
by Simon Glass 06 Jul '17

06 Jul '17

Move the display options code into a separate function so that the U-Boot banner can be obtained from other code. Adjust the 'version' command to use it. Signed-off-by: Simon Glass <sjg(a)chromium.org> Reviewed-by: Bin Meng <bmeng.cn(a)gmail.com> Tested-by: Bin Meng <bmeng.cn(a)gmail.com> --- Changes in v4: - Fix /** nit in header file Changes in v3: None Changes in v2: - Update the code to be more testable - Avoid a warning on spring, etc. cmd/version.c | 4 +++- include/display_options.h | 19 +++++++++++++++++++ lib/display_options.c | 36 +++++++++++++++++++++++++++++++----- 3 files changed, 53 insertions(+), 6 deletions(-) diff --git a/cmd/version.c b/cmd/version.c index 1be0667f09..15aab5dc18 100644 --- a/cmd/version.c +++ b/cmd/version.c @@ -17,7 +17,9 @@ const char __weak version_string[] = U_BOOT_VERSION_STRING; static int do_version(cmd_tbl_t *cmdtp, int flag, int argc, char * const argv[]) { - printf("\n%s\n", version_string); + char buf[DISPLAY_OPTIONS_BANNER_LENGTH]; + + printf(display_options_get_banner(false, buf, sizeof(buf))); #ifdef CC_VERSION_STRING puts(CC_VERSION_STRING "\n"); #endif diff --git a/include/display_options.h b/include/display_options.h index ac44c459b3..d9c8f6dbd4 100644 --- a/include/display_options.h +++ b/include/display_options.h @@ -56,4 +56,23 @@ int print_buffer(ulong addr, const void *data, uint width, uint count, */ int display_options(void); +/* Suggested length of the buffer to pass to display_options_get_banner() */ +#define DISPLAY_OPTIONS_BANNER_LENGTH 200 + +/** + * display_options_get_banner() - Get the U-Boot banner as a string + * + * This returns the U-Boot banner string + * + * @newlines: true to include two newlines at the start + * @buf: place to put string + * @size: Size of buf (string is truncated to fit) + * @return buf + */ +char *display_options_get_banner(bool newlines, char *buf, int size); + +/* This function is used for testing only */ +char *display_options_get_banner_priv(bool newlines, const char *build_tag, + char *buf, int size); + #endif diff --git a/lib/display_options.c b/lib/display_options.c index 29343fc00e..4ea27ca99d 100644 --- a/lib/display_options.c +++ b/lib/display_options.c @@ -13,13 +13,39 @@ #include <linux/ctype.h> #include <asm/io.h> -int display_options (void) +char *display_options_get_banner_priv(bool newlines, const char *build_tag, + char *buf, int size) { -#if defined(BUILD_TAG) - printf ("\n\n%s, Build: %s\n\n", version_string, BUILD_TAG); -#else - printf ("\n\n%s\n\n", version_string); + int len; + + len = snprintf(buf, size, "%s%s", newlines ? "\n\n" : "", + version_string); + if (build_tag && len < size) + len += snprintf(buf + len, size - len, ", Build: %s", + build_tag); + if (len > size - 3) + len = size - 3; + strcpy(buf + len, "\n\n"); + + return buf; +} + +#ifndef BUILD_TAG +#define BUILD_TAG NULL #endif + +char *display_options_get_banner(bool newlines, char *buf, int size) +{ + return display_options_get_banner_priv(newlines, BUILD_TAG, buf, size); +} + +int display_options(void) +{ + char buf[DISPLAY_OPTIONS_BANNER_LENGTH]; + + display_options_get_banner(true, buf, sizeof(buf)); + printf("%s", buf); + return 0; } -- 2.13.1.518.g3df882009-goog

2 1

[U-Boot] [PATCH 00/11] dtoc: Tidy up and add tests
by Simon Glass 06 Jul '17

06 Jul '17

When dtoc was originally written it was an idea without much expectation (on my part at least) that it would be widely used. However there have been several features requests and it is being relied on for several rockchip boards. At present dtoc has no tests of its own. It relies on sandbox_spl which uses CONFIG_SPL_OF_PLATDATA to read the resulting C structures and output their contents. This is a rather round-about way to test the tool. It is better to have tests which specifically test dtoc's functionality. In addition, before expanding the code it would be best to refactor it a bit to make it easier to maintain. This series: - Tidies up and refactors dtoc (pylint, static functions, splitting code) - Adds test for the main features it supports Simon Glass (11): dtoc: Use self._options instead of the global options dtoc: Add a comment at the top dtoc: Split out the main class into its own file dtoc: Fix pylint warnings dtoc: Don't handle properties with / in them dtoc: Pass include_disabled explicitly dtoc: Move static functions out of the class dtoc: Move the main logic into the dtb_platdata file dtoc: Add a comment about string replace in conv_name_to_c() sandbox: Stop printing platdata at the start of SPL dtoc: Add tests arch/sandbox/cpu/spl.c | 11 - test/py/tests/test_ofplatdata.py | 30 +-- tools/dtoc/dtb_platdata.py | 456 ++++++++++++++++++++++++++++++++++++++ tools/dtoc/dtoc.py | 466 ++++----------------------------------- tools/dtoc/dtoc_test.dts | 12 + tools/dtoc/dtoc_test_aliases.dts | 18 ++ tools/dtoc/dtoc_test_empty.dts | 12 + tools/dtoc/dtoc_test_phandle.dts | 23 ++ tools/dtoc/dtoc_test_simple.dts | 48 ++++ tools/dtoc/test_dtoc.py | 271 +++++++++++++++++++++++ 10 files changed, 887 insertions(+), 460 deletions(-) create mode 100644 tools/dtoc/dtb_platdata.py create mode 100644 tools/dtoc/dtoc_test.dts create mode 100644 tools/dtoc/dtoc_test_aliases.dts create mode 100644 tools/dtoc/dtoc_test_empty.dts create mode 100644 tools/dtoc/dtoc_test_phandle.dts create mode 100644 tools/dtoc/dtoc_test_simple.dts create mode 100644 tools/dtoc/test_dtoc.py -- 2.13.1.518.g3df882009-goog

2 28

[U-Boot] [PATCH] binman: Put our local modules ahead of system modules
by Simon Glass 06 Jul '17

06 Jul '17

If a system module is named the same as one of those used by binman we currently pick the system module. Adjust the ordering so that our modules are chosen instead. The module conflict reported was 'tools' from jira-python. I cannot access that package to test it. Signed-off-by: Simon Glass <sjg(a)chromium.org> Reported-by: Kevin Hilman <khilman(a)baylibre.com> --- tools/binman/binman.py | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/tools/binman/binman.py b/tools/binman/binman.py index 95d3a048d8..09dc36a3f7 100755 --- a/tools/binman/binman.py +++ b/tools/binman/binman.py @@ -17,15 +17,14 @@ import unittest # Bring in the patman and dtoc libraries our_path = os.path.dirname(os.path.realpath(__file__)) -sys.path.append(os.path.join(our_path, '../patman')) -sys.path.append(os.path.join(our_path, '../dtoc')) -sys.path.append(os.path.join(our_path, '../')) +for dirname in ['../patman', '../dtoc', '..']: + sys.path.insert(0, os.path.join(our_path, dirname)) # Bring in the libfdt module -sys.path.append('tools') +sys.path.insert(0, 'tools') # Also allow entry-type modules to be brought in from the etype directory. -sys.path.append(os.path.join(our_path, 'etype')) +sys.path.insert(0, os.path.join(our_path, 'etype')) import cmdline import command -- 2.13.1.611.g7e3b11ae1-goog

3 3