November 2020 - U-Boot

[PATCH] binman: Fix typo in kernel-doc format for binman_symname()
by Michal Simek 04 Nov '20

04 Nov '20

Fix typo. Fixes: 19790632648b ("binman: Support accessing binman tables at run time") Signed-off-by: Michal Simek <michal.simek(a)xilinx.com> --- include/binman_sym.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/binman_sym.h b/include/binman_sym.h index ffb7c15798a7..72e6765fe520 100644 --- a/include/binman_sym.h +++ b/include/binman_sym.h @@ -16,7 +16,7 @@ #ifdef CONFIG_BINMAN /** - * binman_symname() - Internal fnuction to get a binman symbol name + * binman_symname() - Internal function to get a binman symbol name * * @entry_name: Name of the entry to look for (e.g. 'u_boot_spl') * @_prop_name: Property value to get from that entry (e.g. 'pos') -- 2.28.0

3 5

[PATCH v3 0/8] sandbox: implement cold reset
by Heinrich Schuchardt 04 Nov '20

04 Nov '20

The command to shut down a device is 'poweroff'. It is a deficit of the sandbox that it does not support resetting yet but shuts down upong seeing the 'reset' command. The patch series implements the cold reset function as a relaunch of the U-Boot binary via execv(). Unit tests are adjusted. A bug leading to closing the console input file descriptor is resolved. Add some of the sandbox API to the HTML documentation. v3: update commit message for patch 5 v2: use O_CLOEXEC in os_open() avoid longjmp() add more comments add the os.h to the HTML documentation Heinrich Schuchardt (8): sandbox: eth-raw: do not close the console input sandbox: enable poweroff command test/py: test poweroff sandbox: use O_CLOEXEC in os_open() sandbox: implement reset test: adjust sysreset tests sandbox: update function descriptions in os.h doc: add sandbox API arch/Kconfig | 3 +- arch/sandbox/cpu/eth-raw-os.c | 8 +- arch/sandbox/cpu/os.c | 16 +- arch/sandbox/cpu/start.c | 26 +++ arch/sandbox/cpu/state.c | 1 + arch/sandbox/include/asm/u-boot-sandbox.h | 10 + doc/api/index.rst | 1 + doc/api/sandbox.rst | 9 + drivers/sysreset/sysreset_sandbox.c | 3 + include/os.h | 236 +++++++++++++--------- test/dm/sysreset.c | 11 +- test/py/tests/test_sandbox_exit.py | 8 +- 12 files changed, 227 insertions(+), 105 deletions(-) create mode 100644 doc/api/sandbox.rst -- 2.28.0

2 22

[PATCH v3 00/29] patman: Collect review tags and comments from Patchwork
by Simon Glass 04 Nov '20

04 Nov '20

Patman is a useful tool for creating, checking sending out patches. It automates the creation of patches, simplifies checking them and handles cover letters and change logs. However once patches are sent and reviewers add Reviewed-by tags, etc., these must be manually added into the commits using git before the next version of the series is sent. Also, the only way to look at patches is in the web interface. It would be nice if patman could talk to Patchwork and collect the review tags itself. Even better if it could show the review comments in a command-line view patch-by-patch to speed up addressing comments. This series adds these features to patman, talking directly to the Patchwork REST API. While pwclient seems like it could handle some of this, or at least provide a library to patman, the documentation[1] doesn't really explain what the commands do and it doesn't seem to work with the current Patchwork (e.g. pwclient git-am requires a patch number but Patchwork seems to use an ID now). With these changes it is possible to type 'patman status' and see a list of new review tags. It is also possible to create a new branch with those tags and list out the review comments as small snippets on the command line. I have been using these features for a little while and they seem useful, so herewith a series. Some of the patches here clean up style problems in patman and other minor problems. See the last four patches for the main changes. You can find it at u-boot-dm/patman-working[2] [1] https://patchwork.readthedocs.io/projects/pwclient/en/latest/usage/ [2] https://gitlab.denx.de/u-boot/custodians/u-boot-dm/-/tree/patman-working Changes in v3: - Rename GetMetaData() function to snake case - Fix incorrect commenting on a line in prepare_patches() - Allow tags to be inserted in the middle of the commit message Changes in v2: - Adjust the algorithm to handle patch/commit mismatch - Correct Python style problems - Use REST API instead of web pages - Many changes to the whole series, including new patches - Use of REST API instead of the web site Simon Glass (29): patman: Correct operation of -n azure/gitLab/travis: Add pygit2 as a dependency for tests patman: Update how tests are run patman: Fix whitespace errors in func_test patman: Use capture_sys_output() consistently patman: Fix remaining pylint3 warnings in func_test patman: Allow linking a series with patchwork patman: Fix indenting in patchstream patman: Fix constant style in patchstream patman: Rename functions in patchstream patman: Rename variables in patchstream patman: Drop unused args in patchstream patman: Fix up argument/return docs in patchstream patman: Move warning collection to a function patman: Attach warnings to individual patches patman: Convert 'Series-xxx' tag errors into warnings patman: Drop unused signoff member patman: Add a test for PatchStream tags patman: Add some tests for warnings patman: Convert testBasic() to use an interator patman: Fix spelling of plural for warning patman: Don't ignore lines starting with hash patman: Allow showing a Commit as a string patman: Improve handling of files patman: Detect missing upstream in CountCommitsToBranch patman: Support checking for review tags in patchwork patman: Support updating a branch with review tags patman: Support parsing of review snippets patman: Support listing comments from patchwork .azure-pipelines.yml | 2 +- .gitlab-ci.yml | 2 +- .travis.yml | 1 + tools/buildman/control.py | 6 +- tools/patman/README | 110 +++- tools/patman/checkpatch.py | 18 +- tools/patman/commit.py | 5 + tools/patman/control.py | 71 ++- tools/patman/func_test.py | 983 +++++++++++++++++++++++++++++--- tools/patman/gitutil.py | 10 +- tools/patman/main.py | 42 +- tools/patman/patchstream.py | 485 +++++++++++----- tools/patman/series.py | 8 +- tools/patman/status.py | 482 ++++++++++++++++ tools/patman/terminal.py | 21 +- tools/patman/test_checkpatch.py | 6 +- 16 files changed, 1981 insertions(+), 271 deletions(-) create mode 100644 tools/patman/status.py -- 2.29.1.341.ge80a0c044ae-goog

1 58

[PATCH 1/1] test: test/bloblist.c depends on asm/state.h
by Heinrich Schuchardt 04 Nov '20

04 Nov '20

Building test/bloblist.c fails for non sandbox devices: test/bloblist.c:10:10: fatal error: asm/state.h: No such file or directory #include <asm/state.h> ^~~~~~~~~~~~~ Build the test only on the sandbox. Signed-off-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> --- test/Makefile | 2 ++ 1 file changed, 2 insertions(+) diff --git a/test/Makefile b/test/Makefile index 1c930b3148..39ae04a3d2 100644 --- a/test/Makefile +++ b/test/Makefile @@ -2,7 +2,9 @@ # # (C) Copyright 2012 The Chromium Authors +ifneq ($(CONFIG_SANDBOX),) obj-$(CONFIG_$(SPL_)CMDLINE) += bloblist.o +endif obj-$(CONFIG_$(SPL_)CMDLINE) += cmd/ obj-$(CONFIG_$(SPL_)CMDLINE) += cmd_ut.o obj-$(CONFIG_$(SPL_)CMDLINE) += command_ut.o -- 2.20.1

2 2

[PATCH 1/1] test: linking test/compression.c fails
by Heinrich Schuchardt 04 Nov '20

04 Nov '20

Building U-Boot with unit tests on a non-sandbox systems fails: ld.bfd: test/built-in.o: in function `compress_using_gzip': test/compression.c:138: undefined reference to `gzip' ld.bfd: test/built-in.o: in function `uncompress_using_bzip2': test/compression.c:187: undefined reference to `BZ2_bzBuffToBuffDecompress' ld.bfd: test/built-in.o: in function `uncompress_using_lzma': test/compression.c:222: undefined reference to `lzmaBuffToBuffDecompress' ld.bfd: test/built-in.o: in function `uncompress_using_lzo': test/compression.c:257: undefined reference to `lzop_decompress' ld.bfd: test/built-in.o: in function `uncompress_using_lz4': test/compression.c:292: undefined reference to `ulz4fn Add the missing dependencies. Signed-off-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> --- test/Kconfig | 9 +++++++++ test/Makefile | 2 +- 2 files changed, 10 insertions(+), 1 deletion(-) diff --git a/test/Kconfig b/test/Kconfig index 2646e7d825..ab3ac54a1b 100644 --- a/test/Kconfig +++ b/test/Kconfig @@ -50,6 +50,15 @@ config UT_LIB_RSA endif +config UT_COMPRESSION + bool "Unit test for compression" + depends on UNIT_TEST + depends on CMDLINE && GZIP_COMPRESSED && BZIP2 && LZMA && LZO && LZ4 + default y + help + Enables tests for compression and decompression routines for simple + sanity and for buffer overflow conditions. + config UT_LOG bool "Unit tests for logging functions" depends on UNIT_TEST diff --git a/test/Makefile b/test/Makefile index 39ae04a3d2..8296734eb3 100644 --- a/test/Makefile +++ b/test/Makefile @@ -8,7 +8,7 @@ endif obj-$(CONFIG_$(SPL_)CMDLINE) += cmd/ obj-$(CONFIG_$(SPL_)CMDLINE) += cmd_ut.o obj-$(CONFIG_$(SPL_)CMDLINE) += command_ut.o -obj-$(CONFIG_$(SPL_)CMDLINE) += compression.o +obj-$(CONFIG_$(SPL_)UT_COMPRESSION) += compression.o obj-y += dm/ obj-$(CONFIG_$(SPL_)CMDLINE) += print_ut.o obj-$(CONFIG_$(SPL_)CMDLINE) += str_ut.o -- 2.20.1

2 2

[PATCH 1/2] binman: Update intel_ifwi to store padded section
by Simon Glass 04 Nov '20

04 Nov '20

With a recent change this entry stores only part of the section data, leaving out the padding at the end. Fix this by using GetPaddedData() to get the data. Add this function to the base Entry class also. Fixes: d1d3ad7d1fe ("binman: Move section padding to the parent") Signed-off-by: Simon Glass <sjg(a)chromium.org> --- tools/binman/entry.py | 16 ++++++++++++++++ tools/binman/etype/intel_ifwi.py | 2 +- 2 files changed, 17 insertions(+), 1 deletion(-) diff --git a/tools/binman/entry.py b/tools/binman/entry.py index 8946d2bc02f..68e694e7d15 100644 --- a/tools/binman/entry.py +++ b/tools/binman/entry.py @@ -456,6 +456,22 @@ class Entry(object): self.Detail('GetData: size %s' % ToHexSize(self.data)) return self.data + def GetPaddedData(self, data=None): + """Get the data for an entry including any padding + + Gets the entry data and uses its section's pad-byte value to add padding + before and after as defined by the pad-before and pad-after properties. + + This does not consider alignment. + + Returns: + Contents of the entry along with any pad bytes before and + after it (bytes) + """ + if data is None: + data = self.GetData() + return self.section.GetPaddedDataForEntry(self, data) + def GetOffsets(self): """Get the offsets for siblings diff --git a/tools/binman/etype/intel_ifwi.py b/tools/binman/etype/intel_ifwi.py index 76b3357c252..1a0e481c198 100644 --- a/tools/binman/etype/intel_ifwi.py +++ b/tools/binman/etype/intel_ifwi.py @@ -71,7 +71,7 @@ class Entry_intel_ifwi(Entry_blob_ext): for entry in self._ifwi_entries.values(): # First get the input data and put it in a file - data = entry.GetData() + data = entry.GetPaddedData() uniq = self.GetUniqueName() input_fname = tools.GetOutputFilename('input.%s' % uniq) tools.WriteFile(input_fname, data) -- 2.29.1.341.ge80a0c044ae-goog

1 3

[PATCH] mmc: atmel-sdhci: fix the clk_enable call in case of no ops
by Eugen Hristev 03 Nov '20

03 Nov '20

If the clock driver does not offer a clk_enable ops, then the system will return -ENOSYS. The clk_enable works with CCF (common clock framework). Some clocks in some cases (like the generic clock for some products: sama5d2) do not have the clk_enable primitive, and in this case probing of the driver will fail. This patch changes the behavior to return an error in case there is really an error, and not a missing primitive. If the clock driver does not have an enable primitive, most likely clocks are always enabled or enabled in the set_rate primitives. Fixes: 81f16438d4 ("mmc: atmel-sdhci: enable the required generic clock") Signed-off-by: Eugen Hristev <eugen.hristev(a)microchip.com> --- drivers/mmc/atmel_sdhci.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/mmc/atmel_sdhci.c b/drivers/mmc/atmel_sdhci.c index f56ae63bc2..82dfc345c0 100644 --- a/drivers/mmc/atmel_sdhci.c +++ b/drivers/mmc/atmel_sdhci.c @@ -86,7 +86,8 @@ static int atmel_sdhci_probe(struct udevice *dev) return -EINVAL; ret = clk_enable(&clk); - if (ret) + /* return error only if the clock really has a clock enable func */ + if (ret != -ENOSYS) return ret; ret = mmc_of_parse(dev, &plat->cfg); -- 2.25.1

2 1

[PATCH] env: sf: fix init function behaviour
by Heiko Schocher 03 Nov '20

03 Nov '20

Michael wrote: commit 92765f45bb95 ("env: Access Environment in SPI flashes before relocation") at least breaks the Kontron sl28 board. I guess it also breaks others which use a (late) SPI environment. reason is, that env_init() sets the init bit, if there is no init function defined in an environment driver, and use default return value -ENOENT in this case later for setting the default environment. Change: Environment driver can now implement an init function and return, if this function does nothing, simply -ENOENT. env_init() now handles -ENOENT correct by setting the inited bit for the environment driver. And if there is no other environment driver whose init function returns 0, load than the default environment. This prevents that each environment driver needs to set the default environment. Fixes: 92765f45bb95 ("env: Access Environment in SPI flashes before relocation") Reported-by: Michael Walle <michael(a)walle.cc> Signed-off-by: Heiko Schocher <hs(a)denx.de> --- travis build: https://travis-ci.org/github/hsdenx/u-boot-test/builds/741049813 env/env.c | 2 ++ env/sf.c | 8 +++++--- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/env/env.c b/env/env.c index 42c7d8155e..37b4b54cb7 100644 --- a/env/env.c +++ b/env/env.c @@ -329,6 +329,8 @@ int env_init(void) for (prio = 0; (drv = env_driver_lookup(ENVOP_INIT, prio)); prio++) { if (!drv->init || !(ret = drv->init())) env_set_inited(drv->location); + if (ret == -ENOENT) + env_set_inited(drv->location); debug("%s: Environment %s init done (ret=%d)\n", __func__, drv->name, ret); diff --git a/env/sf.c b/env/sf.c index 2eb2de1a4e..91ed2860ed 100644 --- a/env/sf.c +++ b/env/sf.c @@ -393,10 +393,12 @@ static int env_sf_init(void) return env_sf_init_early(); #endif /* - * we must return with 0 if there is nothing done, - * else env_set_inited() get not called in env_init() + * return here -ENOENT, so env_init() + * can set the init bit and later if no + * other Environment storage is defined + * can set the default environment */ - return 0; + return -ENOENT; } U_BOOT_ENV_LOCATION(sf) = { -- 2.25.4

3 2

Speed/mode setting via "sf probe" command
by Ralph Siemsen 03 Nov '20

03 Nov '20

Hi, The "sf probe" command is documented to take optional speed/mode args: sf probe [[bus:]cs] [hz] [mode] - init flash device on given SPI bus and chip select This worked correctly in older u-boot versions, but as of 2019.07 the speed/mode arguments appear to be effectively ignored. As this feature is quite useful for testing purposes, I would like to find a way to support it again. But the logic is surprisingly complicated, so I would like to get some advice/ideas. The user-specified speed/mode are parsed in do_spi_flash_probe() and eventually are passed to spi_get_bus_and_cs(). In 2019.04 and earlier, the logic here was pretty straightforward: plat = dev_get_parent_platdata(dev); if (!speed) { speed = plat->max_hz; mode = plat->mode; } ret = spi_set_speed_mode(bus, speed, mode); if (ret) goto err; So this calls spi_set_speed_mode() with the user-specified speed, or a default value. As of commit 0cc1b846fcb310c0ac2f8cbeb4ed5947dc52912 ("dm: spi: Read default speed and mode values from DT") the logic has changed. The user-specified speed value is now stored into plat->max_hz, but *only* when no chip select has been configured. In practice this means only the first call to spi_get_bus_and_cs() uses the speed parameter. Thereafter the speed will not change. Related commit f7dd5370986087af9b9cfa601f34b344ec910b87 ("spi: prevent overriding established bus settings") removes the call to spi_set_speed_mode() entirely. So the speed is now set by dm_spi_claim_bus() which uses slave->max_hz, which in turn is set from plat->max_hz in the spi_child_pre_probe() function. The first call to spi_get_bus_and_cs() typically occurs when searching for u-boot environment, and uses the speed specified in DT. This becomes the only speed, as subsequent "sf probe" do not update plat->max_hz. One potential work-around would be to clear the chip select prior to re-binding the device. This allows plat->max_hz to be updated, andalso means that device_bind_driver() will be repeateed. However I am not entirely certain if this is correct approach. In particular, is using -1 for the cs appropriate in all cases? It seems it can come from DT. diff --git a/cmd/sf.c b/cmd/sf.c index d18f6a888c..8cb70f6487 100644 --- a/cmd/sf.c +++ b/cmd/sf.c @@ -128,6 +128,9 @@ static int do_spi_flash_probe(int argc, char *const argv[]) /* Remove the old device, otherwise probe will just be a nop */ ret = spi_find_bus_and_cs(bus, cs, &bus_dev, &new); if (!ret) { + struct dm_spi_slave_platdata *plat; + plat = dev_get_parent_platdata(new); + plat->cs = -1; device_remove(new, DM_REMOVE_NORMAL); } flash = NULL;

2 2

test/Kconfig
by Heinrich Schuchardt 03 Nov '20

03 Nov '20

Hello Simon, since your patch ba96be48ad34180debcfbc11434be7329d530701 dm: test: Build tests for SPL test/Kconfig is a bit in disarray. menuconfig UNIT_TEST marks a submenu. But there is no submenu anymore. Instead all items that used to be in the submenu are now shown on the same indentation level as UNIT_TEST. I would prefer if we had a submenu for UNIT_TEST as we had before and a separate submenu for SPL_UNIT_TEST. In the Makefile we could replace the remaining UT_<OPTION> by UT_$(SPL_)<OPTION>. This would allow us to independently select the tests built for SPL and for main U-Boot. Would this make sense for you? Best regards Heinrich

2 1