[PATCH v8 3/8] test: cmd: add unit tests for part command
Patrice CHOTARD
patrice.chotard at foss.st.com
Tue May 12 11:45:49 CEST 2026
On 5/12/26 09:08, Dario Binacchi wrote:
> Hi everyone,
>
> I am resending my previous reply because I received a delivery failure
> notification last week indicating that Patrice's mailbox was full.
>
> Patrice, did you have a chance to verify if the tests are passing fine now?
> Please let me know if everything looks good or if you need
>
> me to make any further modifications.
>
> Thanks and regards
>
> Dario
Hi Dario
My mailbox has been cleaned ;-), i expected to respin this series for test tomorrow or next week.
Thanks
Patrice
>
> On Tue, May 5, 2026 at 9:55 AM Dario Binacchi <
> dario.binacchi at amarulasolutions.com> wrote:
>
>> Hi Patrice,
>>
>> On Tue, May 5, 2026 at 9:26 AM Patrice CHOTARD <
>> patrice.chotard at foss.st.com> wrote:
>>
>>>
>>>
>>> On 4/30/26 10:06, Dario Binacchi wrote:
>>>> Add unit tests for the 'part' command, specifically for the 'number',
>>>> 'start', and 'size' subcommands.
>>>>
>>>> These tests establish a baseline for the current partition lookup
>>>> functionality by name. This foundation will be used by subsequent
>>>> patches to extend the command, ensuring consistent behavior as new
>>>> features are introduced.
>>>>
>>>> Signed-off-by: Dario Binacchi <dario.binacchi at amarulasolutions.com>
>>>> Reviewed-by: Simon Glass <sjg at chromium.org>
>>>>
>>>> ---
>>>>
>>>> Changes in v8:
>>>> - Skip the test in case the CONFIG_MCC is not enabled or the mmc
>>>> device is not found.
>>>>
>>>> Changes in v6:
>>>> - Add Reviewed-by of Simon Glass to patch 3/8 "test: cmd: add unit
>>>> tests for part command"
>>>>
>>>> Changes in v4:
>>>> - Place the part command tests in the cmd suite rather than the dm
>>>> suite.
>>>> - Order the include files according the U-Boot coding style.
>>>>
>>>> Changes in v3:
>>>> - Wrap the lines exceeding 80 columns
>>>> - Combine run_command() and ut_asserteq() -> ut_asserteq(1, run_command(
>>>>
>>>> test/cmd/Makefile | 1 +
>>>> test/cmd/part.c | 183 ++++++++++++++++++++++++++++++++++++++++++++++
>>>> 2 files changed, 184 insertions(+)
>>>> create mode 100644 test/cmd/part.c
>>>>
>>>> diff --git a/test/cmd/Makefile b/test/cmd/Makefile
>>>> index 5f2815b1bb65..8d6932f1176c 100644
>>>> --- a/test/cmd/Makefile
>>>> +++ b/test/cmd/Makefile
>>>> @@ -26,6 +26,7 @@ obj-$(CONFIG_CMD_LOADM) += loadm.o
>>>> obj-$(CONFIG_CMD_MEMINFO) += meminfo.o
>>>> obj-$(CONFIG_CMD_MEMORY) += mem_copy.o
>>>> obj-$(CONFIG_CMD_MEM_SEARCH) += mem_search.o
>>>> +obj-$(CONFIG_CMD_PART) += part.o
>>>> ifdef CONFIG_CMD_PCI
>>>> obj-$(CONFIG_CMD_PCI_MPS) += pci_mps.o
>>>> endif
>>>> diff --git a/test/cmd/part.c b/test/cmd/part.c
>>>> new file mode 100644
>>>> index 000000000000..cac57a1a0e35
>>>> --- /dev/null
>>>> +++ b/test/cmd/part.c
>>>> @@ -0,0 +1,183 @@
>>>> +// SPDX-License-Identifier: GPL-2.0+
>>>> +/*
>>>> + * Test for part command
>>>> + *
>>>> + * Copyright (C) 2026 Amarula Solutions
>>>> + * Written by Dario Binacchi <dario.binacchi at amarulasolutions.com>
>>>> + */
>>>> +
>>>> +#include <command.h>
>>>> +#include <dm.h>
>>>> +#include <env.h>
>>>> +#include <part.h>
>>>> +#include <vsprintf.h>
>>>> +#include <dm/test.h>
>>>> +#include <test/cmd.h>
>>>> +#include <test/test.h>
>>>> +#include <test/ut.h>
>>>> +
>>>> +static struct disk_partition gpt_parts[] = {
>>>> + {
>>>> + .start = 48,
>>>> + .size = 1,
>>>> + .name = "test1",
>>>> + .uuid = "c5bce7a2-03f0-4d03-9048-01ff23b9d527",
>>>> + },
>>>> + {
>>>> + .start = 49,
>>>> + .size = 2,
>>>> + .name = "test2",
>>>> + .uuid = "9df346e8-2c53-4cd8-b9ac-3af83f9a9b74",
>>>> + },
>>>> +};
>>>> +
>>>> +static char disk_guid[UUID_STR_LEN + 1] =
>>>> + "8d60b397-1bb6-4d33-80ee-b1587d24c2f8";
>>>> +
>>>> +static int setup_gpt_partitions(struct unit_test_state *uts,
>>>> + unsigned int mmc_dev_num)
>>>> +{
>>>> + struct blk_desc *mmc_dev_desc;
>>>> + char dev_str[10];
>>>> + int i, ret;
>>>> +
>>>> + if (!CONFIG_IS_ENABLED(MMC))
>>>> + return -EAGAIN;
>>>
>>> Hi Dario
>>>
>>> Is there any reason to return -EAGAIN ? -ENODEV seems more appropriate ?
>>>
>>
>> In the testing framework, returning -EAGAIN explicitly signals that the
>> test should be skipped*.*
>>
>> As highlighted in the documentation for ut_run_test():
>>
>> *Return: 0 if all tests passed, -EAGAIN if the test should be skipped, -1
>> if any failed*
>>
>>
>> Since the necessary preconditions weren't met for some configuration in
>> our case, I decided to skip.
>>
>> Thanks and regards,
>>
>> Dario
>>
>>
>>> Thanks
>>> Patrice
>>>
>>>> +
>>>> + snprintf(dev_str, sizeof(dev_str), "%u", mmc_dev_num);
>>>> +
>>>> + ret = blk_get_device_by_str("mmc", dev_str, &mmc_dev_desc);
>>>> + if (ret == -ENODEV)
>>>> + return -EAGAIN;
>>>> +
>>>> + ut_asserteq(mmc_dev_num, ret);
>>>> +
>>>> + if (CONFIG_IS_ENABLED(RANDOM_UUID)) {
>>>> + for (i = 0; i < ARRAY_SIZE(gpt_parts); i++)
>>>> + gen_rand_uuid_str(gpt_parts[i].uuid,
>>>> + UUID_STR_FORMAT_STD);
>>>> +
>>>> + gen_rand_uuid_str(disk_guid, UUID_STR_FORMAT_STD);
>>>> + }
>>>> +
>>>> + ut_assertok(gpt_restore(mmc_dev_desc, disk_guid, gpt_parts,
>>>> + ARRAY_SIZE(gpt_parts)));
>>>> + return 0;
>>>> +}
>>>> +
>>>> +static int cmd_test_part_number(struct unit_test_state *uts)
>>>> +{
>>>> + unsigned int mmc_dev_num = 2;
>>>> + char expected[10];
>>>> + int i, ret;
>>>> +
>>>> + ret = setup_gpt_partitions(uts, mmc_dev_num);
>>>> + if (ret == -EAGAIN)
>>>> + return ret;
>>>> +
>>>> + ut_assertok(ret);
>>>> +
>>>> + for (i = 0; i < ARRAY_SIZE(gpt_parts); i++) {
>>>> + env_set("partnum", NULL);
>>>> + ut_assertok(run_commandf("part number mmc %u %s partnum",
>>>> + mmc_dev_num, gpt_parts[i].name));
>>>> + snprintf(expected, sizeof(expected), "0x%x", i + 1);
>>>> + ut_asserteq_str(expected, env_get("partnum"));
>>>> + }
>>>> +
>>>> + env_set("partnum", NULL);
>>>> + ut_asserteq(1, run_commandf("part number mmc %u bogus partnum",
>>>> + mmc_dev_num));
>>>> + ut_assertnull(env_get("partnum"));
>>>> +
>>>> + return 0;
>>>> +}
>>>> +CMD_TEST(cmd_test_part_number, UTF_CONSOLE);
>>>> +
>>>> +static int cmd_test_part_start(struct unit_test_state *uts)
>>>> +{
>>>> + unsigned int mmc_dev_num = 2;
>>>> + char expected[32];
>>>> + int i, ret;
>>>> +
>>>> + ret = setup_gpt_partitions(uts, mmc_dev_num);
>>>> + if (ret == -EAGAIN)
>>>> + return ret;
>>>> +
>>>> + ut_assertok(ret);
>>>> +
>>>> + for (i = 0; i < ARRAY_SIZE(gpt_parts); i++) {
>>>> + env_set("partstart", NULL);
>>>> + ut_assertok(run_commandf("part start mmc %u %d partstart",
>>>> + mmc_dev_num, i + 1));
>>>> + snprintf(expected, sizeof(expected), "%lx",
>>>> + (unsigned long)gpt_parts[i].start);
>>>> + ut_asserteq_str(expected, env_get("partstart"));
>>>> + }
>>>> +
>>>> + env_set("partstart", NULL);
>>>> + ut_asserteq(1, run_commandf("part start mmc %u 3 partstart",
>>>> + mmc_dev_num));
>>>> + ut_assertnull(env_get("partstart"));
>>>> +
>>>> + for (i = 0; i < ARRAY_SIZE(gpt_parts); i++) {
>>>> + env_set("partstart", NULL);
>>>> + ut_assertok(run_commandf("part start mmc %u %s partstart",
>>>> + mmc_dev_num, gpt_parts[i].name));
>>>> + snprintf(expected, sizeof(expected), "%lx",
>>>> + (unsigned long)gpt_parts[i].start);
>>>> + ut_asserteq_str(expected, env_get("partstart"));
>>>> + }
>>>> +
>>>> + env_set("partstart", NULL);
>>>> + ut_asserteq(1, run_commandf("part start mmc %u bogus partstart",
>>>> + mmc_dev_num));
>>>> + ut_assertnull(env_get("partstart"));
>>>> +
>>>> + return 0;
>>>> +}
>>>> +CMD_TEST(cmd_test_part_start, UTF_CONSOLE);
>>>> +
>>>> +static int cmd_test_part_size(struct unit_test_state *uts)
>>>> +{
>>>> + unsigned int mmc_dev_num = 2;
>>>> + char expected[32];
>>>> + int i, ret;
>>>> +
>>>> + ret = setup_gpt_partitions(uts, mmc_dev_num);
>>>> + if (ret == -EAGAIN)
>>>> + return ret;
>>>> +
>>>> + ut_assertok(ret);
>>>> +
>>>> + for (i = 0; i < ARRAY_SIZE(gpt_parts); i++) {
>>>> + env_set("partsize", NULL);
>>>> + ut_assertok(run_commandf("part size mmc %u %d partsize",
>>>> + mmc_dev_num, i + 1));
>>>> + snprintf(expected, sizeof(expected), "%lx",
>>>> + (unsigned long)gpt_parts[i].size);
>>>> + ut_asserteq_str(expected, env_get("partsize"));
>>>> + }
>>>> +
>>>> + env_set("partsize", NULL);
>>>> + ut_asserteq(1, run_commandf("part size mmc %u 3 partsize",
>>>> + mmc_dev_num));
>>>> + ut_assertnull(env_get("partsize"));
>>>> +
>>>> + for (i = 0; i < ARRAY_SIZE(gpt_parts); i++) {
>>>> + env_set("partsize", NULL);
>>>> + ut_assertok(run_commandf("part size mmc %u %s partsize",
>>>> + mmc_dev_num, gpt_parts[i].name));
>>>> + snprintf(expected, sizeof(expected), "%lx",
>>>> + (unsigned long)gpt_parts[i].size);
>>>> + ut_asserteq_str(expected, env_get("partsize"));
>>>> + }
>>>> +
>>>> + env_set("partsize", NULL);
>>>> + ut_asserteq(1, run_commandf("part size mmc %u bogus partsize",
>>>> + mmc_dev_num));
>>>> + ut_assertnull(env_get("partsize"));
>>>> +
>>>> + return 0;
>>>> +}
>>>> +CMD_TEST(cmd_test_part_size, UTF_CONSOLE);
>>>
>>>
>>
>> --
>> *Dario Binacchi*
>> Senior Embedded Software Engineer
>> M. +39 328 0625246
>> dario.binacchi at amarulasolutions.com
>> ―――――――――――――――
>> Amarula Solutions SRL
>> Via Felice Cavallotti 25D, 41012 Carpi, MO, IT
>> info at amarulasolutions.com
>> www.amarulasolutions.com
>>
>
>
More information about the U-Boot
mailing list