From: Bastian Krause <bst@pengutronix.de>
To: Ahmad Fatoum <a.fatoum@pengutronix.de>, barebox@lists.infradead.org
Subject: Re: [PATCH 3/5] common: machine_id: deprecate machine_id_set_hashable
Date: Mon, 28 Jun 2021 11:50:34 +0200 [thread overview]
Message-ID: <bea1987e-8336-6593-0666-e263f4078c58@pengutronix.de> (raw)
In-Reply-To: <20210628064036.25991-3-a.fatoum@pengutronix.de>
On 6/28/21 8:40 AM, Ahmad Fatoum wrote:
> The Kconfig symbol already warns users that barebox updates
> that add new instances of machine_id_set_hashable may cause the machine
> ID to change making the future less useful. Recent changes allow the
> board DT to identify a specific nvmem cell to use for supplying a machine
> ID, making the old way of drivers providing machine IDs and possibly
> overriding each other no longer necessary or recommended.
I cannot parse the first sentence of the commit message.
Regards,
Bastian
>
> Make this explicit by allowing the old code to be disabled. As we only
> have a single user now and won't accept any new ones, we can remove the
> warning about it possibly changing after update.
>
> Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
> ---
> common/Kconfig | 18 +++++++++++-------
> common/machine_id.c | 11 ++++++++---
> drivers/nvmem/Kconfig | 1 +
> drivers/nvmem/ocotp.c | 7 ++++++-
> include/machine_id.h | 16 ----------------
> 5 files changed, 26 insertions(+), 27 deletions(-)
> delete mode 100644 include/machine_id.h
>
> diff --git a/common/Kconfig b/common/Kconfig
> index a4a109f04f08..a2aa0b2568de 100644
> --- a/common/Kconfig
> +++ b/common/Kconfig
> @@ -1069,13 +1069,11 @@ config MACHINE_ID
> select NVMEM
> help
> Compute a persistent machine-specific id and store it to $global.machine_id.
> - The id is a hash of device-specific information added either via
> - machine_id_set_hashable() or by hashing the nvmem cell referenced by the
> - /chosen/barebox,machine-id device tree property.
> -
> - With machine_id_set_hashable(), the last call prior to the late initcall
> - set_machine_id() willl be used to generate the machine id. This means
> - updating barebox may change the machine id!
> + The id is a hash of device-specific information. This information
> + comes from the nvmem cell device tree node path described by the
> + /chosen/barebox,machine-id-path property. As a special case, the i.MX6 OCOTP
> + driver supplies the SoC UID as hashable for when /chosen/barebox,machine-id-path
> + is not specified.
>
> global.bootm.provide_machine_id may be used to automatically set
> the linux.bootargs.machine_id global variable with a value of
> @@ -1084,6 +1082,12 @@ config MACHINE_ID
> Note: if no hashable information is available no machine id will be passed
> to the kernel.
>
> +config MACHINE_ID_LEGACY
> + bool
> + help
> + Selected by i.MX OCOTP driver, so it can set the SoC UID as hashable.
> + New platforms should use /chosen/barebox,machine-id-path instead.
> +
> config SYSTEMD_OF_WATCHDOG
> bool "inform devicetree-enabled kernel of used watchdog"
> depends on WATCHDOG && OFTREE && FLEXIBLE_BOOTARGS
> diff --git a/common/machine_id.c b/common/machine_id.c
> index fd2f0888a6cd..8303c0b7aa87 100644
> --- a/common/machine_id.c
> +++ b/common/machine_id.c
> @@ -9,23 +9,24 @@
> #include <globalvar.h>
> #include <magicvar.h>
> #include <crypto/sha.h>
> -#include <machine_id.h>
> #include <linux/err.h>
> #include <linux/nvmem-consumer.h>
> #include <of.h>
>
> #define MACHINE_ID_LENGTH 32
>
> +#ifdef CONFIG_MACHINE_ID_LEGACY
> static void *__machine_id_hashable;
> static size_t __machine_id_hashable_length;
>
> -
> +/* Shouldn't be called from new code */
> +void machine_id_set_hashable(const void *hashable, size_t len);
> void machine_id_set_hashable(const void *hashable, size_t len)
> {
> -
> __machine_id_hashable = xmemdup(hashable, len);
> __machine_id_hashable_length = len;
> }
> +#endif
>
> static const void *machine_id_get_hashable(size_t *len)
> {
> @@ -52,8 +53,12 @@ static const void *machine_id_get_hashable(size_t *len)
> return ret;
>
> no_cell:
> +#ifdef CONFIG_MACHINE_ID_LEGACY
> *len = __machine_id_hashable_length;
> return __machine_id_hashable;
> +#else
> + return NULL;
> +#endif
> }
>
> static int machine_id_set_globalvar(void)
> diff --git a/drivers/nvmem/Kconfig b/drivers/nvmem/Kconfig
> index 3781f7a839fc..08a5765573a8 100644
> --- a/drivers/nvmem/Kconfig
> +++ b/drivers/nvmem/Kconfig
> @@ -25,6 +25,7 @@ config NVMEM_SNVS_LPGPR
> config IMX_OCOTP
> tristate "i.MX6 On Chip OTP controller"
> depends on ARCH_IMX6 || ARCH_VF610 || ARCH_IMX8M || ARCH_IMX7
> + select MACHINE_ID_LEGACY
> depends on OFDEVICE
> help
> This adds support for the i.MX6 On-Chip OTP controller. Currently the
> diff --git a/drivers/nvmem/ocotp.c b/drivers/nvmem/ocotp.c
> index b2fad3c68770..0b10e52a86ba 100644
> --- a/drivers/nvmem/ocotp.c
> +++ b/drivers/nvmem/ocotp.c
> @@ -29,7 +29,6 @@
> #include <regmap.h>
> #include <linux/clk.h>
> #include <mach/ocotp.h>
> -#include <machine_id.h>
> #include <mach/ocotp-fusemap.h>
> #include <linux/nvmem-provider.h>
>
> @@ -689,6 +688,7 @@ static int imx_ocotp_read(void *ctx, unsigned offset, void *val, size_t bytes)
>
> static void imx_ocotp_set_unique_machine_id(void)
> {
> + extern void machine_id_set_hashable(const void *hashable, size_t len);
> uint32_t unique_id_parts[UNIQUE_ID_NUM];
> int i;
>
> @@ -785,6 +785,11 @@ static int imx_ocotp_probe(struct device_d *dev)
> ethaddr->value, ethaddr);
> }
>
> + /* Special case: new machine id providers should provide a nvmem cell
> + * and reference its path via /chosen/barebox,machine-id-path.
> + * For i.MX, we support the legacy way of the driver supplying the
> + * hash instead
> + */
> if (IS_ENABLED(CONFIG_MACHINE_ID))
> imx_ocotp_set_unique_machine_id();
>
> diff --git a/include/machine_id.h b/include/machine_id.h
> deleted file mode 100644
> index 31d5e0bb2851..000000000000
> --- a/include/machine_id.h
> +++ /dev/null
> @@ -1,16 +0,0 @@
> -#ifndef __MACHINE_ID_H__
> -#define __MACHINE_ID_H__
> -
> -#if IS_ENABLED(CONFIG_MACHINE_ID)
> -
> -void machine_id_set_hashable(const void *hashable, size_t len);
> -
> -#else
> -
> -static inline void machine_id_set_hashable(const void *hashable, size_t len)
> -{
> -}
> -
> -#endif /* CONFIG_MACHINE_ID */
> -
> -#endif /* __MACHINE_ID_H__ */
>
--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox
next prev parent reply other threads:[~2021-06-28 9:52 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-28 6:40 [PATCH 1/5] common: machine_id: support /chosen/barebox, machine-id-path override Ahmad Fatoum
2021-06-28 6:40 ` [PATCH 2/5] ARM: stm32mp: migrate to barebox,machine-id-path Ahmad Fatoum
2021-06-28 6:40 ` [PATCH 3/5] common: machine_id: deprecate machine_id_set_hashable Ahmad Fatoum
2021-06-28 9:50 ` Bastian Krause [this message]
2021-06-28 10:12 ` Ahmad Fatoum
2021-06-28 6:40 ` [PATCH 4/5] sandbox: dts: populate $global.machine_id Ahmad Fatoum
2021-06-28 6:40 ` [PATCH 5/5] ARM: dts: stm32mp: retire barebox, provide-mac-address in favor of NVMEM Ahmad Fatoum
2021-06-28 9:28 ` [PATCH 1/5] common: machine_id: support /chosen/barebox,machine-id-path override Ahmad Fatoum
2021-06-28 9:35 ` [PATCH 1/5] common: machine_id: support /chosen/barebox, machine-id-path override Bastian Krause
2021-06-28 10:11 ` Ahmad Fatoum
2021-06-28 20:20 ` Sascha Hauer
[not found] ` <CAMHeXxOT__KBUKG6GkNAEkqz4tMBBzuZ7OgnKa0_OX5hz-JEig@mail.gmail.com>
2021-06-30 10:27 ` Ahmad Fatoum
2021-06-30 20:13 ` Trent Piepho
2021-09-15 10:55 ` Ahmad Fatoum
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=bea1987e-8336-6593-0666-e263f4078c58@pengutronix.de \
--to=bst@pengutronix.de \
--cc=a.fatoum@pengutronix.de \
--cc=barebox@lists.infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox