From: Lucas Stach <l.stach@pengutronix.de>
To: Andrey Smirnov <andrew.smirnov@gmail.com>, barebox@lists.infradead.org
Subject: Re: [PATCH v2 09/17] of: Port latest of_device_make_bus_id() implementation
Date: Wed, 10 Oct 2018 18:51:59 +0200 [thread overview]
Message-ID: <1539190319.3515.48.camel@pengutronix.de> (raw)
In-Reply-To: <20181009173805.26181-10-andrew.smirnov@gmail.com>
Hi Andrey,
Am Dienstag, den 09.10.2018, 10:37 -0700 schrieb Andrey Smirnov:
> Code implementing of_device_make_bus_id() in Barebox uses rather old
> implementation from Linux kernel and has a very significan limitation
> in that it will produce identical names for different DT nodes that
> happen to have the same node name as well as "reg" property.
>
> > One such example, that tirggered this change, is "switch@0" nodes that
> can be found in dts/src/arm/vf610-zii-dev-rev-c.dts
>
> This commit replaces original code with the current Linux kernel
> implementation that traverses DT hierarchy from leaf node to its root
> concatenating node names in the process.
I think this change is what is breaking the imx_thermal driver here.
This is a bad interaction between the AIODEV framework using the
hwdev->device_node->name as the device name for the aoidev and the
parent device getting the same name with this change. Before this
change the tempmon platform device (which doesn't has an address by
itself) would get a made up bus_no_reg_magic appended.
Due to both devices having the same name, the device registration for
the aiodev now fails. I've fixed it locally by not using the node name
in the AIODEV framework, but you may prefer to fix this differently.
Something to keep in mind for v3 of this series.
Regards,
Lucas
> > Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
> ---
> drivers/of/platform.c | 80 ++++++++++++-------------------------------
> 1 file changed, 22 insertions(+), 58 deletions(-)
>
> diff --git a/drivers/of/platform.c b/drivers/of/platform.c
> index 4fd3ce2b7..17052f419 100644
> --- a/drivers/of/platform.c
> +++ b/drivers/of/platform.c
> @@ -43,71 +43,35 @@ EXPORT_SYMBOL(of_find_device_by_node);
> * of_device_make_bus_id - Use the device node data to assign a unique name
> * @dev: pointer to device structure that is linked to a device tree node
> *
> - * This routine will first try using either the dcr-reg or the reg property
> - * value to derive a unique name. As a last resort it will use the node
> - * name followed by a unique number.
> + * This routine will first try using the translated bus address to
> + * derive a unique name. If it cannot, then it will prepend names from
> + * parent nodes until a unique name can be derived.
> */
> static void of_device_make_bus_id(struct device_d *dev)
> {
> > - static int bus_no_reg_magic;
> > - struct device_node *np = dev->device_node;
> > - const __be32 *reg, *addrp;
> > + struct device_node *node = dev->device_node;
> > + const __be32 *reg;
> > u64 addr;
> > - char *name, *at;
> -
> > - name = xstrdup(np->name);
> > - at = strchr(name, '@');
> > - if (at)
> > - *at = '\0';
> -
> -#ifdef CONFIG_PPC_DCR
> > - /*
> > - * If it's a DCR based device, use 'd' for native DCRs
> > - * and 'D' for MMIO DCRs.
> > - */
> > - reg = of_get_property(np, "dcr-reg", NULL);
> > - if (reg) {
> -#ifdef CONFIG_PPC_DCR_NATIVE
> > - dev_set_name(dev, "d%x.%s", *reg, name);
> -#else /* CONFIG_PPC_DCR_NATIVE */
> > - u64 addr = of_translate_dcr_address(np, *reg, NULL);
> > - if (addr != OF_BAD_ADDR) {
> > - dev_set_name(dev, "D%llx.%s",
> > - (unsigned long long)addr, name);
> > - free(name);
> > - return;
> > - }
> -#endif /* !CONFIG_PPC_DCR_NATIVE */
> > - }
> -#endif /* CONFIG_PPC_DCR */
> -
> > - /*
> > - * For MMIO, get the physical address
> > - */
> > - reg = of_get_property(np, "reg", NULL);
> > - if (reg) {
> > - if (of_can_translate_address(np)) {
> > - addr = of_translate_address(np, reg);
> > - } else {
> > - addrp = of_get_address(np, 0, NULL, NULL);
> > - if (addrp)
> > - addr = of_read_number(addrp, 1);
> > - else
> > - addr = OF_BAD_ADDR;
> > - }
> > - if (addr != OF_BAD_ADDR) {
> > - dev_set_name(dev, "%llx.%s",
> > - (unsigned long long)addr, name);
> > - free(name);
> +
> > + /* Construct the name, using parent nodes if necessary to ensure uniqueness */
> > + while (node->parent) {
> > + /*
> > + * If the address can be translated, then that is as much
> > + * uniqueness as we need. Make it the first component and return
> > + */
> > + reg = of_get_property(node, "reg", NULL);
> > + if (reg && (addr = of_translate_address(node, reg)) != OF_BAD_ADDR) {
> > + dev_set_name(dev, dev->name ? "%llx.%s:%s" : "%llx.%s",
> > + (unsigned long long)addr, node->name,
> > + dev->name);
> > return;
> > }
> > - }
>
> > - /*
> > - * No BusID, use the node name and add a globally incremented counter
> > - */
> > - dev_set_name(dev, "%s.%d", name, bus_no_reg_magic++);
> > - free(name);
> > + /* format arguments only used if dev_name() resolves to NULL */
> > + dev_set_name(dev, dev->name ? "%s:%s" : "%s",
> > + kbasename(node->full_name), dev->name);
> > + node = node->parent;
> > + }
> }
>
> /**
_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox
next prev parent reply other threads:[~2018-10-10 16:52 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-09 17:37 [PATCH v2 00/17] MV88E6xxx switch family support Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 01/17] ARM: Do not expose ARMv8 functions on ARMv7 Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 02/17] clocksource: Add ARM global timer support Andrey Smirnov
2018-10-11 8:28 ` Sascha Hauer
2018-10-09 17:37 ` [PATCH v2 03/17] VFxxx: Select CLOCKSOURCE_ARM_GLOBAL_TIMER Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 04/17] i.MX: Move GPT driver to drivers/clocksource Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 05/17] clocksource: Introduce ARCH_HAS_IMX_GPT Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 06/17] of: Demote "Bad cell count for" to debug Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 07/17] drivers: Introduce dev_set_name() Andrey Smirnov
2018-10-10 2:57 ` Andrey Smirnov
2018-10-11 8:56 ` Sascha Hauer
2018-10-12 2:01 ` Andrey Smirnov
2018-10-10 9:41 ` Lucas Stach
2018-10-10 15:33 ` Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 08/17] linux: string: Port kbasename() Andrey Smirnov
2018-10-11 8:31 ` Sascha Hauer
2018-10-13 20:47 ` Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 09/17] of: Port latest of_device_make_bus_id() implementation Andrey Smirnov
2018-10-10 16:51 ` Lucas Stach [this message]
2018-10-10 16:54 ` Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 10/17] mdio_bus: Fix documentation for mdio_bus_match() Andrey Smirnov
2018-10-09 17:37 ` [PATCH v2 11/17] include: linux: phy: Add missing PHY_INTERFACE_* constants Andrey Smirnov
2018-10-09 17:38 ` [PATCH v2 12/17] include: linux: ethtool: Add missing *_UNKNOWN constants Andrey Smirnov
2018-10-09 17:38 ` [PATCH v2 13/17] net: phy: Check phy_mask in get_phy_device() Andrey Smirnov
2018-10-09 17:38 ` [PATCH v2 14/17] mdio_bus: Allow for non PHY-devices on MDIO buses Andrey Smirnov
2018-10-09 17:38 ` [PATCH v2 15/17] net: phy: Add basic driver for MV88E6XXX switches from Marvell Andrey Smirnov
2018-10-09 17:38 ` [PATCH v2 16/17] net: phy: mv88e6xxx: Port EEPROM support code Andrey Smirnov
2018-10-09 17:38 ` [PATCH v2 17/17] net: phy: mv88e6xxx: Add support for MAC ports Andrey Smirnov
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=1539190319.3515.48.camel@pengutronix.de \
--to=l.stach@pengutronix.de \
--cc=andrew.smirnov@gmail.com \
--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