From: Guenter Roeck <linux@roeck-us.net>
To: Daniel Walker <danielwa@cisco.com>
Cc: "xe-kernel@external.cisco.com" <xe-kernel@external.cisco.com>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Jean Delvare <khali@linux-fr.org>,
lm-sensors@lm-sensors.org, linux-kernel@vger.kernel.org
Subject: Re: unstable /sys interface from hmon
Date: Mon, 30 Jan 2017 15:11:49 -0800 [thread overview]
Message-ID: <20170130231149.GB14987@roeck-us.net> (raw)
In-Reply-To: <99aae36e-ddfd-dcd2-e544-9c28ae373959@cisco.com>
On Mon, Jan 30, 2017 at 12:37:14PM -0800, Daniel Walker wrote:
> Hi,
>
> Cisco recently upgraded to kernel 4.4.. We have applications which use the /sys interfaces for some hmon drivers.
> We found that the following patches,
>
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bab2243ce1897865e31ea6d59b0478391f51812b
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c3b7cddc70075d525db6b3068d8b2b9158eedc84
>
> Altered the interface as seen from userspace. These patches changed the paths in the following way,
>
> Before,
> /sys/bus/i2c/devices/14-0050/
>
> After,
> /sys/bus/i2c/devices/14-0050/hwmon/hwmon7
>
Actually, the stable link to use would be /sys/class/hwmon/hwmon7
in both cases, and to check if the name attribute resides in hwmon7
or in hwmon7/device (which is the real change in your case - it moved
from hwmon7/device to hwmon7, ie from the parent device to the hwmon
device).
> My impression was that /sys was suppose to be stable, and debugfs was unstable. I'm not sure what the current guidelines are
> for this, but we would request that hmon drivers maintain a stable interface from /sys. This change has cause upgrade issues
> for us which we would like to avoid in the future.
>
The key transition is the move from attaching the attributes to the parent
device (here: the i2c device) to the hwmon device. libsensors handles this
transition transparently; I assume you are not using it ?
> A patch to add in a compatibility layer would be welcome.
>
A compatibility layer would have to link each attribute back to the parent
device. Interesting idea, actually. I'll check if that is possible.
Thanks,
Guenter
prev parent reply other threads:[~2017-01-30 23:12 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-01-30 20:37 Daniel Walker
2017-01-30 23:11 ` Guenter Roeck [this message]
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=20170130231149.GB14987@roeck-us.net \
--to=linux@roeck-us.net \
--cc=danielwa@cisco.com \
--cc=gregkh@linuxfoundation.org \
--cc=khali@linux-fr.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lm-sensors@lm-sensors.org \
--cc=xe-kernel@external.cisco.com \
/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
all inboxes | Powered by JetHome®