Witherspoon: phosphor-hwmon-readd: terminate called after throwing an instance of 'std::filesystem::__cxx11::filesystem_error'

[gkeishin]

Build: https://openpower.xyz/job/openbmc-build/2015/

Version: "2.7.0-dev-90-gae29d00"

Mar 15 09:47:51 witherspoon phosphor-hwmon-readd[1378]: terminate called after throwing an instance of 'std::filesystem::__cxx11::filesystem_error'
Mar 15 09:47:51 witherspoon phosphor-hwmon-readd[1378]:   what():  filesystem error: recursive directory iterator cannot open directory: No such file or directory [/ahb/apb/i2c@1e78a000/i2c-bus@380/tmp275@4a]

Mar 15 10:27:04 witherspoon phosphor-hwmon-readd[2522]: Unable to find hwmon directory from the dev path
Mar 15 10:27:04 witherspoon phosphor-hwmon-readd[2522]: Usage: phosphor-hwmon-readd [options]
Mar 15 10:27:04 witherspoon phosphor-hwmon-readd[2522]: Options:
Mar 15 10:27:04 witherspoon phosphor-hwmon-readd[2522]:     --help               print this menu
Mar 15 10:27:04 witherspoon phosphor-hwmon-readd[2522]:     --path=<path>        sysfs location to monitor
Mar 15 10:27:04 witherspoon phosphor-hwmon-readd[2522]:     --dev-path=<path>    device path to monitor
Mar 15 10:27:04 witherspoon phosphor-hwmon-readd[2522]: ERROR: Path not specified or invalid.

[gkeishin]

I see this on other systems too.. this is pervassive

[spinler]

The only hwmon change lately was openbmc/phosphor-hwmon@9e997b4 - "use filesystem instead of experimental". I wonder if std::filesystem behaves different than the experimental::filesystem.
FYI @pstrinkle

[pstrinkle]

I'll take a look this morning.

On Fri, Mar 15, 2019 at 07:05 Matt Spinler ***@***.***> wrote: The only hwmon change lately was ***@***.*** <openbmc/phosphor-hwmon@9e997b4> - "use filesystem instead of experimental". I wonder if std::filesystem behaves different than the experimental::filesystem. FYI @pstrinkle <https://github.com/pstrinkle> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3499 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABv02PdmDvnQA5Jj0atfXYX_eAsGS6Msks5vW6iegaJpZM4b2Pwo> .

-- "If computers get too powerful, we can organize them into a committee--that will do them in." - Bradley's Bromide http://www.xkcd.com/488/ -- end DRM http://www.xkcd.com/175/ http://xkcd.com/303/

[spinler]

though that directory it's looking at clearly doesn't exist: /ahb/apb/i2c@1e78a000/i2c-bus@380/tmp275@4a

[spinler]

Turns out the difference is in the /= operator.

From cpppreference.com, only for std::filesystem:

path& operator/=(const path& p);
1) If p.is_absolute() || (p.has_root_name() && p.root_name() != root_name()), then replaces the current path with p as if by operator=(p) and finishes.

It doesn't say that for experimental::filesystem.

that breaks this line: https://github.com/openbmc/phosphor-hwmon/blob/master/sysfs.cpp#L170

[pstrinkle]

Thanks for running that down.

…

On Fri, Mar 15, 2019 at 9:16 AM Matt Spinler ***@***.***> wrote: Turns out the difference is in the /= operator. From cpppreference.com, only for std::filesystem: path& operator/=(const path& p); 1) If p.is_absolute() || (p.has_root_name() && p.root_name() != root_name()), then replaces the current path with p as if by operator=(p) and finishes. It doesn't say that for experimental::filesystem. that breaks this line: https://github.com/openbmc/phosphor-hwmon/blob/master/sysfs.cpp#L170 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3499 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABv02CMuNcaOlRM0DcCG2-qYjtBhtSCKks5vW8dHgaJpZM4b2Pwo> .

-- "If computers get too powerful, we can organize them into a committee--that will do them in." - Bradley's Bromide http://www.xkcd.com/488/ -- end DRM http://www.xkcd.com/175/ http://xkcd.com/303/

[spinler]

i can put up a fix in a bit

[pstrinkle]

Thanks!

On Fri, Mar 15, 2019 at 09:25 Matt Spinler ***@***.***> wrote: i can put up a fix in a bit — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3499 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABv02GQiqXFc4lJNYIg27ou1YbPsuoRcks5vW8lpgaJpZM4b2Pwo> .

-- "If computers get too powerful, we can organize them into a committee--that will do them in." - Bradley's Bromide http://www.xkcd.com/488/ -- end DRM http://www.xkcd.com/175/ http://xkcd.com/303/

[spinler]

Resolved by https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-hwmon/+/19288/