| Message ID | 155017550946.954637.8586902409974005660.stgit@dwillia2-desk3.amr.corp.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] nfit/ars: Allow root to busy-poll the ARS state machine | expand |
On Thu, Feb 14, 2019 at 12:32 PM Dan Williams <dan.j.williams@intel.com> wrote: > > The ARS implementation implements exponential back-off on the poll > interval to prevent high-frequency access to the DIMM / platform > interface. Depending on when the ARS completes the poll interval may > exceed the completion event by minutes. Allow root to reset the timeout > each time it probes the status. A one-second timeout is still enforced, > but root can otherwise can control the poll interval. > > Reported-by: Erwin Tsaur <erwin.tsaur@oracle.com> > Signed-off-by: Dan Williams <dan.j.williams@intel.com> > --- > Change since v1: Use mod_delayed_work() instead of queue_delayed_work() > to modify the timeout for existing work. > > drivers/acpi/nfit/core.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c > index 934be96dc149..b1ab593a808a 100644 > --- a/drivers/acpi/nfit/core.c > +++ b/drivers/acpi/nfit/core.c > @@ -1328,6 +1328,12 @@ static ssize_t scrub_show(struct device *dev, > rc = sprintf(buf, "%d%s", acpi_desc->scrub_count, > acpi_desc->scrub_busy > && !acpi_desc->cancel ? "+\n" : "\n"); > + /* Allow an admin to poll the busy state at a higher rate */ > + if (acpi_desc->scrub_busy && !acpi_desc->cancel > + && capable(CAP_SYS_RAWIO)) { > + acpi_desc->scrub_tmo = 1; > + mod_delayed_work(nfit_wq, &acpi_desc->dwork, HZ); > + } I added support to ndctl to specify a poll-interval to 'ndctl wait-scrub'. The support highlighted an infinite loop problem as mod_delayed_work() called in a loop prevented the workqueue from ever running. A revised patch-set adds a new ARS_POLL flag to ensure that the workqueue runs at least once after every mod_delayed_work().
diff --git a/drivers/acpi/nfit/core.c b/drivers/acpi/nfit/core.c index 934be96dc149..b1ab593a808a 100644 --- a/drivers/acpi/nfit/core.c +++ b/drivers/acpi/nfit/core.c @@ -1328,6 +1328,12 @@ static ssize_t scrub_show(struct device *dev, rc = sprintf(buf, "%d%s", acpi_desc->scrub_count, acpi_desc->scrub_busy && !acpi_desc->cancel ? "+\n" : "\n"); + /* Allow an admin to poll the busy state at a higher rate */ + if (acpi_desc->scrub_busy && !acpi_desc->cancel + && capable(CAP_SYS_RAWIO)) { + acpi_desc->scrub_tmo = 1; + mod_delayed_work(nfit_wq, &acpi_desc->dwork, HZ); + } mutex_unlock(&acpi_desc->init_mutex); } device_unlock(dev);
The ARS implementation implements exponential back-off on the poll interval to prevent high-frequency access to the DIMM / platform interface. Depending on when the ARS completes the poll interval may exceed the completion event by minutes. Allow root to reset the timeout each time it probes the status. A one-second timeout is still enforced, but root can otherwise can control the poll interval. Reported-by: Erwin Tsaur <erwin.tsaur@oracle.com> Signed-off-by: Dan Williams <dan.j.williams@intel.com> --- Change since v1: Use mod_delayed_work() instead of queue_delayed_work() to modify the timeout for existing work. drivers/acpi/nfit/core.c | 6 ++++++ 1 file changed, 6 insertions(+)