From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Rafael J. Wysocki" Subject: Re: [PATCH 1/2] cpufreq: try to resume policies which failed on last resume Date: Thu, 26 Dec 2013 02:05:50 +0100 Message-ID: <5562479.pVWRuDL0y6@vostro.rjw.lan> References: Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: Sender: linux-pm-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="utf-8" To: Viresh Kumar , bjorn@mork.no Cc: linaro-kernel@lists.linaro.org, patches@linaro.org, cpufreq@vger.kernel.org, linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org On Tuesday, December 24, 2013 07:11:00 AM Viresh Kumar wrote: > __cpufreq_add_dev() can fail sometimes while we are resuming our syst= em. > Currently we are clearing all sysfs nodes for cpufreq's failed policy= as that > could make userspace unstable. But if we suspend/resume again, we sho= uld atleast > try to bring back those policies. >=20 > This patch fixes this issue by clearing fallback data on failure and = trying to > allocate a new struct cpufreq_policy on second resume. >=20 > Reported-and-tested-by: Bj=C3=B8rn Mork > Signed-off-by: Viresh Kumar Well, while I appreciate the work done here, I don't like the changelog= , I don't really like the way the code is written and I don't like the co= mments. Sorry about that. Bjorn, can you please test the patch below instead along with the [2/2] from this series (on top of linux-pm.git/pm-cpufreq)? Rafael --- =46rom: Rafael J. Wysocki Subject: cpufreq: Clean up after a failing light-weight initialization If cpufreq_policy_restore() returns NULL during system resume, __cpufreq_add_dev() should just fall back to the full initialization instead of returning an error, because that may actually make things work. Moreover, it should not leave stale fallback data behind after it has failed to restore a previously existing policy. This change is based on Viresh Kumar's work. Reported-by: Bj=C3=B8rn Mork Signed-off-by: Rafael J. Wysocki --- drivers/cpufreq/cpufreq.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) Index: linux-pm/drivers/cpufreq/cpufreq.c =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D --- linux-pm.orig/drivers/cpufreq/cpufreq.c +++ linux-pm/drivers/cpufreq/cpufreq.c @@ -1016,15 +1016,17 @@ static int __cpufreq_add_dev(struct devi read_unlock_irqrestore(&cpufreq_driver_lock, flags); #endif =20 - if (frozen) - /* Restore the saved policy when doing light-weight init */ - policy =3D cpufreq_policy_restore(cpu); - else + /* + * Restore the saved policy when doing light-weight init and fall bac= k + * to the full init if that fails. + */ + policy =3D frozen ? cpufreq_policy_restore(cpu) : NULL; + if (!policy) { + frozen =3D false; policy =3D cpufreq_policy_alloc(); - - if (!policy) - goto nomem_out; - + if (!policy) + goto nomem_out; + } =20 /* * In the resume path, since we restore a saved policy, the assignmen= t @@ -1118,8 +1120,11 @@ err_get_freq: if (cpufreq_driver->exit) cpufreq_driver->exit(policy); err_set_policy_cpu: - if (frozen) + if (frozen) { + /* Do not leave stale fallback data behind. */ + per_cpu(cpufreq_cpu_data_fallback, cpu) =3D NULL; cpufreq_policy_put_kobj(policy); + } cpufreq_policy_free(policy); =20 nomem_out: