linux-kernel - [RFC PATCH] cpufreq: mark duplicate frequencies as invalid and continue as normal

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20191022173215.13350-1-sudeep.holla@arm.com>
Date:   Tue, 22 Oct 2019 18:32:15 +0100
From:   Sudeep Holla <sudeep.holla@....com>
To:     Viresh Kumar <viresh.kumar@...aro.org>,
        "Rafael J . Wysocki" <rjw@...ysocki.net>
Cc:     Sudeep Holla <sudeep.holla@....com>, linux-pm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: [RFC PATCH] cpufreq: mark duplicate frequencies as invalid and continue as normal

Currently if we encounter duplicate frequency table entries, we abort
the validation and return error immediately. Instead of failing, we
can mark the entry as invalid and continue to function normal.

Cc: "Rafael J. Wysocki" <rjw@...ysocki.net>
Cc: Viresh Kumar <viresh.kumar@...aro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@....com>
---
 drivers/cpufreq/freq_table.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

Hi Viresh,

Since commit da0c6dc00c69 ("cpufreq: Handle sorted frequency tables more
efficiently"), I seem to have modified the firmware entry on my TC2 to
drop 500MHz and had not seen the issue with duplicate entries and had
totally forgotten about it.

Recently I reverted back to original setting as I corrupted it and
started seeing this issues. I don't know the background for raising
duplicates as fatal error but we did allow it when we add arm_big_little.c
and hence this RFC. If there are known issues with this approach, I can
continue with changed firmware config.

With switcher, we have:
(little cluster)
Virt: 175 MHz, 200 MHz, 250 MHz, 300 MHz, 350 MHz, 400 MHz, 450 MHz, 500 MHz
Actu: 350 MHz, 400 MHz, 500 MHz, 600 MHz, 700 MHz, 800 MHz, 900 MHz, 1000 MHz
(big cluster)
500 MHz, 600 MHz, 700 MHz, 800 MHz, 900 MHz, 1000 MHz, 1.10 GHz, 1.20 GHz

with 500 MHz duplicate in merged table.

Regards,
Sudeep

diff --git a/drivers/cpufreq/freq_table.c b/drivers/cpufreq/freq_table.c
index ded427e0a488..e9bf287846d6 100644
--- a/drivers/cpufreq/freq_table.c
+++ b/drivers/cpufreq/freq_table.c
@@ -305,9 +305,10 @@ static int set_freq_table_sorted(struct cpufreq_policy *policy)
 		}

 		if (pos->frequency == prev->frequency) {
-			pr_warn("Duplicate freq-table entries: %u\n",
+			pr_warn("Duplicate freq-table entries: %u marking it invalid\n,",
 				pos->frequency);
-			return -EINVAL;
+			pos->frequency = CPUFREQ_ENTRY_INVALID;
+			continue;
 		}

 		/* Frequency increased from prev to pos */
-- 
2.17.1