[PATCH v3] cpufreq/pasemi: fix an use-after-free in pas_cpufreq_cpu_init()

STALE2639d

5 messages, 2 authors, 2019-07-19 · open the first message on its own page

[PATCH v3] cpufreq/pasemi: fix an use-after-free in pas_cpufreq_cpu_init()

From: Christian Zigotzky <hidden>
Date: 2019-07-08 21:03:38

Hello Wen,

Thanks for your patch!

Did you test your patch with a P.A. Semi board?

Cheers,
Christian

Re:[PATCH v3] cpufreq/pasemi: fix an use-after-free inpas_cpufreq_cpu_init()

From: <hidden>
Date: 2019-07-09 01:41:40

Hello Wen,

Thanks for your patch!

Did you test your patch with a P.A. Semi board?
Hello Christian, thank you.
We don't have a P.A. Semi board yet, so we didn't test it.
If you have such a board, could you please kindly help to test it?

--
Thanks and regards,
Wen

[PATCH v3] cpufreq/pasemi: fix an use-after-free inpas_cpufreq_cpu_init()

From: Christian Zigotzky <hidden>
Date: 2019-07-10 07:58:05

Is a final patch available for testing? Please do not release it without testing.

- Christian

On 09-07-19, 16:04, Wen Yang wrote:
quoted hunk
The cpu variable is still being used in the of_get_property() call
after the of_node_put() call, which may result in use-after-free.

Fixes: a9acc26b75f ("cpufreq/pasemi: fix possible object reference leak")
Signed-off-by: Wen Yang <wen.yang99 at zte.com.cn>
Cc: "Rafael J. Wysocki" <rjw at rjwysocki.net>
Cc: Viresh Kumar <viresh.kumar at linaro.org>
Cc: Michael Ellerman <mpe at ellerman.id.au>
Cc: linuxppc-dev at lists.ozlabs.org
Cc: linux-pm at vger.kernel.org
Cc: linux-kernel at vger.kernel.org
---
v5: put together the code to get, use, and release cpu device_node.
v4: restore the blank line.
v3: fix a leaked reference.
v2: clean up the code according to the advice of viresh.

 drivers/cpufreq/pasemi-cpufreq.c | 21 +++++++++------------
 1 file changed, 9 insertions(+), 12 deletions(-)
diff --git a/drivers/cpufreq/pasemi-cpufreq.c b/drivers/cpufreq/pasemi-cpufreq.c
index 6b1e4ab..1f0beb7 100644
--- a/drivers/cpufreq/pasemi-cpufreq.c
+++ b/drivers/cpufreq/pasemi-cpufreq.c
@@ -131,10 +131,17 @@ static int pas_cpufreq_cpu_init(struct cpufreq_policy *policy)
     int err = -ENODEV;

     cpu = of_get_cpu_node(policy->cpu, NULL);
-
-    of_node_put(cpu);
     if (!cpu)
         goto out;
I would have loved a blank line here :)
quoted hunk
+    max_freqp = of_get_property(cpu, "clock-frequency", NULL);
+    of_node_put(cpu);
+    if (!max_freqp) {
+        err = -EINVAL;
+        goto out;
+    }
+
+    /* we need the freq in kHz */
+    max_freq = *max_freqp / 1000;

     dn = of_find_compatible_node(NULL, NULL, "1682m-sdc");
     if (!dn)
@@ -171,16 +178,6 @@ static int pas_cpufreq_cpu_init(struct cpufreq_policy *policy)
     }

     pr_debug("init cpufreq on CPU %d\n", policy->cpu);
-
-    max_freqp = of_get_property(cpu, "clock-frequency", NULL);
-    if (!max_freqp) {
-        err = -EINVAL;
-        goto out_unmap_sdcpwr;
-    }
-
-    /* we need the freq in kHz */
-    max_freq = *max_freqp / 1000;
-
     pr_debug("max clock-frequency is at %u kHz\n", max_freq);
     pr_debug("initializing frequency table\n");
Though, enough versions have happened now.

Acked-by: Viresh Kumar <viresh.kumar at linaro.org>

-- 
viresh

[PATCH v7] cpufreq/pasemi: fix an use-after-free inpas_cpufreq_cpu_init()

From: Christian Zigotzky <hidden>
Date: 2019-07-18 12:49:10

On 09.07.2019 at 03:39am, wen.yang99@zte.com.cn wrote:
quoted
Hello Wen,

Thanks for your patch!

Did you test your patch with a P.A. Semi board?
Hello Christian, thank you.
We don't have a P.A. Semi board yet, so we didn't test it.
If you have such a board, could you please kindly help to test it?

--
Thanks and regards,
Wen
Hello Wen,

I successfully tested your pasemi cpufreq modifications with my P.A. Semi board [1] today.

First I patched the latest Git kernel with Viresh Kumar's patch [2]. After that I was able to patch the latest Git kernel with your v7 patch [3].

Then the kernel compiled without any errors.

Afterwards I successfully tested the new Git kernel with some cpufreq governors on openSUSE Tumbleweed 20190521 PowerPC64 [4] and on ubuntu MATE 16.04.6 LTS PowerPC32.

Thanks a lot for your work!

Tested-by: Christian Zigotzky <redacted>

Cheers,
Christian

[1] https://en.wikipedia.org/wiki/AmigaOne_X1000
[2] https://lore.kernel.org/lkml/ee8cf5fb4b4a01fdf9199037ff6d835b935cfd13.1562902877.git.viresh.kumar@linaro.org/#Z30drivers:cpufreq:pasemi-cpufreq.c
[3] https://lists.ozlabs.org/pipermail/linuxppc-dev/2019-July/193735.html
[4] Screenshots: https://i.pinimg.com/originals/37/66/93/37669306cbc909a9d79270a849d18aa6.png and https://i.pinimg.com/originals/fe/f8/bf/fef8bfc90d95b5ae9cf31e175e8ba2da.png

Re:[PATCH v7] cpufreq/pasemi: fix an use-after-freeinpas_cpufreq_cpu_init()

From: <hidden>
Date: 2019-07-19 05:57:37

quoted
quoted
Hello Wen,

Thanks for your patch!

Did you test your patch with a P.A. Semi board?
Hello Christian, thank you.
We don't have a P.A. Semi board yet, so we didn't test it.
If you have such a board, could you please kindly help to test it?

--
Thanks and regards,
Wen
Hello Wen,

I successfully tested your pasemi cpufreq modifications with my P.A.
Semi board [1] today.

First I patched the latest Git kernel with Viresh Kumar's patch [2].
After that I was able to patch the latest Git kernel with your v7 patch [3].

Then the kernel compiled without any errors.

Afterwards I successfully tested the new Git kernel with some cpufreq
governors on openSUSE Tumbleweed 20190521 PowerPC64 [4] and on ubuntu
MATE 16.04.6 LTS PowerPC32.

Thanks a lot for your work!

Tested-by: Christian Zigotzky <redacted>

Cheers,
Christian
Thank you very much!
--
Cheers,
Wen
[1] https://en.wikipedia.org/wiki/AmigaOne_X1000
[2]
https://lore.kernel.org/lkml/ee8cf5fb4b4a01fdf9199037ff6d835b935cfd13.1562902877.git.viresh.kumar@linaro.org/#Z30drivers:cpufreq:pasemi-cpufreq.c
[3] https://lists.ozlabs.org/pipermail/linuxppc-dev/2019-July/193735.html
[4] Screenshots:
https://i.pinimg.com/originals/37/66/93/37669306cbc909a9d79270a849d18aa6.png
and
https://i.pinimg.com/originals/fe/f8/bf/fef8bfc90d95b5ae9cf31e175e8ba2da.png
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help