On 8/28/2024 7:25 PM, Brian Masney wrote:
On Wed, Aug 28, 2024 at 4:13 PM Ernesto A. Fernández
<ernesto.mnd.fernandez@xxxxxxxxx> wrote:
On Wed, Aug 28, 2024 at 12:03:57PM -0700, Trilok Soni wrote:
On 8/28/2024 11:40 AM, Ernesto A. Fernández wrote:
Hi, I have a bug to report.
I'm getting a null pointer dereference inside qcom_rng_probe() when this
driver gets loaded. The problem comes from here:
rng->of_data = (struct qcom_rng_of_data *)of_device_get_match_data(&pdev->dev);
because of_device_get_match_data() will just return null for acpi. It seems
that acpi was left behind by the changes in commit f29cd5bb64c2 ("crypto:
qcom-rng - Add hw_random interface support").
Which Qualcomm platform you are testing w/ the ACPI? Most of our platforms
uses the devicetree.
Amberwing.
We have a few Amberwing servers in the lab at Red Hat. I verified that
qcom-rng was crashing on boot with an upstream kernel, and it's now
fixed with this:
https://lore.kernel.org/linux-arm-msm/20240829012005.382715-1-bmasney@xxxxxxxxxx/T/#t
Brian
Interesting, I haven't seen this in my testing. I'll go swing back and
try to figure out why.
For future reference, I'm still supporting Amberwing/QDF2400.
Addressing emails to me is a good way to get my attention.
-Jeff
