On 10/22/2024 08:59, Ilpo Järvinen wrote:
On Tue, 22 Oct 2024, Mario Limonciello wrote:
On 10/22/2024 08:22, Ilpo Järvinen wrote:
On Tue, 22 Oct 2024, Mario Limonciello wrote:
On 10/22/2024 00:54, Basavaraj Natikar wrote:
On 10/22/2024 2:05 AM, Armin Wolf wrote:
Am 21.10.24 um 18:58 schrieb Basavaraj Natikar:
AMD X3D processors, also known as AMD 3D V-Cache, feature dual Core
Complex Dies (CCDs) and enlarged L3 cache, enabling dynamic mode
switching between Frequency and Cache modes. To optimize
performance,
implement the AMD 3D V-Cache Optimizer, which allows selecting
either:
Frequency mode: cores within the faster CCD are prioritized before
those in the slower CCD.
Cache mode: cores within the larger L3 CCD are prioritized before
those in the smaller L3 CCD.
Co-developed-by: Perry Yuan <perry.yuan@xxxxxxx>
Signed-off-by: Perry Yuan <perry.yuan@xxxxxxx>
Co-developed-by: Mario Limonciello <mario.limonciello@xxxxxxx>
Signed-off-by: Mario Limonciello <mario.limonciello@xxxxxxx>
Reviewed-by: Shyam Sundar S K <Shyam-sundar.S-k@xxxxxxx>
Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@xxxxxxx>
+static int amd_x3d_resume_handler(struct device *dev)
+{
+ struct amd_x3d_dev *data = dev_get_drvdata(dev);
+ int ret = amd_x3d_get_mode(data);
+
+ ret = amd_x3d_mode_switch(data, ret);
+ if (ret < 0)
+ return ret;
+
+ return 0;
Please directly return the result of amd_x3d_mode_switch() here.
Also i think that maybe there exists a way to avoid locking data->lock
twice
during resume, but i will leave that to you.
I will leave as it is.
Other than that:
Reviewed-by: Armin Wolf <W_Armin@xxxxxx>
Thanks for the feedback. I will re-spin a new version after looking for
more
feedback from others. Thanks, -- Basavaraj
FYI - if no other feedback comes in there is no need to re-spin just to
add
R-b tags. Maintainers will automatically pick them up when they use 'b4
$LORE_URL'.
?? There's the return change requested above.
Basavaraj said he will leave it as is in response to Armin's comment.
To me that seemed related to the locking paragraph, not the return related
one.
Ah I see; thanks for correcting my misreading.
