[PATCH] drm/msm/gpu: Fix missing error check for dev_pm_qos_add_request()

2024-10-26 Thread Jinjie Ruan
dev_pm_qos_add_request() can fail, and it returns -EINVAL in case of wrong parameters, return -ENOMEM if there's not enough memory to allocate for data structures, and return -ENODEV if the device has just been removed from the system. If it fails in msm_devfreq_init(), there is no point in going o

Re: [PATCH] drm/msm/gpu: Fix missing error check for dev_pm_qos_add_request()

2024-10-26 Thread Dmitry Baryshkov
On Sat, Oct 26, 2024 at 05:37:38PM +0800, Jinjie Ruan wrote: > dev_pm_qos_add_request() can fail, and it returns -EINVAL in case of > wrong parameters, return -ENOMEM if there's not enough memory to allocate > for data structures, and return -ENODEV if the device has just been > removed from the sy

[PATCH v2 1/2] arm64: dts: qcom: sm8550: correct MDSS interconnects

2024-10-26 Thread Dmitry Baryshkov
SM8550 lists two interconnects for the display subsystem, mdp0-mem (between MDP and LLCC) and mdp1-mem (between LLCC and EBI, memory). The second interconnect is a misuse. mdpN-mem paths should be used for several outboud MDP interconnects rather than the path between LLCC and memory. This kind of

[PATCH v2 2/2] arm64: dts: qcom: sm8650: correct MDSS interconnects

2024-10-26 Thread Dmitry Baryshkov
SM8650 lists two interconnects for the display subsystem, mdp0-mem (between MDP and LLCC) and mdp1-mem (between LLCC and EBI, memory). The second interconnect is a misuse. mdpN-mem paths should be used for several outboud MDP interconnects rather than the path between LLCC and memory. This kind of

[PATCH v2 0/2] arm64: dts: qcom: sm8[56]50: correct MDSS interconnects

2024-10-26 Thread Dmitry Baryshkov
Both SM8550 and SM8650 misuse mdp1-mem interconnect path for the LLCC->EBI path, while it should only be used for the MDP->EBI paths. This kind of misuse can result in bandwidth underflows, possibly degrading picture quality as the required memory bandwidth is divided between all mdpN-mem paths (a

Re: [PATCH 1/3] drm/msm: Avoid NULL dereference in msm_disp_state_print_regs()

2024-10-26 Thread Dmitry Baryshkov
On Mon, 14 Oct 2024 09:36:08 -0700, Douglas Anderson wrote: > If the allocation in msm_disp_state_dump_regs() failed then > `block->state` can be NULL. The msm_disp_state_print_regs() function > _does_ have code to try to handle it with: > > if (*reg) > dump_addr = *reg; > > [...] Applie