Skip to contentSkip to conversation

drm/msm/dp: Avoid division by zero in msm_dp_ctrl_config_msa()

PATCH Superseded Nathan Chancellor 3 participants 6 messages Jan 8 – Jan 13

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/msm/dp: Avoid division by zero in msm_dp_ctrl_config_msa()

An (admittedly problematic) optimization change in LLVM 20 [1] turns known division by zero into the equivalent of __builtin_unreachable(), which invokes undefined behavior if it is encountered in a control flow graph, destroying code generation. When compile testing for x86_64, objtool flags an instance of this optimization triggering in msm_dp_ctrl_config_msa(), inlined into msm_dp_ctrl_on_stream():

drivers/gpu/drm/msm/msm.o: warning: objtool: msm_dp_ctrl_on_stream(): unexpected end of section .text.msm_dp_ctrl_on_stream

The zero division happens if the else branch in the first if statement in msm_dp_ctrl_config_msa() is taken because pixel_div is initialized to zero and it is not possible for LLVM to eliminate the else branch since rate is still not known after inlining into msm_dp_ctrl_on_stream(). Change pixel_div to one to make the division well defined in the presence of an unsupported rate, relying on the DRM_ERROR print to indicate the error to the user.

---
 drivers/gpu/drm/msm/dp/dp_ctrl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/msm/dp/dp_ctrl.c +1 -1
index cbcc7c2f0ffc..e4731c059ed8 100644
@@ -2395,7 +2395,7 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
 			       bool is_ycbcr_420)
 {
 	u32 pixel_m, pixel_n;
-	u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
+	u32 mvid, nvid, pixel_div = 1, dispcc_input_rate;
 	u32 const nvid_fixed = DP_LINK_CONSTANT_N_VALUE;
 	u32 const link_rate_hbr2 = 540000;
 	u32 const link_rate_hbr3 = 810000;

---
base-commit: 66691e272e40c91305f1704695e0cb340cd162ff
change-id: 20260108-drm-msm-dp_ctrl-avoid-zero-div-be5dc40cbc18

Best regards,

Signature · 1 line
Nathan Chancellor <[email protected]>
#1 lore raw
Quoted 41 lines · Nathan Chancellor

An (admittedly problematic) optimization change in LLVM 20 [1] turns known division by zero into the equivalent of __builtin_unreachable(), which invokes undefined behavior if it is encountered in a control flow graph, destroying code generation. When compile testing for x86_64, objtool flags an instance of this optimization triggering in msm_dp_ctrl_config_msa(), inlined into msm_dp_ctrl_on_stream():

drivers/gpu/drm/msm/msm.o: warning: objtool: msm_dp_ctrl_on_stream(): unexpected end of section .text.msm_dp_ctrl_on_stream

The zero division happens if the else branch in the first if statement in msm_dp_ctrl_config_msa() is taken because pixel_div is initialized to zero and it is not possible for LLVM to eliminate the else branch since rate is still not known after inlining into msm_dp_ctrl_on_stream(). Change pixel_div to one to make the division well defined in the presence of an unsupported rate, relying on the DRM_ERROR print to indicate the error to the user.

---
 drivers/gpu/drm/msm/dp/dp_ctrl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/msm/dp/dp_ctrl.c +1 -2
index cbcc7c2f0ffc..e4731c059ed8 100644
@@ -2395,7 +2395,7 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
 			       bool is_ycbcr_420)
 {
 	u32 pixel_m, pixel_n;
-	u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
+	u32 mvid, nvid, pixel_div = 1, dispcc_input_rate;
 	u32 const nvid_fixed = DP_LINK_CONSTANT_N_VALUE;
 	u32 const link_rate_hbr2 = 540000;
 	u32 const link_rate_hbr3 = 810000;
 
---
Quoted 2 lines

base-commit: 66691e272e40c91305f1704695e0cb340cd162ff
change-id: 20260108-drm-msm-dp_ctrl-avoid-zero-div-be5dc40cbc18

Dmitry, would it be beneficial to throw an actual error when the rate is is mangled? i.e.

drivers/gpu/drm/msm/dp/dp_ctrl.c +22 -12
index aa2303d0e148..4f710b8e6bc6 100644
@@ -2404,9 +2404,9 @@ static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl)
        return msm_dp_ctrl_setup_main_link(ctrl, &training_step);
 }
 
-static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
-                              u32 rate, u32 stream_rate_khz,
-                              bool is_ycbcr_420)
+static int msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
+                                 u32 rate, u32 stream_rate_khz,
+                                 bool is_ycbcr_420)
 {
        u32 pixel_m, pixel_n;
        u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
@@ -2415,14 +2415,21 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
        u32 const link_rate_hbr3 = 810000;
        unsigned long den, num;
 
-       if (rate == link_rate_hbr3)
+       switch (rate) {
+       case link_rate_hbr3:
                pixel_div = 6;
-       else if (rate == 162000 || rate == 270000)
-               pixel_div = 2;
-       else if (rate == link_rate_hbr2)
+               break;
+       case link_rate_hbr2:
                pixel_div = 4;
-       else
+               break;
+       case 270000:
+       case 162000:
+               pixel_div = 2;
+               break;
+       default:
                DRM_ERROR("Invalid pixel mux divider\n");
+               return -EINVAL;
+       }
 
        dispcc_input_rate = (rate * 10) / pixel_div;
 
@@ -2458,6 +2465,8 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
        drm_dbg_dp(ctrl->drm_dev, "mvid=0x%x, nvid=0x%x\n", mvid, nvid);
        msm_dp_write_link(ctrl, REG_DP_SOFTWARE_MVID, mvid);
        msm_dp_write_link(ctrl, REG_DP_SOFTWARE_NVID, nvid);
+
+       return 0;
 }
 
 int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train)
@@ -2525,10 +2534,11 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train
 
        msm_dp_ctrl_configure_source_params(ctrl);
 
-       msm_dp_ctrl_config_msa(ctrl,
-               ctrl->link->link_params.rate,
-               pixel_rate_orig,
-               ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
+       ret = msm_dp_ctrl_config_msa(ctrl, ctrl->link->link_params.rate,
+                                    pixel_rate_orig,
+                                    ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
+       if (ret)
+               return ret;
 
        msm_dp_panel_clear_dsc_dto(ctrl->panel);
 

Konrad

#2 lore raw
Quoted 2 lines · Konrad Dybcio

Dmitry, would it be beneficial to throw an actual error when the rate is is mangled? i.e.

This certainly seems more robust if acceptable.

Quoted 7 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +0 -0
index aa2303d0e148..4f710b8e6bc6 100644
@@ -2404,9 +2404,9 @@ static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl)
        return msm_dp_ctrl_setup_main_link(ctrl, &training_step);
 }

drivers/gpu/drm/msm/dp/dp_ctrl.c static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl) · line 2412

Quoted 10 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +3 -3
 
-static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
-                              u32 rate, u32 stream_rate_khz,
-                              bool is_ycbcr_420)
+static int msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
+                                 u32 rate, u32 stream_rate_khz,
+                                 bool is_ycbcr_420)
 {
        u32 pixel_m, pixel_n;
        u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;

This initialization is now unnecessary with this change.

Quoted 55 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +19 -9
@@ -2415,14 +2415,21 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
        u32 const link_rate_hbr3 = 810000;
        unsigned long den, num;
 
-       if (rate == link_rate_hbr3)
+       switch (rate) {
+       case link_rate_hbr3:
                pixel_div = 6;
-       else if (rate == 162000 || rate == 270000)
-               pixel_div = 2;
-       else if (rate == link_rate_hbr2)
+               break;
+       case link_rate_hbr2:
                pixel_div = 4;
-       else
+               break;
+       case 270000:
+       case 162000:
+               pixel_div = 2;
+               break;
+       default:
                DRM_ERROR("Invalid pixel mux divider\n");
+               return -EINVAL;
+       }
 
        dispcc_input_rate = (rate * 10) / pixel_div;
 
@@ -2458,6 +2465,8 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
        drm_dbg_dp(ctrl->drm_dev, "mvid=0x%x, nvid=0x%x\n", mvid, nvid);
        msm_dp_write_link(ctrl, REG_DP_SOFTWARE_MVID, mvid);
        msm_dp_write_link(ctrl, REG_DP_SOFTWARE_NVID, nvid);
+
+       return 0;
 }
 
 int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train)
@@ -2525,10 +2534,11 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train
 
        msm_dp_ctrl_configure_source_params(ctrl);
 
-       msm_dp_ctrl_config_msa(ctrl,
-               ctrl->link->link_params.rate,
-               pixel_rate_orig,
-               ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
+       ret = msm_dp_ctrl_config_msa(ctrl, ctrl->link->link_params.rate,
+                                    pixel_rate_orig,
+                                    ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
+       if (ret)
+               return ret;
 
        msm_dp_panel_clear_dsc_dto(ctrl->panel);
 

Konrad

#3 lore raw
In reply to Konrad Dybcio
Quoted 78 lines · Konrad Dybcio
Quoted 41 lines · Nathan Chancellor

An (admittedly problematic) optimization change in LLVM 20 [1] turns known division by zero into the equivalent of __builtin_unreachable(), which invokes undefined behavior if it is encountered in a control flow graph, destroying code generation. When compile testing for x86_64, objtool flags an instance of this optimization triggering in msm_dp_ctrl_config_msa(), inlined into msm_dp_ctrl_on_stream():

drivers/gpu/drm/msm/msm.o: warning: objtool: msm_dp_ctrl_on_stream(): unexpected end of section .text.msm_dp_ctrl_on_stream

The zero division happens if the else branch in the first if statement in msm_dp_ctrl_config_msa() is taken because pixel_div is initialized to zero and it is not possible for LLVM to eliminate the else branch since rate is still not known after inlining into msm_dp_ctrl_on_stream(). Change pixel_div to one to make the division well defined in the presence of an unsupported rate, relying on the DRM_ERROR print to indicate the error to the user.

---
 drivers/gpu/drm/msm/dp/dp_ctrl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/msm/dp/dp_ctrl.c +1 -2
index cbcc7c2f0ffc..e4731c059ed8 100644
@@ -2395,7 +2395,7 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
 			       bool is_ycbcr_420)
 {
 	u32 pixel_m, pixel_n;
-	u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
+	u32 mvid, nvid, pixel_div = 1, dispcc_input_rate;
 	u32 const nvid_fixed = DP_LINK_CONSTANT_N_VALUE;
 	u32 const link_rate_hbr2 = 540000;
 	u32 const link_rate_hbr3 = 810000;
 
---
Quoted 2 lines

base-commit: 66691e272e40c91305f1704695e0cb340cd162ff
change-id: 20260108-drm-msm-dp_ctrl-avoid-zero-div-be5dc40cbc18

Dmitry, would it be beneficial to throw an actual error when the rate is is mangled? i.e.

drivers/gpu/drm/msm/dp/dp_ctrl.c +7 -7
index aa2303d0e148..4f710b8e6bc6 100644
@@ -2404,9 +2404,9 @@ static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl)
        return msm_dp_ctrl_setup_main_link(ctrl, &training_step);
 }
 
-static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
-                              u32 rate, u32 stream_rate_khz,
-                              bool is_ycbcr_420)
+static int msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
+                                 u32 rate, u32 stream_rate_khz,
+                                 bool is_ycbcr_420)
 {
        u32 pixel_m, pixel_n;
        u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
@@ -2415,14 +2415,21 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
        u32 const link_rate_hbr3 = 810000;
        unsigned long den, num;
 
-       if (rate == link_rate_hbr3)
+       switch (rate) {
+       case link_rate_hbr3:
                pixel_div = 6;
-       else if (rate == 162000 || rate == 270000)
-               pixel_div = 2;
-       else if (rate == link_rate_hbr2)
+               break;
+       case link_rate_hbr2:

drivers/gpu/drm/msm/dp/dp_ctrl.c static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl, · line 2431

Quoted 10 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +7 -1
                pixel_div = 4;
-       else
+               break;
+       case 270000:
+       case 162000:
+               pixel_div = 2;
+               break;
+       default:
                DRM_ERROR("Invalid pixel mux divider\n");
+               return -EINVAL;

Well... In the ideal world, we can't end up here, PHY's configure_dp_clocks (or qcom_edp_set_vco_div()) will fail if the link rate is is invalid here. I think, we should return an error here, but there is no need to propagate it further.

See the discussion at https://lore.kernel.org/dri-devel/[email protected]/

Quoted 32 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +8 -4
+       }
 
        dispcc_input_rate = (rate * 10) / pixel_div;
 
@@ -2458,6 +2465,8 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
        drm_dbg_dp(ctrl->drm_dev, "mvid=0x%x, nvid=0x%x\n", mvid, nvid);
        msm_dp_write_link(ctrl, REG_DP_SOFTWARE_MVID, mvid);
        msm_dp_write_link(ctrl, REG_DP_SOFTWARE_NVID, nvid);
+
+       return 0;
 }
 
 int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train)
@@ -2525,10 +2534,11 @@ int msm_dp_ctrl_on_stream(struct msm_dp_ctrl *msm_dp_ctrl, bool force_link_train
 
        msm_dp_ctrl_configure_source_params(ctrl);
 
-       msm_dp_ctrl_config_msa(ctrl,
-               ctrl->link->link_params.rate,
-               pixel_rate_orig,
-               ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
+       ret = msm_dp_ctrl_config_msa(ctrl, ctrl->link->link_params.rate,
+                                    pixel_rate_orig,
+                                    ctrl->panel->msm_dp_mode.out_fmt_is_yuv_420);
+       if (ret)
+               return ret;
 
        msm_dp_panel_clear_dsc_dto(ctrl->panel);
 

Konrad

Signature · 2 lines
With best wishes
Dmitry
#4 lore raw
Quoted 50 lines · Dmitry Baryshkov
Quoted 33 lines · Konrad Dybcio

Dmitry, would it be beneficial to throw an actual error when the rate is is mangled? i.e.

drivers/gpu/drm/msm/dp/dp_ctrl.c +7 -7
index aa2303d0e148..4f710b8e6bc6 100644
@@ -2404,9 +2404,9 @@ static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl)
        return msm_dp_ctrl_setup_main_link(ctrl, &training_step);
 }
 
-static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
-                              u32 rate, u32 stream_rate_khz,
-                              bool is_ycbcr_420)
+static int msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
+                                 u32 rate, u32 stream_rate_khz,
+                                 bool is_ycbcr_420)
 {
        u32 pixel_m, pixel_n;
        u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
@@ -2415,14 +2415,21 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
        u32 const link_rate_hbr3 = 810000;
        unsigned long den, num;
 
-       if (rate == link_rate_hbr3)
+       switch (rate) {
+       case link_rate_hbr3:
                pixel_div = 6;
-       else if (rate == 162000 || rate == 270000)
-               pixel_div = 2;
-       else if (rate == link_rate_hbr2)
+               break;
+       case link_rate_hbr2:

drivers/gpu/drm/msm/dp/dp_ctrl.c static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl, · line 2431

Quoted 10 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +7 -1
                pixel_div = 4;
-       else
+               break;
+       case 270000:
+       case 162000:
+               pixel_div = 2;
+               break;
+       default:
                DRM_ERROR("Invalid pixel mux divider\n");
+               return -EINVAL;

Well... In the ideal world, we can't end up here, PHY's configure_dp_clocks (or qcom_edp_set_vco_div()) will fail if the link rate is is invalid here. I think, we should return an error here, but there is no need to propagate it further.

Quoted 1 line

See the discussion at https://lore.kernel.org/dri-devel/[email protected]/

I interpret that as approving of the above hunk but omitting the hunk that modifies msm_dp_ctrl_on_stream(). In that case, what is the point of changing the return type of msm_dp_ctrl_config_msa()? Wouldn't the below diff have the same exact effect as a smaller change? I don't mind rolling this up as a v2.

Cheers,
Nathan

drivers/gpu/drm/msm/dp/dp_ctrl.c +13 -6
index aa2303d0e148..d8ea73b89f7c 100644
@@ -2409,20 +2409,27 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
 			       bool is_ycbcr_420)
 {
 	u32 pixel_m, pixel_n;
-	u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
+	u32 mvid, nvid, pixel_div, dispcc_input_rate;
 	u32 const nvid_fixed = DP_LINK_CONSTANT_N_VALUE;
 	u32 const link_rate_hbr2 = 540000;
 	u32 const link_rate_hbr3 = 810000;
 	unsigned long den, num;
 
-	if (rate == link_rate_hbr3)
+	switch (rate) {
+	case link_rate_hbr3:
 		pixel_div = 6;
-	else if (rate == 162000 || rate == 270000)
-		pixel_div = 2;
-	else if (rate == link_rate_hbr2)
+		break;
+	case link_rate_hbr2:
 		pixel_div = 4;
-	else
+		break;
+	case 162000:
+	case 270000:
+		pixel_div = 2;
+		break;
+	default:
 		DRM_ERROR("Invalid pixel mux divider\n");
+		return;
+	}
 
 	dispcc_input_rate = (rate * 10) / pixel_div;
 
#5 lore raw
Quoted 86 lines · Nathan Chancellor
Quoted 50 lines · Dmitry Baryshkov
Quoted 33 lines · Konrad Dybcio

Dmitry, would it be beneficial to throw an actual error when the rate is is mangled? i.e.

drivers/gpu/drm/msm/dp/dp_ctrl.c +7 -7
index aa2303d0e148..4f710b8e6bc6 100644
@@ -2404,9 +2404,9 @@ static int msm_dp_ctrl_link_retrain(struct msm_dp_ctrl_private *ctrl)
         return msm_dp_ctrl_setup_main_link(ctrl, &training_step);
  }
 
-static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
-                              u32 rate, u32 stream_rate_khz,
-                              bool is_ycbcr_420)
+static int msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
+                                 u32 rate, u32 stream_rate_khz,
+                                 bool is_ycbcr_420)
  {
         u32 pixel_m, pixel_n;
         u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
@@ -2415,14 +2415,21 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
         u32 const link_rate_hbr3 = 810000;
         unsigned long den, num;
 
-       if (rate == link_rate_hbr3)
+       switch (rate) {
+       case link_rate_hbr3:
                 pixel_div = 6;
-       else if (rate == 162000 || rate == 270000)
-               pixel_div = 2;
-       else if (rate == link_rate_hbr2)
+               break;
+       case link_rate_hbr2:

drivers/gpu/drm/msm/dp/dp_ctrl.c static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl, · line 2431

Quoted 10 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +7 -1
                 pixel_div = 4;
-       else
+               break;
+       case 270000:
+       case 162000:
+               pixel_div = 2;
+               break;
+       default:
                 DRM_ERROR("Invalid pixel mux divider\n");
+               return -EINVAL;

Well... In the ideal world, we can't end up here, PHY's configure_dp_clocks (or qcom_edp_set_vco_div()) will fail if the link rate is is invalid here. I think, we should return an error here, but there is no need to propagate it further.

Quoted 1 line

See the discussion at https://lore.kernel.org/dri-devel/[email protected]/

I interpret that as approving of the above hunk but omitting the hunk that modifies msm_dp_ctrl_on_stream(). In that case, what is the point of changing the return type of msm_dp_ctrl_config_msa()? Wouldn't the below diff have the same exact effect as a smaller change? I don't mind rolling this up as a v2.

Cheers,
Nathan

drivers/gpu/drm/msm/dp/dp_ctrl.c +5 -5
index aa2303d0e148..d8ea73b89f7c 100644
@@ -2409,20 +2409,27 @@ static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl,
  			       bool is_ycbcr_420)
  {
  	u32 pixel_m, pixel_n;
-	u32 mvid, nvid, pixel_div = 0, dispcc_input_rate;
+	u32 mvid, nvid, pixel_div, dispcc_input_rate;
  	u32 const nvid_fixed = DP_LINK_CONSTANT_N_VALUE;
  	u32 const link_rate_hbr2 = 540000;
  	u32 const link_rate_hbr3 = 810000;
  	unsigned long den, num;
 
-	if (rate == link_rate_hbr3)
+	switch (rate) {
+	case link_rate_hbr3:
  		pixel_div = 6;
-	else if (rate == 162000 || rate == 270000)
-		pixel_div = 2;
-	else if (rate == link_rate_hbr2)
+		break;
+	case link_rate_hbr2:

drivers/gpu/drm/msm/dp/dp_ctrl.c static void msm_dp_ctrl_config_msa(struct msm_dp_ctrl_private *ctrl, · line 2431

Quoted 10 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +7 -1
  		pixel_div = 4;
-	else
+		break;
+	case 162000:
+	case 270000:
+		pixel_div = 2;
+		break;
+	default:
  		DRM_ERROR("Invalid pixel mux divider\n");
+		return;

Please add a comment, stating that we can't actually reach it. LGTM otherwise.

Quoted 4 lines
drivers/gpu/drm/msm/dp/dp_ctrl.c +1 -0
+	}
 
  	dispcc_input_rate = (rate * 10) / pixel_div;
 
Signature · 2 lines
With best wishes
Dmitry