Skip to contentSkip to conversation

drm/ast: Treat AST2600 like AST2500 in most places

PATCH In review Thomas Zimmermann 3 participants 8 messages Jun 7, 2022 – Jun 9, 2022

Subsystems (4)
lore
Share & tools
lore

Thomas Zimmermann

author maintainer
#0 lore raw

PATCH 1/1 drm/ast: Treat AST2600 like AST2500 in most places

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

Fixes f9bd00e0ea9d ("drm/ast: Create chip AST2600")
Cc KuoHsiang Chou <[email protected]>
Cc Dave Airlie <[email protected]>
Cc <[email protected]> # v5.11+
---
 drivers/gpu/drm/ast/ast_main.c | 4 ++--
 drivers/gpu/drm/ast/ast_mode.c | 6 +++---
 drivers/gpu/drm/ast/ast_post.c | 6 +++---
 3 files changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/ast/ast_main.c +2 -2
index d770d5a23c1a..56b2ac138375 100644
@@ -307,7 +307,7 @@ static int ast_get_dram_info(struct drm_device *dev)
 	default:
 		ast->dram_bus_width = 16;
 		ast->dram_type = AST_DRAM_1Gx16;
-		if (ast->chip == AST2500)
+		if ((ast->chip == AST2500) || (ast->chip == AST2600))
 			ast->mclk = 800;
 		else
 			ast->mclk = 396;
@@ -319,7 +319,7 @@ static int ast_get_dram_info(struct drm_device *dev)
 	else
 		ast->dram_bus_width = 32;
 
-	if (ast->chip == AST2500) {
+	if ((ast->chip == AST2600) || (ast->chip == AST2500)) {
 		switch (mcr_cfg & 0x03) {
 		case 0:
 			ast->dram_type = AST_DRAM_1Gx16;
drivers/gpu/drm/ast/ast_mode.c +3 -3
index 323af2746aa9..1dde30b98317 100644
@@ -310,7 +310,7 @@ static void ast_set_crtc_reg(struct ast_private *ast,
 	u8 jreg05 = 0, jreg07 = 0, jreg09 = 0, jregAC = 0, jregAD = 0, jregAE = 0;
 	u16 temp, precache = 0;
 
-	if ((ast->chip == AST2500) &&
+	if (((ast->chip == AST2600) || (ast->chip == AST2500)) &&
 	    (vbios_mode->enh_table->flags & AST2500PreCatchCRT))
 		precache = 40;
 
@@ -428,7 +428,7 @@ static void ast_set_dclk_reg(struct ast_private *ast,
 {
 	const struct ast_vbios_dclk_info *clk_info;
 
-	if (ast->chip == AST2500)
+	if ((ast->chip == AST2600) || (ast->chip == AST2500))
 		clk_info = &dclk_table_ast2500[vbios_mode->enh_table->dclk_index];
 	else
 		clk_info = &dclk_table[vbios_mode->enh_table->dclk_index];
@@ -476,7 +476,7 @@ static void ast_set_crtthd_reg(struct ast_private *ast)
 		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0xe0);
 		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0xa0);
 	} else if (ast->chip == AST2300 || ast->chip == AST2400 ||
-	    ast->chip == AST2500) {
+		   ast->chip == AST2500 || ast->chip == AST2600) {
 		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0x78);
 		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0x60);
 	} else if (ast->chip == AST2100 ||
drivers/gpu/drm/ast/ast_post.c +3 -3
index 0aa9cf0fb5c3..eb1ff9084034 100644
@@ -80,7 +80,7 @@ ast_set_def_ext_reg(struct drm_device *dev)
 		ast_set_index_reg(ast, AST_IO_CRTC_PORT, i, 0x00);
 
 	if (ast->chip == AST2300 || ast->chip == AST2400 ||
-	    ast->chip == AST2500) {
+	    ast->chip == AST2500 || ast->chip == AST2600) {
 		if (pdev->revision >= 0x20)
 			ext_reg_info = extreginfo_ast2300;
 		else
@@ -105,7 +105,7 @@ ast_set_def_ext_reg(struct drm_device *dev)
 	/* Enable RAMDAC for A1 */
 	reg = 0x04;
 	if (ast->chip == AST2300 || ast->chip == AST2400 ||
-	    ast->chip == AST2500)
+	    ast->chip == AST2500 || ast->chip == AST2600)
 		reg |= 0x20;
 	ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb6, 0xff, reg);
 }
@@ -382,7 +382,7 @@ void ast_post_gpu(struct drm_device *dev)
 	if (ast->chip == AST2600) {
 		ast_dp_launch(dev, 1);
 	} else if (ast->config_mode == ast_use_p2a) {
-		if (ast->chip == AST2500)
+		if (ast->chip == AST2500 || ast->chip == AST2600)
 			ast_post_chip_2500(dev);
 		else if (ast->chip == AST2300 || ast->chip == AST2400)
 			ast_post_chip_2300(dev);
Signature · 1 line
2.36.1
author maintainer
#1 lore raw

Jocelyn, do you have a way of getting this patch tested?

Am 07.06.22 um 14:02 schrieb Thomas Zimmermann:

Quoted 109 lines

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

Fixes f9bd00e0ea9d ("drm/ast: Create chip AST2600")
Cc KuoHsiang Chou <[email protected]>
Cc Dave Airlie <[email protected]>
Cc <[email protected]> # v5.11+
---
  drivers/gpu/drm/ast/ast_main.c | 4 ++--
  drivers/gpu/drm/ast/ast_mode.c | 6 +++---
  drivers/gpu/drm/ast/ast_post.c | 6 +++---
  3 files changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/ast/ast_main.c +2 -2
index d770d5a23c1a..56b2ac138375 100644
@@ -307,7 +307,7 @@ static int ast_get_dram_info(struct drm_device *dev)
  	default:
  		ast->dram_bus_width = 16;
  		ast->dram_type = AST_DRAM_1Gx16;
-		if (ast->chip == AST2500)
+		if ((ast->chip == AST2500) || (ast->chip == AST2600))
  			ast->mclk = 800;
  		else
  			ast->mclk = 396;
@@ -319,7 +319,7 @@ static int ast_get_dram_info(struct drm_device *dev)
  	else
  		ast->dram_bus_width = 32;
 
-	if (ast->chip == AST2500) {
+	if ((ast->chip == AST2600) || (ast->chip == AST2500)) {
  		switch (mcr_cfg & 0x03) {
  		case 0:
  			ast->dram_type = AST_DRAM_1Gx16;
drivers/gpu/drm/ast/ast_mode.c +3 -3
index 323af2746aa9..1dde30b98317 100644
@@ -310,7 +310,7 @@ static void ast_set_crtc_reg(struct ast_private *ast,
  	u8 jreg05 = 0, jreg07 = 0, jreg09 = 0, jregAC = 0, jregAD = 0, jregAE = 0;
  	u16 temp, precache = 0;
 
-	if ((ast->chip == AST2500) &&
+	if (((ast->chip == AST2600) || (ast->chip == AST2500)) &&
  	    (vbios_mode->enh_table->flags & AST2500PreCatchCRT))
  		precache = 40;
 
@@ -428,7 +428,7 @@ static void ast_set_dclk_reg(struct ast_private *ast,
  {
  	const struct ast_vbios_dclk_info *clk_info;
 
-	if (ast->chip == AST2500)
+	if ((ast->chip == AST2600) || (ast->chip == AST2500))
  		clk_info = &dclk_table_ast2500[vbios_mode->enh_table->dclk_index];
  	else
  		clk_info = &dclk_table[vbios_mode->enh_table->dclk_index];
@@ -476,7 +476,7 @@ static void ast_set_crtthd_reg(struct ast_private *ast)
  		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0xe0);
  		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0xa0);
  	} else if (ast->chip == AST2300 || ast->chip == AST2400 ||
-	    ast->chip == AST2500) {
+		   ast->chip == AST2500 || ast->chip == AST2600) {
  		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0x78);
  		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0x60);
  	} else if (ast->chip == AST2100 ||
drivers/gpu/drm/ast/ast_post.c +3 -3
index 0aa9cf0fb5c3..eb1ff9084034 100644
@@ -80,7 +80,7 @@ ast_set_def_ext_reg(struct drm_device *dev)
  		ast_set_index_reg(ast, AST_IO_CRTC_PORT, i, 0x00);
 
  	if (ast->chip == AST2300 || ast->chip == AST2400 ||
-	    ast->chip == AST2500) {
+	    ast->chip == AST2500 || ast->chip == AST2600) {
  		if (pdev->revision >= 0x20)
  			ext_reg_info = extreginfo_ast2300;
  		else
@@ -105,7 +105,7 @@ ast_set_def_ext_reg(struct drm_device *dev)
  	/* Enable RAMDAC for A1 */
  	reg = 0x04;
  	if (ast->chip == AST2300 || ast->chip == AST2400 ||
-	    ast->chip == AST2500)
+	    ast->chip == AST2500 || ast->chip == AST2600)
  		reg |= 0x20;
  	ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb6, 0xff, reg);
  }
@@ -382,7 +382,7 @@ void ast_post_gpu(struct drm_device *dev)
  	if (ast->chip == AST2600) {
  		ast_dp_launch(dev, 1);
  	} else if (ast->config_mode == ast_use_p2a) {
-		if (ast->chip == AST2500)
+		if (ast->chip == AST2500 || ast->chip == AST2600)
  			ast_post_chip_2500(dev);
  		else if (ast->chip == AST2300 || ast->chip == AST2400)
  			ast_post_chip_2300(dev);
Signature · 6 lines
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Ivo Totev
#2 lore raw
Quoted 1 line · Thomas Zimmermann

Jocelyn, do you have a way of getting this patch tested?

Thanks for sending this patch.
I'm sorry I'm not able to test it directly.
It's a bit complex to try an upstream kernel on the machine we found the regression, I will do my best to get it tested, but no promise.

Also be careful when backporting to stable kernel, as the patch may apply, but some other "if (AST2500)" logic might be missing the AST2600 check.

Quoted 108 lines

Am 07.06.22 um 14:02 schrieb Thomas Zimmermann:

Quoted 105 lines

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

Fixes f9bd00e0ea9d ("drm/ast: Create chip AST2600")
Cc KuoHsiang Chou <[email protected]>
Cc Dave Airlie <[email protected]>
Cc <[email protected]> # v5.11+

---

  drivers/gpu/drm/ast/ast_main.c | 4 ++--
  drivers/gpu/drm/ast/ast_mode.c | 6 +++---
  drivers/gpu/drm/ast/ast_post.c | 6 +++---
  3 files changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/ast/ast_main.c +0 -0
index d770d5a23c1a..56b2ac138375 100644
@@ -307,7 +307,7 @@ static int ast_get_dram_info(struct drm_device *dev)

      default:
          ast->dram_bus_width = 16;
          ast->dram_type = AST_DRAM_1Gx16;
-        if (ast->chip == AST2500)
+        if ((ast->chip == AST2500) || (ast->chip == AST2600)) ast->mclk = 800;
          else
              ast->mclk = 396;

drivers/gpu/drm/ast/ast_main.c +0 -0
@@ -319,7 +319,7 @@ static int ast_get_dram_info(struct drm_device *dev)

      else
          ast->dram_bus_width = 32;
-    if (ast->chip == AST2500) {
+    if ((ast->chip == AST2600) || (ast->chip == AST2500)) { switch (mcr_cfg & 0x03) {
          case 0:
              ast->dram_type = AST_DRAM_1Gx16;

drivers/gpu/drm/ast/ast_mode.c +2 -2
index 323af2746aa9..1dde30b98317 100644
@@ -310,7 +310,7 @@ static void ast_set_crtc_reg(struct ast_private *ast,
       u8 jreg05 = 0, jreg07 = 0, jreg09 = 0, jregAC = 0, jregAD = 0, jregAE = 0;
       u16 temp, precache = 0;
-    if ((ast->chip == AST2500) &&
+    if (((ast->chip == AST2600) || (ast->chip == AST2500)) &&
           (vbios_mode->enh_table->flags & AST2500PreCatchCRT))
           precache = 40;
@@ -428,7 +428,7 @@ static void ast_set_dclk_reg(struct ast_private *ast,
   {
       const struct ast_vbios_dclk_info *clk_info;
-    if (ast->chip == AST2500)
+    if ((ast->chip == AST2600) || (ast->chip == AST2500))
           clk_info = &dclk_table_ast2500[vbios_mode->enh_table->dclk_index];

      else
          clk_info = &dclk_table[vbios_mode->enh_table->dclk_index];

drivers/gpu/drm/ast/ast_mode.c +1 -1
@@ -476,7 +476,7 @@ static void ast_set_crtthd_reg(struct ast_private 
 *ast)
           ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0xe0);
           ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0xa0);
       } else if (ast->chip == AST2300 || ast->chip == AST2400 ||
-        ast->chip == AST2500) {
+           ast->chip == AST2500 || ast->chip == AST2600) {
           ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0x78);
           ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0x60);

      } else if (ast->chip == AST2100 ||

drivers/gpu/drm/ast/ast_post.c +0 -0
index 0aa9cf0fb5c3..eb1ff9084034 100644
@@ -80,7 +80,7 @@ ast_set_def_ext_reg(struct drm_device *dev)
           ast_set_index_reg(ast, AST_IO_CRTC_PORT, i, 0x00);

      if (ast->chip == AST2300 || ast->chip == AST2400 || -        ast->chip == AST2500) {
+        ast->chip == AST2500 || ast->chip == AST2600) { if (pdev->revision >= 0x20) ext_reg_info = extreginfo_ast2300; else

drivers/gpu/drm/ast/ast_post.c +0 -0
@@ -105,7 +105,7 @@ ast_set_def_ext_reg(struct drm_device *dev)
       /* Enable RAMDAC for A1 */
       reg = 0x04;

      if (ast->chip == AST2300 || ast->chip == AST2400 || -        ast->chip == AST2500)
+        ast->chip == AST2500 || ast->chip == AST2600)
          reg |= 0x20;
      ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb6, 0xff, reg); }

drivers/gpu/drm/ast/ast_post.c +1 -1
@@ -382,7 +382,7 @@ void ast_post_gpu(struct drm_device *dev)
       if (ast->chip == AST2600) {
           ast_dp_launch(dev, 1);
       } else if (ast->config_mode == ast_use_p2a) {
-        if (ast->chip == AST2500)
+        if (ast->chip == AST2500 || ast->chip == AST2600)
               ast_post_chip_2500(dev);
           else if (ast->chip == AST2300 || ast->chip == AST2400)
               ast_post_chip_2300(dev);
#3 lore raw
In reply to Thomas Zimmermann

Hi Thomas

Thanks for your suggestions!

I answer each revision inline that followed by [KH]:.

Regards,

Kuo-Hsiang Chou

-----Original Message-----

From: Thomas Zimmermann [mailto:[email protected]]

Sent: Tuesday, June 07, 2022 8:03 PM

To: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Kuo-Hsiang Chou <[email protected]>

Subject: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

[KH]: Based on CVE_2019_6260 item3, P2A is disallowed anymore.

P2A (PCIe to AMBA) is a bridge that is able to revise any BMC registers.

Yes, P2A is dangerous on security issue, because Host open a backdoor and someone malicious SW/APP will be easy to take control of BMC.

Therefore, P2A is disabled forever.

Now, return to this patch, there is no need to add AST2600 condition on the P2A flow.

Fixes f9bd00e0ea9d ("drm/ast: Create chip AST2600")
Cc KuoHsiang Chou <[email protected]>
Cc Dave Airlie <[email protected]>
Cc <[email protected]> # v5.11+

---

drivers/gpu/drm/ast/ast_main.c | 4 ++-- drivers/gpu/drm/ast/ast_mode.c | 6 +++--- drivers/gpu/drm/ast/ast_post.c | 6 +++---

3 files changed, 8 insertions(+), 8 deletions(-)
drivers/gpu/drm/ast/ast_main.c index d770d5a23c1a..56b2ac138375 100644 renamed from drivers/gpu/drm/ast/ast_main.c +0 -0

--- a/drivers/gpu/drm/ast/ast_main.c

+++ b/drivers/gpu/drm/ast/ast_main.c

@@ -307,7 +307,7 @@ static int ast_get_dram_info(struct drm_device *dev)

      default:

              ast->dram_bus_width = 16;

              ast->dram_type = AST_DRAM_1Gx16;

- if (ast->chip == AST2500)

+ if ((ast->chip == AST2500) || (ast->chip == AST2600))

                      ast->mclk = 800;

              else

                      ast->mclk = 396;

@@ -319,7 +319,7 @@ static int ast_get_dram_info(struct drm_device *dev)

      else

              ast->dram_bus_width = 32;

- if (ast->chip == AST2500) {

+ if ((ast->chip == AST2600) || (ast->chip == AST2500)) {

              switch (mcr_cfg & 0x03) {

              case 0:

                      ast->dram_type = AST_DRAM_1Gx16;

[KH]: P2A is disabled, there is no need to take care of BMC DRAM setting.

drivers/gpu/drm/ast/ast_mode.c index 323af2746aa9..1dde30b98317 100644 renamed from drivers/gpu/drm/ast/ast_mode.c +0 -0

--- a/drivers/gpu/drm/ast/ast_mode.c

+++ b/drivers/gpu/drm/ast/ast_mode.c

@@ -310,7 +310,7 @@ static void ast_set_crtc_reg(struct ast_private *ast,

      u8 jreg05 = 0, jreg07 = 0, jreg09 = 0, jregAC = 0, jregAD = 0, jregAE = 0;

      u16 temp, precache = 0;

- if ((ast->chip == AST2500) &&

+ if (((ast->chip == AST2600) || (ast->chip == AST2500)) &&

          (vbios_mode->enh_table->flags & AST2500PreCatchCRT))

              precache = 40;

[KH]: after checking on register value, AST2600 doesn't run this.

@@ -428,7 +428,7 @@ static void ast_set_dclk_reg(struct ast_private *ast, {

      const struct ast_vbios_dclk_info *clk_info;

- if (ast->chip == AST2500)

+ if ((ast->chip == AST2600) || (ast->chip == AST2500))

              clk_info = &dclk_table_ast2500[vbios_mode->enh_table->dclk_index];

      else

              clk_info = &dclk_table[vbios_mode->enh_table->dclk_index];

[KH]: after checking on register value, AST2600 doesn't run this.

The difference between 2 table of " dclk_table_ast2500" and " dclk_table" are the setting of Reduce Blanking.

@@ -476,7 +476,7 @@ static void ast_set_crtthd_reg(struct ast_private *ast)

              ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0xe0);

              ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0xa0);

      } else if (ast->chip == AST2300 || ast->chip == AST2400 ||

- ast->chip == AST2500) {

+ ast->chip == AST2500 || ast->chip == AST2600) {

              ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa7, 0x78);

              ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa6, 0x60);

      } else if (ast->chip == AST2100 ||

[KH]: Yes, the patch is "drm/ast: Create threshold values for AST2600" that is the root cause of whites lines on AST2600
commit
bcc77411e8a65929655cef7b63a36000724cdc4b<https://cgit.freedesktop.org/drm/drm/commit/?id=bcc77411e8a65929655cef7b63a36000724cdc4b> (patch<https://cgit.freedesktop.org/drm/drm/patch/?id=bcc77411e8a65929655cef7b63a36000724cdc4b>)

drivers/gpu/drm/ast/ast_post.c index 0aa9cf0fb5c3..eb1ff9084034 100644 renamed from drivers/gpu/drm/ast/ast_post.c +0 -0

--- a/drivers/gpu/drm/ast/ast_post.c

+++ b/drivers/gpu/drm/ast/ast_post.c

@@ -80,7 +80,7 @@ ast_set_def_ext_reg(struct drm_device *dev)

              ast_set_index_reg(ast, AST_IO_CRTC_PORT, i, 0x00);

       if (ast->chip == AST2300 || ast->chip == AST2400 ||

- ast->chip == AST2500) {

+ ast->chip == AST2500 || ast->chip == AST2600) {

              if (pdev->revision >= 0x20)

                      ext_reg_info = extreginfo_ast2300;

              else

[KH]: after checking on register value, AST2600 doesn't run this.

@@ -105,7 +105,7 @@ ast_set_def_ext_reg(struct drm_device *dev)

      /* Enable RAMDAC for A1 */

      reg = 0x04;

      if (ast->chip == AST2300 || ast->chip == AST2400 ||

- ast->chip == AST2500)

+ ast->chip == AST2500 || ast->chip == AST2600)

              reg |= 0x20;

      ast_set_index_reg_mask(ast, AST_IO_CRTC_PORT, 0xb6, 0xff, reg);  }

[KH]: Yes, it should be a patch, because the value is set by GOP/VBIOS during booting.

ALTHOUGH, after checking on register value, AST2600 doesn't run this.

@@ -382,7 +382,7 @@ void ast_post_gpu(struct drm_device *dev)

      if (ast->chip == AST2600) {

              ast_dp_launch(dev, 1);

      } else if (ast->config_mode == ast_use_p2a) {

- if (ast->chip == AST2500)

+ if (ast->chip == AST2500 || ast->chip == AST2600)

                      ast_post_chip_2500(dev);

              else if (ast->chip == AST2300 || ast->chip == AST2400)

                      ast_post_chip_2300(dev);

[KH]: NO, AST2600 has its flow that is ast_dp_lauch()

And "ast_use_p2a" means using P2A bridge that is no longer allowed.

Again, Thanks all who works on drm/ast. Thank you!

Regards,

Kuo-Hsiang Chou

--

2.36.1

#4 lore raw
Quoted 5 lines · Kuo-Hsiang Chou

Hi Thomas

Thanks for your suggestions!

I answer each revision inline that followed by [KH]:.

Thanks for reviewing this.

Quoted 31 lines

Regards,

        Kuo-Hsiang Chou

-----Original Message-----

From: Thomas Zimmermann [mailto:[email protected]]

Sent: Tuesday, June 07, 2022 8:03 PM

To: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Kuo-Hsiang Chou <[email protected]>

Subject: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

[KH]: Based on CVE_2019_6260 item3, P2A is disallowed anymore.

P2A (PCIe to AMBA) is a bridge that is able to revise any BMC registers.

Yes, P2A is dangerous on security issue, because Host open a backdoor and someone malicious SW/APP will be easy to take control of BMC.

Therefore, P2A is disabled forever.

Now, return to this patch, there is no need to add AST2600 condition on the P2A flow.

[snip]

Quoted 8 lines

[KH]: Yes, the patch is "drm/ast: Create threshold values for AST2600" that is the root cause of whites lines on AST2600

commit

bcc77411e8a65929655cef7b63a36000724cdc4b <https://cgit.freedesktop.org/drm/drm/commit/?id=bcc77411e8a65929655cef7b63a36000724cdc4b> (patch <https://cgit.freedesktop.org/drm/drm/patch/?id=bcc77411e8a65929655cef7b63a36000724cdc4b>)

So basically this commit should be enough to fix the white lines and flickering with VGA output on AST2600 ? I will try to have it tested, and if it's good, we may want to have it on stable kernel.

Best regards,

Signature · 1 line
Jocelyn
#5 lore raw

Hi Jocelyn Falempe,

-----Original Message-----
From: Jocelyn Falempe [mailto:[email protected]]
Sent: Wednesday, June 08, 2022 9:17 PM
To: Kuo-Hsiang Chou <[email protected]>; Thomas Zimmermann <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]
Cc: [email protected]; [email protected]; Luke Chen <[email protected]>; Hungju Huang <[email protected]>; Charles Kuan <[email protected]>
Subject: Re: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Quoted 5 lines · Kuo-Hsiang Chou

Hi Thomas

Thanks for your suggestions!

I answer each revision inline that followed by [KH]:.

Thanks for reviewing this.

Quoted 40 lines

Regards,

        Kuo-Hsiang Chou

-----Original Message-----

From: Thomas Zimmermann [mailto:[email protected]]

Sent: Tuesday, June 07, 2022 8:03 PM

To: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Kuo-Hsiang Chou <[email protected]>

Subject: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

[KH]: Based on CVE_2019_6260 item3, P2A is disallowed anymore.

P2A (PCIe to AMBA) is a bridge that is able to revise any BMC registers.

Yes, P2A is dangerous on security issue, because Host open a backdoor and someone malicious SW/APP will be easy to take control of BMC.

Therefore, P2A is disabled forever.

Now, return to this patch, there is no need to add AST2600 condition on the P2A flow.

[snip]

Quoted 4 lines

[KH]: Yes, the patch is "drm/ast: Create threshold values for AST2600" that is the root cause of whites lines on AST2600

Quoted 9 lines

commit

bcc77411e8a65929655cef7b63a36000724cdc4b
<https://cgit.freedesktop.org/drm/drm/commit/?id=bcc77411e8a65929655ce
f7b63a36000724cdc4b> (patch
<https://cgit.freedesktop.org/drm/drm/patch/?id=bcc77411e8a65929655cef
7b63a36000724cdc4b>)

So basically this commit should be enough to fix the white lines and flickering with VGA output on AST2600 ? [KH]: Yes.
You are welcome to tell me something if you consider there is other strange issue. Thanks for your efforts on drm/ast project!
Regards,
Kuo-Hsiang Chou

I will try to have it tested, and if it's good, we may want to have it on stable kernel.

Best regards,

Signature · 1 line
Jocelyn
#6 lore raw
Quoted 74 lines · Kuo-Hsiang Chou

Hi Jocelyn Falempe,

-----Original Message-----
From: Jocelyn Falempe [mailto:[email protected]]
Sent: Wednesday, June 08, 2022 9:17 PM
To: Kuo-Hsiang Chou <[email protected]>; Thomas Zimmermann <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]
Cc: [email protected]; [email protected]; Luke Chen <[email protected]>; Hungju Huang <[email protected]>; Charles Kuan <[email protected]>
Subject: Re: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Quoted 5 lines · Kuo-Hsiang Chou

Hi Thomas

Thanks for your suggestions!

I answer each revision inline that followed by [KH]:.

Thanks for reviewing this.

Quoted 40 lines

Regards,

        Kuo-Hsiang Chou

-----Original Message-----

From: Thomas Zimmermann [mailto:[email protected]]

Sent: Tuesday, June 07, 2022 8:03 PM

To: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Kuo-Hsiang Chou <[email protected]>

Subject: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

[KH]: Based on CVE_2019_6260 item3, P2A is disallowed anymore.

P2A (PCIe to AMBA) is a bridge that is able to revise any BMC registers.

Yes, P2A is dangerous on security issue, because Host open a backdoor and someone malicious SW/APP will be easy to take control of BMC.

Therefore, P2A is disabled forever.

Now, return to this patch, there is no need to add AST2600 condition on the P2A flow.

[snip]

Quoted 4 lines

[KH]: Yes, the patch is "drm/ast: Create threshold values for AST2600" that is the root cause of whites lines on AST2600

Quoted 9 lines

commit

bcc77411e8a65929655cef7b63a36000724cdc4b
<https://cgit.freedesktop.org/drm/drm/commit/?id=bcc77411e8a65929655ce
f7b63a36000724cdc4b> (patch
<https://cgit.freedesktop.org/drm/drm/patch/?id=bcc77411e8a65929655cef
7b63a36000724cdc4b>)

Quoted 6 lines

So basically this commit should be enough to fix the white lines and flickering with VGA output on AST2600 ? [KH]: Yes.
You are welcome to tell me something if you consider there is other strange issue. Thanks for your efforts on drm/ast project!
Regards,
Kuo-Hsiang Chou

I've got confirmation that this commit is enough to fix the issue introduced with f9bd00e0ea9d.

So let's drop this patch, and submit bcc77411e8a6 "drm/ast: Create threshold values for AST2600" to stable kernel instead.

Thanks,

Quoted 5 lines

I will try to have it tested, and if it's good, we may want to have it on stable kernel.

Best regards,

author maintainer
#7 lore raw

Am 09.06.22 um 09:06 schrieb Jocelyn Falempe:

Quoted 82 lines · Kuo-Hsiang Chou
Quoted 74 lines

Hi Jocelyn Falempe,

-----Original Message-----
From: Jocelyn Falempe [mailto:[email protected]]
Sent: Wednesday, June 08, 2022 9:17 PM
To: Kuo-Hsiang Chou <[email protected]>; Thomas Zimmermann <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]
Cc: [email protected]; [email protected]; Luke Chen <[email protected]>; Hungju Huang <[email protected]>; Charles Kuan <[email protected]>
Subject: Re: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Quoted 5 lines · Kuo-Hsiang Chou

Hi Thomas

Thanks for your suggestions!

I answer each revision inline that followed by [KH]:.

Thanks for reviewing this.

Quoted 40 lines

Regards,

          Kuo-Hsiang Chou

-----Original Message-----

From: Thomas Zimmermann [mailto:[email protected]]

Sent: Tuesday, June 07, 2022 8:03 PM

To: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; Kuo-Hsiang Chou <[email protected]>

Subject: [PATCH] drm/ast: Treat AST2600 like AST2500 in most places

Include AST2600 in most of the branches for AST2500. Thereby revert most effects of commit f9bd00e0ea9d ("drm/ast: Create chip AST2600").

The AST2600 used to be treated like an AST2500, which at least gave usable display output. After introducing AST2600 in the driver without further updates, lots of functions take the wrong branches.

Handling AST2600 in the AST2500 branches reverts back to the original settings. The exception are cases where AST2600 meanwhile got its own branch.

[KH]: Based on CVE_2019_6260 item3, P2A is disallowed anymore.

P2A (PCIe to AMBA) is a bridge that is able to revise any BMC registers.

Yes, P2A is dangerous on security issue, because Host open a backdoor and someone malicious SW/APP will be easy to take control of BMC.

Therefore, P2A is disabled forever.

Now, return to this patch, there is no need to add AST2600 condition on the P2A flow.

[snip]

Quoted 4 lines

[KH]: Yes, the patch is "drm/ast: Create threshold values for AST2600" that is the root cause of whites lines on AST2600

Quoted 9 lines

commit

bcc77411e8a65929655cef7b63a36000724cdc4b
<https://cgit.freedesktop.org/drm/drm/commit/?id=bcc77411e8a65929655ce
f7b63a36000724cdc4b> (patch
<https://cgit.freedesktop.org/drm/drm/patch/?id=bcc77411e8a65929655cef
7b63a36000724cdc4b>)

Quoted 6 lines

So basically this commit should be enough to fix the white lines  and flickering with VGA output on AST2600 ? [KH]: Yes.
    You are welcome to tell me something if you consider there is other strange issue. Thanks for your efforts on drm/ast project!
Regards,
    Kuo-Hsiang Chou

Quoted 3 lines

I've got confirmation that this commit is enough to fix the issue introduced with f9bd00e0ea9d.

So let's drop this patch, and submit bcc77411e8a6 "drm/ast: Create threshold values for AST2600" to stable kernel instead.

Thanks to both of you for your help with this bug.

Best regards
Thomas

Quoted 9 lines

Thanks,

Quoted 5 lines

I will try to have it tested, and if it's good, we may want to have it on stable kernel.

Best regards,

Signature · 6 lines
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Ivo Totev