Skip to contentSkip to conversation

Does obmc-ikvm work with bmcweb?

DISCUSSION xiuzhi 5 participants 26 messages Mar 14, 2019 – Apr 16, 2019

Subsystems (4)
lore
Share & tools
lore

xiuzhi

author
#0 lore raw

Hi Jae, Ed,
I updated the source to openbmc latest version which uses bmcweb instead of phosphor-rest.
It seemed that ikvm cannot work on bmcweb.
The websit error:
app.bundle.js:43 Refused to connect to 'wss://192.168.200.132/kvm/0' because it violates the following Content Security Policy directive: "default-src 'self'". Note that 'connect-src' was not explicitly set, so 'default-src' is used as a fallback. Then I modify the bmcweb file include/security_headers_middleware.hpp L40:

        // The KVM currently needs to load images from base64 encoded strings.
        // img-src 'self' data: is used to allow that.
        // https://stackoverflow.com/questions/18447970/content-security-policy-data-not-working-for-base64-images-in-chrome-28
        res.addHeader("Content-Security-Policy",
                      "default-src 'self'; img-src 'self' data:");

to:
policy-data-not-working-for-base64-images-in-chrome-28

        res.addHeader("Content-Security-Policy",
                      "script-src 'self'; img-src 'self' data:");

The website erro:
"app.bundle.js:43 WebSocket on-error event"
"WebSocket connection to 'wss://192.168.200.132/kvmws' failed: WebSocket opening handshake timed out"
What is wrong with it?
Best,
Xiuzhi

#1 lore raw

Hi xiuzhi,

These are my repo information:
bmcweb: 9474b3788017bddd70e493e2b9b7674be30abc87
phosphor-webui: 511a2bbc5540fccb883b8d0005c2b387a23c9d05
Cherry picked
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-webui/+/10268/
libvncserver: 3348a7e42e86dfb98dd7458ad29def476cf6096f
obmc-ikvm: 2bc661d34abd1fda92a9d2b256ed88ca0e90d09a

I can see working KVM through https://my.bmc.ip.addr/#/server-control/kvm

-Jae

Quoted 26 lines · xiuzhi

Hi Jae, Ed,
    I updated the source to openbmc latest version which uses bmcweb
instead of phosphor-rest.
   It seemed that ikvm cannot work on bmcweb.
The websit error:
app.bundle.js:43 Refused to connect to 'wss://192.168.200.132/kvm/0' because it violates the following Content Security Policy directive: "default-src 'self'". Note that 'connect-src' was not explicitly set, so 'default-src' is used as a fallback. Then I modify the bmcweb file include/security_headers_middleware.hpp L40:
        // The KVM currently needs to load images from base64 encoded
strings.
        // img-src 'self' data: is used to allow that.
        //
https://stackoverflow.com/questions/18447970/content-security-policy-data-not-working-for-base64-images-in-chrome-28
        res.addHeader("Content-Security-Policy",
                      "default-src 'self'; img-src 'self' data:");
to:
policy-data-not-working-for-base64-images-in-chrome-28
        res.addHeader("Content-Security-Policy",
                      "script-src 'self'; img-src 'self' data:");

The website erro:
"app.bundle.js:43 WebSocket on-error event"
"WebSocket connection to 'wss://192.168.200.132/kvmws' failed: WebSocket opening handshake timed out"
  What is wrong with it?
   Best,
Xiuzhi

author
#2 lore raw

Hi Jae,
Thanks for giving me the source commit numbers related to the ikvm feature .
But It didn't work after I updated them.
Would you like to tell me your github.com/openbmc/openbmc commit number and the target machine of TEMPLATCONF?
Best,
Xiuzhi

Hi xiuzhi,

These are my repo information:
bmcweb: 9474b3788017bddd70e493e2b9b7674be30abc87
phosphor-webui: 511a2bbc5540fccb883b8d0005c2b387a23c9d05
Cherry picked
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-webui/+/10268/
libvncserver: 3348a7e42e86dfb98dd7458ad29def476cf6096f
obmc-ikvm: 2bc661d34abd1fda92a9d2b256ed88ca0e90d09a

I can see working KVM through https://my.bmc.ip.addr/#/server-control/kvm

-Jae

Quoted 30 lines · xiuzhi

Hi Jae, Ed,
I updated the source to openbmc latest version which uses bmcweb
instead of phosphor-rest.
It seemed that ikvm cannot work on bmcweb.
The websit error:
app.bundle.js:43 Refused to connect to 'wss://192.168.200.132/kvm/0' because it violates the following Content Security Policy directive:
"default-src 'self'". Note that 'connect-src' was not explicitly set, so 'default-src' is used as a fallback.
Then I modify the bmcweb file include/security_headers_middleware.hpp L40:

         // The KVM currently needs to load images from base64 encoded 
strings.
         // img-src 'self' data: is used to allow that.
         // 
https://stackoverflow.com/questions/18447970/content-security-policy-data-not-working-for-base64-images-in-chrome-28
         res.addHeader("Content-Security-Policy",
                       "default-src 'self'; img-src 'self' data:");
to:
policy-data-not-working-for-base64-images-in-chrome-28
         res.addHeader("Content-Security-Policy",
                       "script-src 'self'; img-src 'self' data:");

The website erro:
"app.bundle.js:43 WebSocket on-error event"
"WebSocket connection to 'wss://192.168.200.132/kvmws' failed: WebSocket opening handshake timed out"
What is wrong with it?
Best,
Xiuzhi

#3 lore raw

Hi Xiuzhi,

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

I'm using a custom build config for my machine. As you already know, for enabling the KVM feature, customization on kernel configs, adding a 'start-ipkvm' service, adding of libvncserver package and back-porting of aspeed video driver are also needed so there is no one-shot buildable TEMPLATECONF for KVM test.

Cheers,
Jae

Quoted 56 lines · xiuzhi

Hi Jae,
   Thanks for giving me the source  commit numbers related to the  ikvm
feature  .
But It didn't work after I updated them.
 Would you like to tell me your github.com/openbmc/openbmc commit
number and the target machine of  TEMPLATCONF?
Best,
Xiuzhi

Hi xiuzhi,

These are my repo information:
bmcweb: 9474b3788017bddd70e493e2b9b7674be30abc87
phosphor-webui: 511a2bbc5540fccb883b8d0005c2b387a23c9d05
                 Cherry picked
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-webui/+/10268/
libvncserver: 3348a7e42e86dfb98dd7458ad29def476cf6096f
obmc-ikvm: 2bc661d34abd1fda92a9d2b256ed88ca0e90d09a

I can see working KVM through https://my.bmc.ip.addr/#/server-control/kvm

-Jae

Quoted 10 lines · xiuzhi

Hi Jae, Ed,
      I updated the source to openbmc latest version which uses bmcweb instead of phosphor-rest.
     It seemed that ikvm cannot work on bmcweb.
The websit error:
app.bundle.js:43 Refused to connect to 'wss://192.168.200.132/kvm/0' because it violates the following Content Security Policy directive:
"default-src 'self'". Note that 'connect-src' was not explicitly set, so 'default-src' is used as a fallback.
Then I modify the bmcweb file include/security_headers_middleware.hpp

L40:

Quoted 5 lines

          // The KVM currently needs to load images from base64 encoded strings.
          // img-src 'self' data: is used to allow that.
          //

https://stackoverflow.com/questions/18447970/content-security-policy-data-not-working-for-base64-images-in-chrome-28

Quoted 15 lines

          res.addHeader("Content-Security-Policy", "default-src 'self'; img-src 'self' data:"); to:
policy-data-not-working-for-base64-images-in-chrome-28 res.addHeader("Content-Security-Policy", "script-src 'self'; img-src 'self' data:");

The website erro:
"app.bundle.js:43 WebSocket on-error event"
"WebSocket connection to 'wss://192.168.200.132/kvmws' failed: WebSocket opening handshake timed out"
    What is wrong with it?
     Best,
Xiuzhi

#4 lore raw
Quoted 7 lines

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

I'm using a custom build config for my machine. As you already know, for enabling the KVM feature, customization on kernel configs, adding a 'start-ipkvm' service, adding of libvncserver package and back-porting of aspeed video driver are also needed so there is no one-shot buildable TEMPLATECONF for KVM test.

Could you kindly put everything together on a branch so others could refer the branch to know what the changes are for KVM to work?
Thanks a lot!

#5 lore raw
Quoted 8 lines · Lei YU
Quoted 7 lines

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

I'm using a custom build config for my machine. As you already know, for enabling the KVM feature, customization on kernel configs, adding a 'start-ipkvm' service, adding of libvncserver package and back-porting of aspeed video driver are also needed so there is no one-shot buildable TEMPLATECONF for KVM test.

Quoted 4 lines

Could you kindly put everything together on a branch so others could refer the branch to know what the changes are for KVM to work?
Thanks a lot!

Well, I've already shared SHAs that are relating to KVM in this email thread, I'm sharing it again with some more details.

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

bmcweb: 9474b3788017bddd70e493e2b9b7674be30abc87

phosphor-webui: 511a2bbc5540fccb883b8d0005c2b387a23c9d05
Cherry picked
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-webui/+/10268/

libvncserver: 3348a7e42e86dfb98dd7458ad29def476cf6096f Need to remove 'x11' and 'gtk+' to reduce package size so I made:
libvncserver_%.bbappend
----------------------------------------------------------------
PACKAGECONFIG_remove = "gcrypt gnutls png sdl zlib"

TARGET_CXXFLAGS += " -Dflto"

do_install_append() {
    rm -rf ${D}${libdir}/libvncclient*
}

inherit cmake

# Use the latest to support obmc-ikvm DEPENDS += "openssl"
SRC_URI = "git://github.com/LibVNC/libvncserver" SRCREV = "3348a7e42e86dfb98dd7458ad29def476cf6096f" S = "${WORKDIR}/git"

# Remove x11 and gtk+ that cause big image size # Actually, these aren't needed to support obmc-ikvm REQUIRED_DISTRO_FEATURES_remove = "x11" DEPENDS_remove = "gtk+" RDEPENDS_${PN}_remove = "gtk+"
----------------------------------------------------------------

start-ipkvm: Cherry pick https://gerrit.openbmc-project.xyz/#/c/openbmc/meta-phosphor/+/13536/
After cherry picking it, I changed start-ipkvm.service file like below:
start-ipkvm.service
----------------------------------------------------------------
[Unit]
Description=OpenBMC ipKVM daemon
StopWhenUnneeded=false

[Service]
Restart=always
ExecStartPre=/usr/bin/create_usbhid.sh
ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1

[Install]
WantedBy=multi-user.target
-----------------------------------------------------------------

obmc-ikvm: 2bc661d34abd1fda92a9d2b256ed88ca0e90d09a

linux-aspeed: 7859d04be772a1d185de335a8e9d38ac03a89f3f Added these kernel configs:
-----------------------------------------------------------------
CONFIG_FRAME_VECTOR=y
CONFIG_MEDIA_SUPPORT=y CONFIG_MEDIA_CAMERA_SUPPORT=y CONFIG_VIDEO_DEV=y
CONFIG_VIDEO_V4L2=y CONFIG_V4L_PLATFORM_DRIVERS=y CONFIG_VIDEO_ASPEED=y
CONFIG_VIDEOBUF2_CORE=y CONFIG_VIDEOBUF2_V4L2=y CONFIG_VIDEOBUF2_MEMOPS=y CONFIG_VIDEOBUF2_DMA_CONTIG=y CONFIG_MEDIA_SUBDRV_AUTOSELECT=y CONFIG_USB_GADGET_VBUS_DRAW=2 CONFIG_USB_GADGET_STORAGE_NUM_BUFFERS=2 CONFIG_USB_LIBCOMPOSITE=y
CONFIG_USB_F_HID=y
CONFIG_USB_GADGET=y
CONFIG_U_SERIAL_CONSOLE=y CONFIG_USB_ASPEED_VHUB=y CONFIG_USB_CONFIGFS=y CONFIG_USB_CONFIGFS_MASS_STORAGE=y CONFIG_USB_CONFIGFS_F_FS=y CONFIG_USB_CONFIGFS_F_HID=y
-------------------------------------------------------------------
Back ported Eddie's aspeed-video driver

MAINTAINERS +8 -0
index b98a5763f6c1..cdd8f824b6da 100644
@@ -2358,6 +2358,14 @@ S:	Maintained
 F:	Documentation/hwmon/asc7621
 F:	drivers/hwmon/asc7621.c
 
+ASPEED VIDEO ENGINE DRIVER
+M:	Eddie James <[email protected]>
+L:	[email protected]
+L:	[email protected] (moderated for non-subscribers)
+S:	Maintained
+F:	drivers/media/platform/aspeed-video.c
+F:	Documentation/devicetree/bindings/media/aspeed-video.txt
+
 ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS
 M:	Corentin Chary <[email protected]>
 L:	[email protected]
arch/arm/boot/dts/aspeed-g5.dtsi +11 -0
index d8a9d20cc4e0..d6da0a29ced3 100644
@@ -243,6 +243,17 @@
 				interrupts = <0x19>;
 			};
 
+			video: video@1e700000 {
+				compatible = "aspeed,ast2500-video-engine";
+				reg = <0x1e700000 0x20000>;
+				clocks = <&syscon ASPEED_CLK_GATE_VCLK>,
+					 <&syscon ASPEED_CLK_GATE_ECLK>;
+				clock-names = "vclk", "eclk";
+				resets = <&syscon ASPEED_RESET_VIDEO>;
+				interrupts = <7>;
+				status = "disabled";
+			};
+
 			adc: adc@1e6e9000 {
 				compatible = "aspeed,ast2500-adc";
 				reg = <0x1e6e9000 0xb0>;
drivers/clk/clk-aspeed.c +1 -1
index 596136793fc4..f16ce7dbd6ab 100644
@@ -87,7 +87,7 @@ struct aspeed_clk_gate {
 /* TODO: ask Aspeed about the actual parent data */
 static const struct aspeed_gate_data aspeed_gates[] = {
 	/*				 clk rst   name			parent	flags */
-	[ASPEED_CLK_GATE_ECLK] =	{  0, -1, "eclk-gate",		"eclk",	0 }, /* Video Engine */
+	[ASPEED_CLK_GATE_ECLK] =	{  0,  6, "eclk-gate",		"eclk",	0 }, /* Video Engine */
 	[ASPEED_CLK_GATE_GCLK] =	{  1,  7, "gclk-gate",		NULL,	0 }, /* 2D 

engine */
[ASPEED_CLK_GATE_MCLK] = { 2, -1, "mclk-gate", "mpll",
CLK_IS_CRITICAL }, /* SDRAM */
[ASPEED_CLK_GATE_VCLK] = { 3, 6, "vclk-gate", NULL, 0 }, /* Video
Capture */
@@ -113,6 +113,24 @@ static const struct aspeed_gate_data aspeed_gates[] = {
[ASPEED_CLK_GATE_LHCCLK] = { 28, -1, "lhclk-gate", "lhclk", 0 }, /*
LPC master/LPC+ */
};

+static const char * const eclk_parent_names[] = {
+ "mpll",
+ "hpll",
+ "dpll",
+};
+
+static const struct clk_div_table ast2500_eclk_div_table[] = {
+ { 0x0, 2 },
+ { 0x1, 2 },
+ { 0x2, 3 },
+ { 0x3, 4 },
+ { 0x4, 5 },
+ { 0x5, 6 },
+ { 0x6, 7 },
+ { 0x7, 8 },
+ { 0 }
+};
+
static const struct clk_div_table ast2500_mac_div_table[] = {
{ 0x0, 4 }, /* Yep, really. Aspeed confirmed this is correct */
{ 0x1, 4 },
@@ -192,18 +210,21 @@ static struct clk_hw *aspeed_ast2500_calc_pll(const char *name, u32 val)

struct aspeed_clk_soc_data {
const struct clk_div_table *div_table;
+ const struct clk_div_table *eclk_div_table;
const struct clk_div_table *mac_div_table; struct clk_hw *(*calc_pll)(const char *name, u32 val);
};

static const struct aspeed_clk_soc_data ast2500_data = {
.div_table = ast2500_div_table,
+ .eclk_div_table = ast2500_eclk_div_table,
.mac_div_table = ast2500_mac_div_table, .calc_pll = aspeed_ast2500_calc_pll,
};

static const struct aspeed_clk_soc_data ast2400_data = {
.div_table = ast2400_div_table,
+ .eclk_div_table = ast2400_div_table,
.mac_div_table = ast2400_div_table, .calc_pll = aspeed_ast2400_calc_pll,
};
@@ -317,6 +338,7 @@ static const u8 aspeed_resets[] = {
[ASPEED_RESET_PECI] = 10,
[ASPEED_RESET_I2C] = 2,
[ASPEED_RESET_AHB] = 1,
+ [ASPEED_RESET_VIDEO] = 6,

/*
* SCUD4 resets start at an offset to separate them from
@@ -522,6 +544,22 @@ static int aspeed_clk_probe(struct platform_device *pdev)
return PTR_ERR(hw);
aspeed_clk_data->hws[ASPEED_CLK_24M] = hw;

+ hw = clk_hw_register_mux(dev, "eclk-mux", eclk_parent_names,
+ ARRAY_SIZE(eclk_parent_names), 0,
+ scu_base + ASPEED_CLK_SELECTION, 2, 0x3, 0,
+ &aspeed_clk_lock);
+ if (IS_ERR(hw))
+ return PTR_ERR(hw);
+ aspeed_clk_data->hws[ASPEED_CLK_ECLK_MUX] = hw;
+
+ hw = clk_hw_register_divider_table(dev, "eclk", "eclk-mux", 0,

+					   scu_base + ASPEED_CLK_SELECTION, 28,
+					   3, 0, soc_data->eclk_div_table,
+					   &aspeed_clk_lock);

+ if (IS_ERR(hw))
+ return PTR_ERR(hw);
+ aspeed_clk_data->hws[ASPEED_CLK_ECLK] = hw;
+
/*
* TODO: There are a number of clocks that not included in this driver
* as more information is required:
@@ -531,7 +569,6 @@ static int aspeed_clk_probe(struct platform_device *pdev)

 	 *   RGMII
 	 *   RMII
 	 *   UART[1..5] clock source mux
-	 *   Video Engine (ECLK) mux and clock divider

*/

for (i = 0; i < ARRAY_SIZE(aspeed_gates); i++) {

drivers/media/platform/Kconfig +9 -0
index 54fe90acb5b2..d6edf2d28f9b 100644
@@ -32,6 +32,15 @@ source "drivers/media/platform/davinci/Kconfig"
 
 source "drivers/media/platform/omap/Kconfig"
 
+config VIDEO_ASPEED
+	tristate "Aspeed AST2400 and AST2500 Video Engine driver"
+	depends on VIDEO_V4L2
+	select VIDEOBUF2_DMA_CONTIG
+	help
+	  Support for the Aspeed Video Engine (VE) embedded in the Aspeed
+	  AST2400 and AST2500 SOCs. The VE can capture and compress video data
+	  from digital or analog sources.
+
 config VIDEO_SH_VOU
 	tristate "SuperH VOU video output driver"
 	depends on MEDIA_CAMERA_SUPPORT
drivers/media/platform/Makefile +1 -0
index 41322ab65802..205c33a004fc 100644
@@ -3,6 +3,7 @@
 # Makefile for the video capture/playback device drivers.
 #
 
+obj-$(CONFIG_VIDEO_ASPEED)		+= aspeed-video.o
 obj-$(CONFIG_VIDEO_CADENCE)		+= cadence/
 obj-$(CONFIG_VIDEO_VIA_CAMERA) += via-camera.o
 obj-$(CONFIG_VIDEO_CAFE_CCIC) += marvell-ccic/
include/dt-bindings/clock/aspeed-clock.h +1 -0
index f43738607d77..15a9059d0303 100644
@@ -50,5 +50,6 @@
 #define ASPEED_RESET_I2C		7
 #define ASPEED_RESET_AHB		8
 #define ASPEED_RESET_CRT1		9
+#define ASPEED_RESET_VIDEO		10
 
 #endif

------------------------------------------------------------------

These are all I did for testing KVM.

Cheers,
Jae

#6 lore raw

Attaching a patch file for back porting of Eddie's aspeed-video driver. Diff in below doesn't include the newly created file so I'm attaching it as a patch.

-Jae

Quoted 296 lines · Jae Hyun Yoo
Quoted 8 lines · Lei YU
Quoted 7 lines

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

I'm using a custom build config for my machine. As you already know, for enabling the KVM feature, customization on kernel configs, adding a 'start-ipkvm' service, adding of libvncserver package and back-porting of aspeed video driver are also needed so there is no one-shot buildable TEMPLATECONF for KVM test.

Quoted 4 lines

Could you kindly put everything together on a branch so others could refer the branch to know what the changes are for KVM to work?
Thanks a lot!

Well, I've already shared SHAs that are relating to KVM in this email thread, I'm sharing it again with some more details.

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

bmcweb: 9474b3788017bddd70e493e2b9b7674be30abc87

phosphor-webui: 511a2bbc5540fccb883b8d0005c2b387a23c9d05
                Cherry picked
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-webui/+/10268/

libvncserver: 3348a7e42e86dfb98dd7458ad29def476cf6096f Need to remove 'x11' and 'gtk+' to reduce package size so I made:
libvncserver_%.bbappend
----------------------------------------------------------------
PACKAGECONFIG_remove = "gcrypt gnutls png sdl zlib"

TARGET_CXXFLAGS += " -Dflto"

do_install_append() {
    rm -rf ${D}${libdir}/libvncclient*
}

inherit cmake

# Use the latest to support obmc-ikvm DEPENDS += "openssl"
SRC_URI = "git://github.com/LibVNC/libvncserver" SRCREV = "3348a7e42e86dfb98dd7458ad29def476cf6096f" S = "${WORKDIR}/git"

# Remove x11 and gtk+ that cause big image size # Actually, these aren't needed to support obmc-ikvm REQUIRED_DISTRO_FEATURES_remove = "x11" DEPENDS_remove = "gtk+" RDEPENDS_${PN}_remove = "gtk+"
----------------------------------------------------------------

start-ipkvm: Cherry pick https://gerrit.openbmc-project.xyz/#/c/openbmc/meta-phosphor/+/13536/
After cherry picking it, I changed start-ipkvm.service file like below:
start-ipkvm.service
----------------------------------------------------------------
[Unit]
Description=OpenBMC ipKVM daemon
StopWhenUnneeded=false

[Service]
Restart=always
ExecStartPre=/usr/bin/create_usbhid.sh
ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1

[Install]
WantedBy=multi-user.target
-----------------------------------------------------------------

obmc-ikvm: 2bc661d34abd1fda92a9d2b256ed88ca0e90d09a

linux-aspeed: 7859d04be772a1d185de335a8e9d38ac03a89f3f Added these kernel configs:
-----------------------------------------------------------------
CONFIG_FRAME_VECTOR=y
CONFIG_MEDIA_SUPPORT=y CONFIG_MEDIA_CAMERA_SUPPORT=y CONFIG_VIDEO_DEV=y
CONFIG_VIDEO_V4L2=y CONFIG_V4L_PLATFORM_DRIVERS=y CONFIG_VIDEO_ASPEED=y
CONFIG_VIDEOBUF2_CORE=y CONFIG_VIDEOBUF2_V4L2=y CONFIG_VIDEOBUF2_MEMOPS=y CONFIG_VIDEOBUF2_DMA_CONTIG=y CONFIG_MEDIA_SUBDRV_AUTOSELECT=y CONFIG_USB_GADGET_VBUS_DRAW=2 CONFIG_USB_GADGET_STORAGE_NUM_BUFFERS=2 CONFIG_USB_LIBCOMPOSITE=y
CONFIG_USB_F_HID=y
CONFIG_USB_GADGET=y
CONFIG_U_SERIAL_CONSOLE=y CONFIG_USB_ASPEED_VHUB=y CONFIG_USB_CONFIGFS=y CONFIG_USB_CONFIGFS_MASS_STORAGE=y CONFIG_USB_CONFIGFS_F_FS=y CONFIG_USB_CONFIGFS_F_HID=y
-------------------------------------------------------------------
Back ported Eddie's aspeed-video driver

MAINTAINERS +8 -0
index b98a5763f6c1..cdd8f824b6da 100644
@@ -2358,6 +2358,14 @@ S:    Maintained
  F:    Documentation/hwmon/asc7621
  F:    drivers/hwmon/asc7621.c
 
+ASPEED VIDEO ENGINE DRIVER
+M:    Eddie James <[email protected]>
+L:    [email protected]
+L:    [email protected] (moderated for non-subscribers)
+S:    Maintained
+F:    drivers/media/platform/aspeed-video.c
+F:    Documentation/devicetree/bindings/media/aspeed-video.txt
+
  ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS
  M:    Corentin Chary <[email protected]>
  L:    [email protected]
arch/arm/boot/dts/aspeed-g5.dtsi +11 -0
index d8a9d20cc4e0..d6da0a29ced3 100644
@@ -243,6 +243,17 @@
                  interrupts = <0x19>;
              };
 
+            video: video@1e700000 {
+                compatible = "aspeed,ast2500-video-engine";
+                reg = <0x1e700000 0x20000>;
+                clocks = <&syscon ASPEED_CLK_GATE_VCLK>,
+                     <&syscon ASPEED_CLK_GATE_ECLK>;
+                clock-names = "vclk", "eclk";
+                resets = <&syscon ASPEED_RESET_VIDEO>;
+                interrupts = <7>;
+                status = "disabled";
+            };
+
              adc: adc@1e6e9000 {
                  compatible = "aspeed,ast2500-adc";
                  reg = <0x1e6e9000 0xb0>;
drivers/clk/clk-aspeed.c +0 -1
index 596136793fc4..f16ce7dbd6ab 100644
@@ -87,7 +87,7 @@ struct aspeed_clk_gate {
  /* TODO: ask Aspeed about the actual parent data */
  static const struct aspeed_gate_data aspeed_gates[] = {
      /*                 clk rst   name            parent    flags */
-    [ASPEED_CLK_GATE_ECLK] =    {  0, -1, "eclk-gate",        

"eclk",    0 }, /* Video Engine */
+    [ASPEED_CLK_GATE_ECLK] =    {  0,  6, "eclk-gate", "eclk",    0 }, /* Video Engine */
     [ASPEED_CLK_GATE_GCLK] =    {  1,  7, "gclk-gate",        NULL,
0 }, /* 2D engine */
     [ASPEED_CLK_GATE_MCLK] =    {  2, -1, "mclk-gate",        "mpll",
CLK_IS_CRITICAL }, /* SDRAM */
     [ASPEED_CLK_GATE_VCLK] =    {  3,  6, "vclk-gate",        NULL,
0 }, /* Video Capture */

drivers/clk/clk-aspeed.c +0 -0
@@ -113,6 +113,24 @@ static const struct aspeed_gate_data aspeed_gates[] 
 = {
      [ASPEED_CLK_GATE_LHCCLK] =    { 28, -1, "lhclk-gate",        

"lhclk", 0 }, /* LPC master/LPC+ */
 };

drivers/clk/clk-aspeed.c +21 -0
+static const char * const eclk_parent_names[] = {
+    "mpll",
+    "hpll",
+    "dpll",
+};
+
+static const struct clk_div_table ast2500_eclk_div_table[] = {
+    { 0x0, 2 },
+    { 0x1, 2 },
+    { 0x2, 3 },
+    { 0x3, 4 },
+    { 0x4, 5 },
+    { 0x5, 6 },
+    { 0x6, 7 },
+    { 0x7, 8 },
+    { 0 }
+};
+
  static const struct clk_div_table ast2500_mac_div_table[] = {
      { 0x0, 4 }, /* Yep, really. Aspeed confirmed this is correct */
      { 0x1, 4 },
@@ -192,18 +210,21 @@ static struct clk_hw 
 *aspeed_ast2500_calc_pll(const char *name, u32 val)
 
  struct aspeed_clk_soc_data {
      const struct clk_div_table *div_table;
+    const struct clk_div_table *eclk_div_table;
      const struct clk_div_table *mac_div_table;
      struct clk_hw *(*calc_pll)(const char *name, u32 val);
  };
 
  static const struct aspeed_clk_soc_data ast2500_data = {
      .div_table = ast2500_div_table,
+    .eclk_div_table = ast2500_eclk_div_table,
      .mac_div_table = ast2500_mac_div_table,
      .calc_pll = aspeed_ast2500_calc_pll,
  };
 
  static const struct aspeed_clk_soc_data ast2400_data = {
      .div_table = ast2400_div_table,
+    .eclk_div_table = ast2400_div_table,
      .mac_div_table = ast2400_div_table,
      .calc_pll = aspeed_ast2400_calc_pll,

 };

drivers/clk/clk-aspeed.c +17 -0
@@ -317,6 +338,7 @@ static const u8 aspeed_resets[] = {
      [ASPEED_RESET_PECI]    = 10,
      [ASPEED_RESET_I2C]    =  2,
      [ASPEED_RESET_AHB]    =  1,
+    [ASPEED_RESET_VIDEO]    =  6,
 
      /*
       * SCUD4 resets start at an offset to separate them from
@@ -522,6 +544,22 @@ static int aspeed_clk_probe(struct platform_device 
 *pdev)
          return PTR_ERR(hw);
      aspeed_clk_data->hws[ASPEED_CLK_24M] = hw;
 
+    hw = clk_hw_register_mux(dev, "eclk-mux", eclk_parent_names,
+                 ARRAY_SIZE(eclk_parent_names), 0,
+                 scu_base + ASPEED_CLK_SELECTION, 2, 0x3, 0,
+                 &aspeed_clk_lock);
+    if (IS_ERR(hw))
+        return PTR_ERR(hw);
+    aspeed_clk_data->hws[ASPEED_CLK_ECLK_MUX] = hw;
+
+    hw = clk_hw_register_divider_table(dev, "eclk", "eclk-mux", 0,
+                       scu_base + ASPEED_CLK_SELECTION, 28,
+                       3, 0, soc_data->eclk_div_table,
+                       &aspeed_clk_lock);
+    if (IS_ERR(hw))
+        return PTR_ERR(hw);
+    aspeed_clk_data->hws[ASPEED_CLK_ECLK] = hw;
+
      /*
       * TODO: There are a number of clocks that not included in this 

driver
      * as more information is required:

drivers/clk/clk-aspeed.c +0 -1
@@ -531,7 +569,6 @@ static int aspeed_clk_probe(struct platform_device 
 *pdev)
       *   RGMII
       *   RMII
       *   UART[1..5] clock source mux
-     *   Video Engine (ECLK) mux and clock divider
       */
 

     for (i = 0; i < ARRAY_SIZE(aspeed_gates); i++) {

drivers/media/platform/Kconfig +9 -0
index 54fe90acb5b2..d6edf2d28f9b 100644
@@ -32,6 +32,15 @@ source "drivers/media/platform/davinci/Kconfig"
 
  source "drivers/media/platform/omap/Kconfig"
 
+config VIDEO_ASPEED
+    tristate "Aspeed AST2400 and AST2500 Video Engine driver"
+    depends on VIDEO_V4L2
+    select VIDEOBUF2_DMA_CONTIG
+    help
+      Support for the Aspeed Video Engine (VE) embedded in the Aspeed
+      AST2400 and AST2500 SOCs. The VE can capture and compress video data
+      from digital or analog sources.
+
  config VIDEO_SH_VOU
      tristate "SuperH VOU video output driver"
      depends on MEDIA_CAMERA_SUPPORT
drivers/media/platform/Makefile +1 -0
index 41322ab65802..205c33a004fc 100644
@@ -3,6 +3,7 @@
  # Makefile for the video capture/playback device drivers.
  #
 
+obj-$(CONFIG_VIDEO_ASPEED)        += aspeed-video.o
  obj-$(CONFIG_VIDEO_CADENCE)        += cadence/
  obj-$(CONFIG_VIDEO_VIA_CAMERA) += via-camera.o
  obj-$(CONFIG_VIDEO_CAFE_CCIC) += marvell-ccic/
include/dt-bindings/clock/aspeed-clock.h +1 -1
index f43738607d77..15a9059d0303 100644
@@ -50,5 +50,6 @@
  #define ASPEED_RESET_I2C        7
  #define ASPEED_RESET_AHB        8
  #define ASPEED_RESET_CRT1        9
+#define ASPEED_RESET_VIDEO        10
 
  #endif
------------------------------------------------------------------
Quoted 4 lines

These are all I did for testing KVM.

Cheers,
Jae

From 58d948177be889762ec2277b7bfb8a60d02f445d Mon Sep 17 00:00:00 2001
From: Jae Hyun Yoo <[email protected]>
Date: Fri, 9 Nov 2018 11:32:27 -0800
Subject: [PATCH] Add Aspeed Video Engine Driver

The Video Engine (VE) embedded in the Aspeed AST2400 and AST2500 SOCs can capture and compress video data from digital or analog sources. With the Aspeed chip acting a service processor, the Video Engine can capture the host processor graphics output.

Add a V4L2 driver to capture video data and compress it to JPEG images. Make the video frames available through the V4L2 streaming interface.

---
 .../devicetree/bindings/media/aspeed-video.txt     |   26 +
 MAINTAINERS                                        |    8 +
 arch/arm/boot/dts/aspeed-g5.dtsi                   |   11 +
 drivers/clk/clk-aspeed.c                           |   41 +-
 drivers/media/platform/Kconfig                     |    9 +
 drivers/media/platform/Makefile                    |    1 +
 drivers/media/platform/aspeed-video.c              | 1729 ++++++++++++++++++++
 include/dt-bindings/clock/aspeed-clock.h           |    1 +
 8 files changed, 1824 insertions(+), 2 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/media/aspeed-video.txt
 create mode 100644 drivers/media/platform/aspeed-video.c
Documentation/devicetree/bindings/media/aspeed-video.txt new file +26 -0
new file mode 100644
index 000000000000..78b464ae2672
@@ -0,0 +1,26 @@
+* Device tree bindings for Aspeed Video Engine
+
+The Video Engine (VE) embedded in the Aspeed AST2400 and AST2500 SOCs can
+capture and compress video data from digital or analog sources.
+
+Required properties:
+ - compatible:		"aspeed,ast2400-video-engine" or
+			"aspeed,ast2500-video-engine"
+ - reg:			contains the offset and length of the VE memory region
+ - clocks:		clock specifiers for the syscon clocks associated with
+			the VE (ordering must match the clock-names property)
+ - clock-names:		"vclk" and "eclk"
+ - resets:		reset specifier for the syscon reset associated with
+			the VE
+ - interrupts:		the interrupt associated with the VE on this platform
+
+Example:
+
+video-engine@1e700000 {
+    compatible = "aspeed,ast2500-video-engine";
+    reg = <0x1e700000 0x20000>;
+    clocks = <&syscon ASPEED_CLK_GATE_VCLK>, <&syscon ASPEED_CLK_GATE_ECLK>;
+    clock-names = "vclk", "eclk";
+    resets = <&syscon ASPEED_RESET_VIDEO>;
+    interrupts = <7>;
+};
MAINTAINERS +8 -0
index b98a5763f6c1..cdd8f824b6da 100644
@@ -2358,6 +2358,14 @@ S:	Maintained
 F:	Documentation/hwmon/asc7621
 F:	drivers/hwmon/asc7621.c
 
+ASPEED VIDEO ENGINE DRIVER
+M:	Eddie James <[email protected]>
+L:	[email protected]
+L:	[email protected] (moderated for non-subscribers)
+S:	Maintained
+F:	drivers/media/platform/aspeed-video.c
+F:	Documentation/devicetree/bindings/media/aspeed-video.txt
+
 ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS
 M:	Corentin Chary <[email protected]>
 L:	[email protected]
arch/arm/boot/dts/aspeed-g5.dtsi +11 -0
index d8a9d20cc4e0..d6da0a29ced3 100644
@@ -243,6 +243,17 @@
 				interrupts = <0x19>;
 			};
 
+			video: video@1e700000 {
+				compatible = "aspeed,ast2500-video-engine";
+				reg = <0x1e700000 0x20000>;
+				clocks = <&syscon ASPEED_CLK_GATE_VCLK>,
+					 <&syscon ASPEED_CLK_GATE_ECLK>;
+				clock-names = "vclk", "eclk";
+				resets = <&syscon ASPEED_RESET_VIDEO>;
+				interrupts = <7>;
+				status = "disabled";
+			};
+
 			adc: adc@1e6e9000 {
 				compatible = "aspeed,ast2500-adc";
 				reg = <0x1e6e9000 0xb0>;
drivers/clk/clk-aspeed.c +39 -2

89 lines not on this page · raw message

drivers/media/platform/Kconfig +9 -0
index 54fe90acb5b2..d6edf2d28f9b 100644
@@ -32,6 +32,15 @@ source "drivers/media/platform/davinci/Kconfig"
 
 source "drivers/media/platform/omap/Kconfig"
 
+config VIDEO_ASPEED
+	tristate "Aspeed AST2400 and AST2500 Video Engine driver"
+	depends on VIDEO_V4L2
+	select VIDEOBUF2_DMA_CONTIG
+	help
+	  Support for the Aspeed Video Engine (VE) embedded in the Aspeed
+	  AST2400 and AST2500 SOCs. The VE can capture and compress video data
+	  from digital or analog sources.
+
 config VIDEO_SH_VOU
 	tristate "SuperH VOU video output driver"
 	depends on MEDIA_CAMERA_SUPPORT
drivers/media/platform/Makefile +1 -0
index 41322ab65802..205c33a004fc 100644
@@ -3,6 +3,7 @@
 # Makefile for the video capture/playback device drivers.
 #
 
+obj-$(CONFIG_VIDEO_ASPEED)		+= aspeed-video.o
 obj-$(CONFIG_VIDEO_CADENCE)		+= cadence/
 obj-$(CONFIG_VIDEO_VIA_CAMERA) += via-camera.o
 obj-$(CONFIG_VIDEO_CAFE_CCIC) += marvell-ccic/
drivers/media/platform/aspeed-video.c new file +1729 -0

1,729 lines not on this page · raw message

include/dt-bindings/clock/aspeed-clock.h +1 -0
index f43738607d77..15a9059d0303 100644
@@ -50,5 +50,6 @@
 #define ASPEED_RESET_I2C		7
 #define ASPEED_RESET_AHB		8
 #define ASPEED_RESET_CRT1		9
+#define ASPEED_RESET_VIDEO		10
 
 #endif
Signature · 1 line
2.7.4
author
#7 lore raw

Hi Joe,
The ikvm works now, thanks very much.
One issue, the right button of mouse operation has no response on bmc ikvm website page. Best,
xiuzhi

------------------ Original ------------------
From: "jae.hyun.yoo";<[email protected]>;
Send time: Tuesday, Mar 19, 2019 1:33 AM
To: "Lei YU"<[email protected]>;
Cc: "xiuzhi"<[email protected]>; "openbmc"<[email protected]>; "ed"<[email protected]>; "Eddie James"<[email protected]>;
Subject: Re: Does obmc-ikvm work with bmcweb?

Attaching a patch file for back porting of Eddie's aspeed-video driver. Diff in below doesn't include the newly created file so I'm attaching it as a patch.

-Jae

Quoted 307 lines · Jae Hyun Yoo
Quoted 8 lines · Lei YU
Quoted 7 lines

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

I'm using a custom build config for my machine. As you already know, for enabling the KVM feature, customization on kernel configs, adding a 'start-ipkvm' service, adding of libvncserver package and back-porting of aspeed video driver are also needed so there is no one-shot buildable TEMPLATECONF for KVM test.

Quoted 5 lines

Could you kindly put everything together on a branch so others could refer the
branch to know what the changes are for KVM to work? Thanks a lot!

Well, I've already shared SHAs that are relating to KVM in this email thread, I'm sharing it again with some more details.

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

bmcweb: 9474b3788017bddd70e493e2b9b7674be30abc87

phosphor-webui: 511a2bbc5540fccb883b8d0005c2b387a23c9d05
Cherry picked
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-webui/+/10268/

libvncserver: 3348a7e42e86dfb98dd7458ad29def476cf6096f Need to remove 'x11' and 'gtk+' to reduce package size so I made:
libvncserver_%.bbappend
----------------------------------------------------------------
PACKAGECONFIG_remove = "gcrypt gnutls png sdl zlib"

TARGET_CXXFLAGS += " -Dflto"

do_install_append() {
     rm -rf ${D}${libdir}/libvncclient*
}

inherit cmake

# Use the latest to support obmc-ikvm DEPENDS += "openssl"
SRC_URI = "git://github.com/LibVNC/libvncserver" SRCREV = "3348a7e42e86dfb98dd7458ad29def476cf6096f" S = "${WORKDIR}/git"

# Remove x11 and gtk+ that cause big image size # Actually, these aren't needed to support obmc-ikvm REQUIRED_DISTRO_FEATURES_remove = "x11" DEPENDS_remove = "gtk+" RDEPENDS_${PN}_remove = "gtk+"
----------------------------------------------------------------

start-ipkvm: Cherry pick
https://gerrit.openbmc-project.xyz/#/c/openbmc/meta-phosphor/+/13536/
After cherry picking it, I changed start-ipkvm.service file like below:
start-ipkvm.service
----------------------------------------------------------------
[Unit]
Description=OpenBMC ipKVM daemon
StopWhenUnneeded=false

[Service]
Restart=always
ExecStartPre=/usr/bin/create_usbhid.sh ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1

[Install]
WantedBy=multi-user.target
-----------------------------------------------------------------

obmc-ikvm: 2bc661d34abd1fda92a9d2b256ed88ca0e90d09a

linux-aspeed: 7859d04be772a1d185de335a8e9d38ac03a89f3f Added these kernel configs:
-----------------------------------------------------------------
CONFIG_FRAME_VECTOR=y
CONFIG_MEDIA_SUPPORT=y CONFIG_MEDIA_CAMERA_SUPPORT=y CONFIG_VIDEO_DEV=y
CONFIG_VIDEO_V4L2=y CONFIG_V4L_PLATFORM_DRIVERS=y CONFIG_VIDEO_ASPEED=y
CONFIG_VIDEOBUF2_CORE=y CONFIG_VIDEOBUF2_V4L2=y CONFIG_VIDEOBUF2_MEMOPS=y CONFIG_VIDEOBUF2_DMA_CONTIG=y CONFIG_MEDIA_SUBDRV_AUTOSELECT=y CONFIG_USB_GADGET_VBUS_DRAW=2 CONFIG_USB_GADGET_STORAGE_NUM_BUFFERS=2 CONFIG_USB_LIBCOMPOSITE=y
CONFIG_USB_F_HID=y
CONFIG_USB_GADGET=y
CONFIG_U_SERIAL_CONSOLE=y CONFIG_USB_ASPEED_VHUB=y CONFIG_USB_CONFIGFS=y CONFIG_USB_CONFIGFS_MASS_STORAGE=y CONFIG_USB_CONFIGFS_F_FS=y CONFIG_USB_CONFIGFS_F_HID=y
-------------------------------------------------------------------
Back ported Eddie's aspeed-video driver

MAINTAINERS +8 -0
index b98a5763f6c1..cdd8f824b6da 100644
@@ -2358,6 +2358,14 @@ S:    Maintained
  F:    Documentation/hwmon/asc7621
  F:    drivers/hwmon/asc7621.c
 
+ASPEED VIDEO ENGINE DRIVER
+M:    Eddie James <[email protected]>
+L:    [email protected]
+L:    [email protected] (moderated for non-subscribers)
+S:    Maintained
+F:    drivers/media/platform/aspeed-video.c
+F:    Documentation/devicetree/bindings/media/aspeed-video.txt
+
  ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS
  M:    Corentin Chary <[email protected]>
  L:    [email protected]
arch/arm/boot/dts/aspeed-g5.dtsi +11 -0
index d8a9d20cc4e0..d6da0a29ced3 100644
@@ -243,6 +243,17 @@
                  interrupts = <0x19>;
              };
 
+            video: video@1e700000 {
+                compatible = "aspeed,ast2500-video-engine";
+                reg = <0x1e700000 0x20000>;
+                clocks = <&syscon ASPEED_CLK_GATE_VCLK>,
+                     <&syscon ASPEED_CLK_GATE_ECLK>;
+                clock-names = "vclk", "eclk";
+                resets = <&syscon ASPEED_RESET_VIDEO>;
+                interrupts = <7>;
+                status = "disabled";
+            };
+
              adc: adc@1e6e9000 {
                  compatible = "aspeed,ast2500-adc";
                  reg = <0x1e6e9000 0xb0>;
drivers/clk/clk-aspeed.c +0 -1
index 596136793fc4..f16ce7dbd6ab 100644
@@ -87,7 +87,7 @@ struct aspeed_clk_gate {
  /* TODO: ask Aspeed about the actual parent data */
  static const struct aspeed_gate_data aspeed_gates[] = {
      /*                 clk rst   name            parent    flags */
-    [ASPEED_CLK_GATE_ECLK] =    {  0, -1, "eclk-gate",        

"eclk", 0 }, /* Video Engine */
+ [ASPEED_CLK_GATE_ECLK] = { 0, 6, "eclk-gate",
"eclk", 0 }, /* Video Engine */

      [ASPEED_CLK_GATE_GCLK] =    {  1,  7, "gclk-gate",        NULL,    
0 }, /* 2D engine */
      [ASPEED_CLK_GATE_MCLK] =    {  2, -1, "mclk-gate",        "mpll", 
CLK_IS_CRITICAL }, /* SDRAM */
      [ASPEED_CLK_GATE_VCLK] =    {  3,  6, "vclk-gate",        NULL,    

0 }, /* Video Capture */

drivers/clk/clk-aspeed.c +0 -0
@@ -113,6 +113,24 @@ static const struct aspeed_gate_data aspeed_gates[] 
 = {
      [ASPEED_CLK_GATE_LHCCLK] =    { 28, -1, "lhclk-gate",        

"lhclk", 0 }, /* LPC master/LPC+ */
};

drivers/clk/clk-aspeed.c +21 -0
+static const char * const eclk_parent_names[] = {
+    "mpll",
+    "hpll",
+    "dpll",
+};
+
+static const struct clk_div_table ast2500_eclk_div_table[] = {
+    { 0x0, 2 },
+    { 0x1, 2 },
+    { 0x2, 3 },
+    { 0x3, 4 },
+    { 0x4, 5 },
+    { 0x5, 6 },
+    { 0x6, 7 },
+    { 0x7, 8 },
+    { 0 }
+};
+
  static const struct clk_div_table ast2500_mac_div_table[] = {
      { 0x0, 4 }, /* Yep, really. Aspeed confirmed this is correct */
      { 0x1, 4 },
@@ -192,18 +210,21 @@ static struct clk_hw 
 *aspeed_ast2500_calc_pll(const char *name, u32 val)
 
  struct aspeed_clk_soc_data {
      const struct clk_div_table *div_table;
+    const struct clk_div_table *eclk_div_table;
      const struct clk_div_table *mac_div_table;
      struct clk_hw *(*calc_pll)(const char *name, u32 val);
  };
 
  static const struct aspeed_clk_soc_data ast2500_data = {
      .div_table = ast2500_div_table,
+    .eclk_div_table = ast2500_eclk_div_table,
      .mac_div_table = ast2500_mac_div_table,
      .calc_pll = aspeed_ast2500_calc_pll,
  };
 
  static const struct aspeed_clk_soc_data ast2400_data = {
      .div_table = ast2400_div_table,
+    .eclk_div_table = ast2400_div_table,
      .mac_div_table = ast2400_div_table,
      .calc_pll = aspeed_ast2400_calc_pll,
  };
drivers/clk/clk-aspeed.c +17 -0
@@ -317,6 +338,7 @@ static const u8 aspeed_resets[] = {
      [ASPEED_RESET_PECI]    = 10,
      [ASPEED_RESET_I2C]    =  2,
      [ASPEED_RESET_AHB]    =  1,
+    [ASPEED_RESET_VIDEO]    =  6,
 
      /*
       * SCUD4 resets start at an offset to separate them from
@@ -522,6 +544,22 @@ static int aspeed_clk_probe(struct platform_device 
 *pdev)
          return PTR_ERR(hw);
      aspeed_clk_data->hws[ASPEED_CLK_24M] = hw;
 
+    hw = clk_hw_register_mux(dev, "eclk-mux", eclk_parent_names,
+                 ARRAY_SIZE(eclk_parent_names), 0,
+                 scu_base + ASPEED_CLK_SELECTION, 2, 0x3, 0,
+                 &aspeed_clk_lock);
+    if (IS_ERR(hw))
+        return PTR_ERR(hw);
+    aspeed_clk_data->hws[ASPEED_CLK_ECLK_MUX] = hw;
+
+    hw = clk_hw_register_divider_table(dev, "eclk", "eclk-mux", 0,
+                       scu_base + ASPEED_CLK_SELECTION, 28,
+                       3, 0, soc_data->eclk_div_table,
+                       &aspeed_clk_lock);
+    if (IS_ERR(hw))
+        return PTR_ERR(hw);
+    aspeed_clk_data->hws[ASPEED_CLK_ECLK] = hw;
+
      /*
       * TODO: There are a number of clocks that not included in this 

driver
* as more information is required:

drivers/clk/clk-aspeed.c +0 -1
@@ -531,7 +569,6 @@ static int aspeed_clk_probe(struct platform_device 
 *pdev)
       *   RGMII
       *   RMII
       *   UART[1..5] clock source mux
-     *   Video Engine (ECLK) mux and clock divider
       */
 
      for (i = 0; i < ARRAY_SIZE(aspeed_gates); i++) {
drivers/media/platform/Kconfig +9 -0
index 54fe90acb5b2..d6edf2d28f9b 100644
@@ -32,6 +32,15 @@ source "drivers/media/platform/davinci/Kconfig"
 
  source "drivers/media/platform/omap/Kconfig"
 
+config VIDEO_ASPEED
+    tristate "Aspeed AST2400 and AST2500 Video Engine driver"
+    depends on VIDEO_V4L2
+    select VIDEOBUF2_DMA_CONTIG
+    help
+      Support for the Aspeed Video Engine (VE) embedded in the Aspeed
+      AST2400 and AST2500 SOCs. The VE can capture and compress video data
+      from digital or analog sources.
+
  config VIDEO_SH_VOU
      tristate "SuperH VOU video output driver"
      depends on MEDIA_CAMERA_SUPPORT
drivers/media/platform/Makefile +1 -0
index 41322ab65802..205c33a004fc 100644
@@ -3,6 +3,7 @@
  # Makefile for the video capture/playback device drivers.
  #
 
+obj-$(CONFIG_VIDEO_ASPEED)        += aspeed-video.o
  obj-$(CONFIG_VIDEO_CADENCE)        += cadence/
  obj-$(CONFIG_VIDEO_VIA_CAMERA) += via-camera.o
  obj-$(CONFIG_VIDEO_CAFE_CCIC) += marvell-ccic/
include/dt-bindings/clock/aspeed-clock.h +1 -1
index f43738607d77..15a9059d0303 100644
@@ -50,5 +50,6 @@
  #define ASPEED_RESET_I2C        7
  #define ASPEED_RESET_AHB        8
  #define ASPEED_RESET_CRT1        9
+#define ASPEED_RESET_VIDEO        10
 
  #endif
------------------------------------------------------------------

These are all I did for testing KVM.

Cheers,
Jae

#9 lore raw
In reply to Jae Hyun Yoo

Hi Jae,

Thanks for sharing the detailed information.
I managed to do a build on Romulus and see the KVM partially working:
* I can see the video output in the KVM page;
* I am unable to do any operation, e.g. pressing a key does not seem to be sent to host
* When I browse to another page and browse back to KVM page, it displays a small black block, instead of showing correct KVM view.

Did you meet such issue before and do you happen to know how to fix this?

Thanks a lot!

Quoted 6 lines · Jae Hyun Yoo

Attaching a patch file for back porting of Eddie's aspeed-video driver. Diff in below doesn't include the newly created file so I'm attaching it as a patch.

-Jae

#8 lore raw
In reply to xiuzhi

Hi Xiuzhi,

Good to hear that. :)
For the mouse button mapping issue, try this patch:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19393

Cheers,
Jae

Quoted 331 lines · xiuzhi

Hi Joe,
   The ikvm works now, thanks very much.
One issue, the right button of mouse operation has no response on bmc ikvm website page. Best,
xiuzhi

------------------ Original ------------------
*From: * "jae.hyun.yoo";<[email protected]>;
*Send time:* Tuesday, Mar 19, 2019 1:33 AM
*To:* "Lei YU"<[email protected]>;
*Cc:* "xiuzhi"<[email protected]>; "openbmc"<[email protected]>; "ed"<[email protected]>; "Eddie James"<[email protected]>; *Subject: * Re: Does obmc-ikvm work with bmcweb?

Attaching a patch file for back porting of Eddie's aspeed-video driver. Diff in below doesn't include the newly created file so I'm attaching it as a patch.

-Jae

Quoted 4 lines · Jae Hyun Yoo
Quoted 3 lines · Lei YU
Quoted 3 lines

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

I'm using a custom build config for my machine. As you already

know, for

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

enabling the KVM feature, customization on kernel configs, adding a 'start-ipkvm' service, adding of libvncserver package and back-porting of aspeed video driver are also needed so there is no one-shot

buildable

Quoted 257 lines
Quoted 7 lines
Quoted 1 line

TEMPLATECONF for KVM test.

Could you kindly put everything together on a branch so others could refer the
branch to know what the changes are for KVM to work? Thanks a lot!

Well, I've already shared SHAs that are relating to KVM in this email thread, I'm sharing it again with some more details.

openbmc: ae29d00f227e55192a69f66c87633e1b53fc6e68

bmcweb: 9474b3788017bddd70e493e2b9b7674be30abc87

phosphor-webui: 511a2bbc5540fccb883b8d0005c2b387a23c9d05 Cherry picked
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-webui/+/10268/

libvncserver: 3348a7e42e86dfb98dd7458ad29def476cf6096f Need to remove 'x11' and 'gtk+' to reduce package size so I made:
libvncserver_%.bbappend
----------------------------------------------------------------
PACKAGECONFIG_remove = "gcrypt gnutls png sdl zlib"

TARGET_CXXFLAGS += " -Dflto"

do_install_append() {
      rm -rf ${D}${libdir}/libvncclient*
}

inherit cmake

# Use the latest to support obmc-ikvm DEPENDS += "openssl"
SRC_URI = "git://github.com/LibVNC/libvncserver" SRCREV = "3348a7e42e86dfb98dd7458ad29def476cf6096f" S = "${WORKDIR}/git"

# Remove x11 and gtk+ that cause big image size # Actually, these aren't needed to support obmc-ikvm REQUIRED_DISTRO_FEATURES_remove = "x11" DEPENDS_remove = "gtk+" RDEPENDS_${PN}_remove = "gtk+"
----------------------------------------------------------------

start-ipkvm: Cherry pick
https://gerrit.openbmc-project.xyz/#/c/openbmc/meta-phosphor/+/13536/
After cherry picking it, I changed start-ipkvm.service file like below:
start-ipkvm.service
----------------------------------------------------------------
[Unit]
Description=OpenBMC ipKVM daemon
StopWhenUnneeded=false

[Service]
Restart=always
ExecStartPre=/usr/bin/create_usbhid.sh ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1

[Install]
WantedBy=multi-user.target
-----------------------------------------------------------------

obmc-ikvm: 2bc661d34abd1fda92a9d2b256ed88ca0e90d09a

linux-aspeed: 7859d04be772a1d185de335a8e9d38ac03a89f3f Added these kernel configs:
-----------------------------------------------------------------
CONFIG_FRAME_VECTOR=y
CONFIG_MEDIA_SUPPORT=y CONFIG_MEDIA_CAMERA_SUPPORT=y CONFIG_VIDEO_DEV=y
CONFIG_VIDEO_V4L2=y CONFIG_V4L_PLATFORM_DRIVERS=y CONFIG_VIDEO_ASPEED=y
CONFIG_VIDEOBUF2_CORE=y CONFIG_VIDEOBUF2_V4L2=y CONFIG_VIDEOBUF2_MEMOPS=y CONFIG_VIDEOBUF2_DMA_CONTIG=y CONFIG_MEDIA_SUBDRV_AUTOSELECT=y CONFIG_USB_GADGET_VBUS_DRAW=2 CONFIG_USB_GADGET_STORAGE_NUM_BUFFERS=2 CONFIG_USB_LIBCOMPOSITE=y
CONFIG_USB_F_HID=y
CONFIG_USB_GADGET=y
CONFIG_U_SERIAL_CONSOLE=y CONFIG_USB_ASPEED_VHUB=y CONFIG_USB_CONFIGFS=y CONFIG_USB_CONFIGFS_MASS_STORAGE=y CONFIG_USB_CONFIGFS_F_FS=y CONFIG_USB_CONFIGFS_F_HID=y
-------------------------------------------------------------------
Back ported Eddie's aspeed-video driver

MAINTAINERS +0 -0
index b98a5763f6c1..cdd8f824b6da 100644
@@ -2358,6 +2358,14 @@ S:    Maintained

   F:    Documentation/hwmon/asc7621
   F:    drivers/hwmon/asc7621.c

MAINTAINERS +8 -0
+ASPEED VIDEO ENGINE DRIVER
+M:    Eddie James <[email protected]>
+L:    [email protected]
+L:    [email protected] (moderated for non-subscribers)
+S:    Maintained
+F:    drivers/media/platform/aspeed-video.c
+F:    Documentation/devicetree/bindings/media/aspeed-video.txt
+

   ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS M:    Corentin Chary <[email protected]> L:    [email protected]

arch/arm/boot/dts/aspeed-g5.dtsi +11 -0
index d8a9d20cc4e0..d6da0a29ced3 100644
@@ -243,6 +243,17 @@
                    interrupts = <0x19>;
                };
 
+            video: video@1e700000 {
+                compatible = "aspeed,ast2500-video-engine";
+                reg = <0x1e700000 0x20000>;
+                clocks = <&syscon ASPEED_CLK_GATE_VCLK>,
+                     <&syscon ASPEED_CLK_GATE_ECLK>;
+                clock-names = "vclk", "eclk";
+                resets = <&syscon ASPEED_RESET_VIDEO>;
+                interrupts = <7>;
+                status = "disabled";
+            };
+
                adc: adc@1e6e9000 {
                    compatible = "aspeed,ast2500-adc";
                    reg = <0x1e6e9000 0xb0>;
drivers/clk/clk-aspeed.c +0 -1
index 596136793fc4..f16ce7dbd6ab 100644
@@ -87,7 +87,7 @@ struct aspeed_clk_gate {
    /* TODO: ask Aspeed about the actual parent data */
    static const struct aspeed_gate_data aspeed_gates[] = {
        /*                 clk rst   name            parent    flags */
-    [ASPEED_CLK_GATE_ECLK] =    {  0, -1, "eclk-gate",

"eclk",    0 }, /* Video Engine */
+    [ASPEED_CLK_GATE_ECLK] =    {  0,  6, "eclk-gate", "eclk",    0 }, /* Video Engine */
       [ASPEED_CLK_GATE_GCLK] =    {  1,  7, "gclk-gate",        NULL, 0 }, /* 2D engine */
       [ASPEED_CLK_GATE_MCLK] =    {  2, -1, "mclk-gate",        "mpll", CLK_IS_CRITICAL }, /* SDRAM */
       [ASPEED_CLK_GATE_VCLK] =    {  3,  6, "vclk-gate",        NULL, 0 }, /* Video Capture */

drivers/clk/clk-aspeed.c +0 -0
@@ -113,6 +113,24 @@ static const struct aspeed_gate_data aspeed_gates[]
 = {
        [ASPEED_CLK_GATE_LHCCLK] =    { 28, -1, "lhclk-gate",

"lhclk", 0 }, /* LPC master/LPC+ */
   };

drivers/clk/clk-aspeed.c +18 -0
+static const char * const eclk_parent_names[] = {
+    "mpll",
+    "hpll",
+    "dpll",
+};
+
+static const struct clk_div_table ast2500_eclk_div_table[] = {
+    { 0x0, 2 },
+    { 0x1, 2 },
+    { 0x2, 3 },
+    { 0x3, 4 },
+    { 0x4, 5 },
+    { 0x5, 6 },
+    { 0x6, 7 },
+    { 0x7, 8 },
+    { 0 }
+};
+

   static const struct clk_div_table ast2500_mac_div_table[] = { { 0x0, 4 }, /* Yep, really. Aspeed confirmed this is correct */ { 0x1, 4 },

drivers/clk/clk-aspeed.c +3 -0
@@ -192,18 +210,21 @@ static struct clk_hw
 *aspeed_ast2500_calc_pll(const char *name, u32 val)
 
    struct aspeed_clk_soc_data {
        const struct clk_div_table *div_table;
+    const struct clk_div_table *eclk_div_table;
        const struct clk_div_table *mac_div_table;
        struct clk_hw *(*calc_pll)(const char *name, u32 val);
    };
 
    static const struct aspeed_clk_soc_data ast2500_data = {
        .div_table = ast2500_div_table,
+    .eclk_div_table = ast2500_eclk_div_table,
        .mac_div_table = ast2500_mac_div_table,
        .calc_pll = aspeed_ast2500_calc_pll,
    };
 
    static const struct aspeed_clk_soc_data ast2400_data = {
        .div_table = ast2400_div_table,
+    .eclk_div_table = ast2400_div_table,
        .mac_div_table = ast2400_div_table,
        .calc_pll = aspeed_ast2400_calc_pll,

   };

drivers/clk/clk-aspeed.c +17 -0
@@ -317,6 +338,7 @@ static const u8 aspeed_resets[] = {
        [ASPEED_RESET_PECI]    = 10,
        [ASPEED_RESET_I2C]    =  2,
        [ASPEED_RESET_AHB]    =  1,
+    [ASPEED_RESET_VIDEO]    =  6,
 
        /*
         * SCUD4 resets start at an offset to separate them from
@@ -522,6 +544,22 @@ static int aspeed_clk_probe(struct platform_device
 *pdev)
            return PTR_ERR(hw);
        aspeed_clk_data->hws[ASPEED_CLK_24M] = hw;
 
+    hw = clk_hw_register_mux(dev, "eclk-mux", eclk_parent_names,
+                 ARRAY_SIZE(eclk_parent_names), 0,
+                 scu_base + ASPEED_CLK_SELECTION, 2, 0x3, 0,
+                 &aspeed_clk_lock);
+    if (IS_ERR(hw))
+        return PTR_ERR(hw);
+    aspeed_clk_data->hws[ASPEED_CLK_ECLK_MUX] = hw;
+
+    hw = clk_hw_register_divider_table(dev, "eclk", "eclk-mux", 0,
+                       scu_base + ASPEED_CLK_SELECTION, 28,
+                       3, 0, soc_data->eclk_div_table,
+                       &aspeed_clk_lock);
+    if (IS_ERR(hw))
+        return PTR_ERR(hw);
+    aspeed_clk_data->hws[ASPEED_CLK_ECLK] = hw;
+
        /*
         * TODO: There are a number of clocks that not included in this

driver
        * as more information is required:

drivers/clk/clk-aspeed.c +0 -1
@@ -531,7 +569,6 @@ static int aspeed_clk_probe(struct platform_device
 *pdev)
         *   RGMII
         *   RMII
         *   UART[1..5] clock source mux
-     *   Video Engine (ECLK) mux and clock divider
         */
 

       for (i = 0; i < ARRAY_SIZE(aspeed_gates); i++) {

drivers/media/platform/Kconfig +0 -0
index 54fe90acb5b2..d6edf2d28f9b 100644
@@ -32,6 +32,15 @@ source "drivers/media/platform/davinci/Kconfig"
 

   source "drivers/media/platform/omap/Kconfig"

Quoted 7 lines

+config VIDEO_ASPEED
+    tristate "Aspeed AST2400 and AST2500 Video Engine driver" +    depends on VIDEO_V4L2
+    select VIDEOBUF2_DMA_CONTIG
+    help
+      Support for the Aspeed Video Engine (VE) embedded in the Aspeed +      AST2400 and AST2500 SOCs. The VE can capture and compress

video data

Quoted 36 lines

+      from digital or analog sources.
+
   config VIDEO_SH_VOU
       tristate "SuperH VOU video output driver"
       depends on MEDIA_CAMERA_SUPPORT

drivers/media/platform/Makefile +1 -0
index 41322ab65802..205c33a004fc 100644
@@ -3,6 +3,7 @@
    # Makefile for the video capture/playback device drivers.
    #
 
+obj-$(CONFIG_VIDEO_ASPEED)        += aspeed-video.o

   obj-$(CONFIG_VIDEO_CADENCE)        += cadence/
   obj-$(CONFIG_VIDEO_VIA_CAMERA) += via-camera.o
   obj-$(CONFIG_VIDEO_CAFE_CCIC) += marvell-ccic/

include/dt-bindings/clock/aspeed-clock.h +1 -1
index f43738607d77..15a9059d0303 100644
@@ -50,5 +50,6 @@
    #define ASPEED_RESET_I2C        7
    #define ASPEED_RESET_AHB        8
    #define ASPEED_RESET_CRT1        9
+#define ASPEED_RESET_VIDEO        10
 
    #endif
------------------------------------------------------------------

These are all I did for testing KVM.

Cheers,
Jae

#10 lore raw
In reply to Lei YU

Hi Lei,

Quoted 7 lines · Lei YU

Hi Jae,

Thanks for sharing the detailed information.
I managed to do a build on Romulus and see the KVM partially working:
* I can see the video output in the KVM page;
* I am unable to do any operation, e.g. pressing a key does not seem to be sent to host

Check if there are /dev/hidg0 and /dev/hidg1. If not, check kernel configs again.
Also, check 'start-ipkvm.service' file whether that launches obmc-ikvm using this options:
ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1

Quoted 2 lines

* When I browse to another page and browse back to KVM page, it displays a small black block, instead of showing correct KVM view.

Try this patch:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392

Cheers,
Jae

Quoted 12 lines

Did you meet such issue before and do you happen to know how to fix this?

Thanks a lot!

Quoted 6 lines · Jae Hyun Yoo

Attaching a patch file for back porting of Eddie's aspeed-video driver. Diff in below doesn't include the newly created file so I'm attaching it as a patch.

-Jae

#11 lore raw
Quoted 6 lines

Check if there are /dev/hidg0 and /dev/hidg1. If not, check kernel configs again.
Also, check 'start-ipkvm.service' file whether that launches obmc-ikvm using this options:
ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1

Confirmed that both /dev/hidg0 and /dev/hidg1 are created, and the service is started with obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1.

Quoted 3 lines

Try this patch:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392

Unfortunately, the patch does not fix the issue, the KVM page still shows a small black block when browsed to another page and back.

#12 lore raw

Hi Jae/Xiuzhi,

I have tried enabling KVM as mentioned in the mail thread. But unable to get the video in the web. Also tried using vnc viewer, but connection refused.

I see the following:

# obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1
26/02/2019 09:41:04 Unable to establish connection with systemd socket
26/02/2019 09:41:04 Listening for VNC connections on TCP port 5900
26/02/2019 09:41:04 Listening for VNC connections on TCP6 port 5900
26/02/2019 09:42:17 Got connection from client 127.0.0.1
26/02/2019 09:42:17   other clients:
26/02/2019 09:42:17 Normal socket connection
# dmesg | grep video
[   18.921547] aspeed-video 1e700000.video: assigned reserved memory node
framebuffer
[   59.799015] aspeed-video 1e700000.video: Timed out; first mode detect

obmc-ikvm recipe:

SUMMARY = "obmc kvm application"
DESCRIPTION = "Interfaces video device and USB HID device on BMC"

FILESEXTRAPATHS_append := ":${THISDIR}/${PN}"

inherit systemd
inherit obmc-phosphor-systemd
inherit autotools pkgconfig

LICENSE = "Apache-2.0"
LIC_FILES_CHKSUM = "file://LICENSE;md5=e3fc50a88d0a364313df4b21ef20c29e"

PV = "1.0+git${SRCPV}"

SRC_URI = "git://github.com/openbmc/obmc-ikvm \
        file://Fix_video_restarting_sync_issue.patch \
        "

SRCREV = "2bc661d34abd1fda92a9d2b256ed88ca0e90d09a"

DEPENDS += "autoconf-archive"
DEPENDS += "libvncserver"
DEPENDS += "sdbusplus"
DEPENDS += "sdeventplus"
DEPENDS += "phosphor-logging"
DEPENDS += "phosphor-dbus-interfaces"
DEPENDS += "systemd"

S = "${WORKDIR}/git"

SERVICE_FILE = "start-ipkvm.service" SYSTEMD_PACKAGES = "${PN}" SYSTEMD_SERVICE_${PN} += "${SERVICE_FILE}" SYSTEMD_DEFAULT_TARGET ?= "${SYSTEMD_DEFAULT_TARGET}"

What am I doing wrong here?
Do we need to set "-DBMCWEB_ENABLE_KVM=ON" in bmcweb?

Regards,
Thaj

Quoted 18 lines · Lei YU
Quoted 6 lines

Check if there are /dev/hidg0 and /dev/hidg1. If not, check kernel configs again.
Also, check 'start-ipkvm.service' file whether that launches obmc-ikvm using this options:
ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1

Confirmed that both /dev/hidg0 and /dev/hidg1 are created, and the service is
started with obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1.

Quoted 3 lines

Try this patch:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392

Unfortunately, the patch does not fix the issue, the KVM page still shows a small black block when browsed to another page and back.

#13 lore raw

Hi Thaju/Lei,

Check journal (journalctl -f) while you're testing KVM if there is any suspicious error. In normal opration, you should see logs like below:

Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Got connection from client 127.0.0.1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12   other clients:
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Normal socket connection
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Client Protocol Version 3.8
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Protocol version sent 3.8, using 3.8
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 rfbProcessClientSecurityType: executing handler for type 1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 rfbProcessClientSecurityType: returning securityResult for client rfb version >= 3.8
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Pixel format for client 127.0.0.1:
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12   32 bpp, depth 24, little endian
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12   true colour: max r 255 g 255 b 255, shift r 16 g 8 b 0
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Using image quality level 6 for client 127.0.0.1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Using JPEG subsampling 0, Q79 for client 127.0.0.1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Using compression level 2 for client 127.0.0.1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Enabling NewFBSize protocol extension for client 127.0.0.1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Enabling LastRect protocol extension for client 127.0.0.1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 rfbProcessClientNormalMessage: ignoring unsupported encoding type Enc(0xFFFFFEFE)
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 rfbProcessClientNormalMessage: ignoring unsupported encoding type Enc(0xFFFFFECC)
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 rfbProcessClientNormalMessage: ignoring unsupported encoding type Enc(0xFFFFFEC8)
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 rfbProcessClientNormalMessage: ignoring unsupported encoding type Enc(0xFFFFFEC7)
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Enabling full-color cursor updates for client 127.0.0.1
Mar 19 16:17:12 wolfpass env[1331]: 19/03/2019 16:17:12 Using tight encoding for client 127.0.0.1
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19 Client 127.0.0.1 gone
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19 Statistics 
      events    Transmit/ RawEquiv ( saved)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19 FramebufferUpdate   :      1 |         0/        0 (  0.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  LastRect 
    :     58 |       696/      696 (  0.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  tight 
    :     58 |   1831655/182452920 ( 99.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  RichCursor 
    :      1 |      1684/     1684 (  0.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  TOTALS 
    :    118 |   1834035/182455300 ( 99.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19 Statistics 
      events    Received/ RawEquiv ( saved)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  PointerEvent 
    :      3 |        18/       18 (  0.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19 FramebufferUpdate   :     59 |       590/      590 (  0.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  SetEncodings 
    :      1 |        68/       68 (  0.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  SetPixelFormat 
    :      1 |        20/       20 (  0.0%)
Mar 19 16:17:19 wolfpass env[1331]: 19/03/2019 16:17:19  TOTALS 
    :     64 |       696/      696 (  0.0%)

Also, try these two patches. I added one more patch.
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433

If it doesn't make help, try external connection using Tight VNC viewer. Currently, external connection is disabled by libvncserver's listening ip setting so apply below patch if you are going to enable external connection.

In obmc-ikvm package:
-----------------------

ikvm_server.cpp +1 -2
index 35310da292be..b7bd5344ed0d 100644
@@ -44,7 +44,7 @@ Server::Server(const Args& args, Input& i, Video& v) :
     server->cursor->xhot = 1;
     server->cursor->yhot = 1;
 
-    rfbStringToAddr(&ip[0], &server->listenInterface);
+    //rfbStringToAddr(&ip[0], &server->listenInterface);
 
     rfbInitServer(server);
-------------------------

Make sure that it's just for test purpose. We intentionally disabled external connection because of security issue. Also, you might need to use Tight VNC viewer only. I checked that other VNC viewers don't support tight encoding properly.

Regards,
Jae

Quoted 83 lines · Thaju

Hi Jae/Xiuzhi,

I have tried enabling KVM as mentioned in the mail thread. But unable to get the video in the web. Also tried using vnc viewer, but connection refused.

I see the following:

# obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1
26/02/2019 09:41:04 Unable to establish connection with systemd socket
26/02/2019 09:41:04 Listening for VNC connections on TCP port 5900
26/02/2019 09:41:04 Listening for VNC connections on TCP6 port 5900
26/02/2019 09:42:17 Got connection from client 127.0.0.1
26/02/2019 09:42:17   other clients:
26/02/2019 09:42:17 Normal socket connection
# dmesg | grep video
[   18.921547] aspeed-video 1e700000.video: assigned reserved memory node framebuffer
[   59.799015] aspeed-video 1e700000.video: Timed out; first mode detect

obmc-ikvm recipe:

SUMMARY = "obmc kvm application"
DESCRIPTION = "Interfaces video device and USB HID device on BMC"

FILESEXTRAPATHS_append := ":${THISDIR}/${PN}"

inherit systemd
inherit obmc-phosphor-systemd
inherit autotools pkgconfig

LICENSE = "Apache-2.0"
LIC_FILES_CHKSUM = "file://LICENSE;md5=e3fc50a88d0a364313df4b21ef20c29e"

PV = "1.0+git${SRCPV}"

SRC_URI = "git://github.com/openbmc/obmc-ikvm <http://github.com/openbmc/obmc-ikvm> \
        file://Fix_video_restarting_sync_issue.patch \
        "
SRCREV = "2bc661d34abd1fda92a9d2b256ed88ca0e90d09a"

DEPENDS += "autoconf-archive"
DEPENDS += "libvncserver"
DEPENDS += "sdbusplus"
DEPENDS += "sdeventplus"
DEPENDS += "phosphor-logging"
DEPENDS += "phosphor-dbus-interfaces"
DEPENDS += "systemd"

S = "${WORKDIR}/git"

SERVICE_FILE = "start-ipkvm.service" SYSTEMD_PACKAGES = "${PN}" SYSTEMD_SERVICE_${PN} += "${SERVICE_FILE}" SYSTEMD_DEFAULT_TARGET ?= "${SYSTEMD_DEFAULT_TARGET}"

What am I doing wrong here?
Do we need to set "-DBMCWEB_ENABLE_KVM=ON" in bmcweb?

Regards,
Thaj

On Wed, Mar 20, 2019 at 11:01 AM Lei YU <[email protected] <mailto:[email protected]>> wrote:

> Check if there are /dev/hidg0 and /dev/hidg1. If not, check kernel > configs again.
> Also, check 'start-ipkvm.service' file whether that launches
obmc-ikvm
> using this options:
> ExecStart=/usr/bin/env obmc-ikvm -v /dev/video0 -f 10 -k
/dev/hidg0 -p
> /dev/hidg1

Confirmed that both /dev/hidg0 and /dev/hidg1 are created, and the service is
started with obmc-ikvm -v /dev/video0 -f 10 -k /dev/hidg0 -p /dev/hidg1.

>
> Try this patch:
> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392

Unfortunately, the patch does not fix the issue, the KVM page still shows a
small black block when browsed to another page and back.

#14 lore raw
Quoted 2 lines

Check journal (journalctl -f) while you're testing KVM if there is any suspicious error. In normal opration, you should see logs like below:

I see the same logs as Thaj, only below logs show when the issue occurs.

Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Unable to
establish connection with systemd socket
Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
VNC connections on TCP port 5900
Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
VNC connections on TCP6 port 5900
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Got connection
from client 127.0.0.1
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06   other clients:
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Normal socket connection

And sometimes it works, then I see logs like:

Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Normal socket connection
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Client Protocol
Version 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Protocol
version sent 3.8, using 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
rfbProcessClientSecurityType: executing handler for type 1
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
rfbProcessClientSecurityType: returning securityResult for client rfb
version >= 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Pixel format
for client 192.168.10.11:
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   32 bpp, depth
24, little endian
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   true colour:
max r 255 g 255 b 255, shift r 16 g 8 b 0
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using
compression level 1 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using image
quality level 6 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using JPEG
subsampling 0, Q79 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
X-style cursor updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
full-color cursor updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling cursor
position updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
LastRect protocol extension for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using tight

encoding for client 192.168.10.11

It's found that the VNC server may work when the screen is changed, e.g. reboot the OS, or enter the OS.
When OS is in idle, usually it will not work.

Quoted 3 lines

Also, try these two patches. I added one more patch.
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433

Tried the patches, still the same issue as above.

#15 lore raw

Hi Lei,

Are you using a web browser from Windows OS? It seems that there is an issue of Windows websockify/novnc. Use a browser from OS other than Windows til the issue is resolved.

Did you test it using Tight VNC viewer after enabling external connection? The reason why I suggested that is to identify whether this issue comes from KVM web proxy or from KVM server.

Need more information of your test environment.
Server platform:
BMC SoC:
Host OS:

Regards,
Jae

Quoted 61 lines · Lei YU
Quoted 2 lines

Check journal (journalctl -f) while you're testing KVM if there is any suspicious error. In normal opration, you should see logs like below:

I see the same logs as Thaj, only below logs show when the issue occurs.

Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Unable to
establish connection with systemd socket
Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
VNC connections on TCP port 5900
Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
VNC connections on TCP6 port 5900
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Got connection
from client 127.0.0.1
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06   other clients:
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Normal socket connection

And sometimes it works, then I see logs like:

Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Normal socket connection
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Client Protocol
Version 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Protocol
version sent 3.8, using 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
rfbProcessClientSecurityType: executing handler for type 1
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
rfbProcessClientSecurityType: returning securityResult for client rfb
version >= 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Pixel format
for client 192.168.10.11:
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   32 bpp, depth
24, little endian
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   true colour:
max r 255 g 255 b 255, shift r 16 g 8 b 0
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using
compression level 1 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using image
quality level 6 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using JPEG
subsampling 0, Q79 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
X-style cursor updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
full-color cursor updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling cursor
position updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
LastRect protocol extension for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using tight

encoding for client 192.168.10.11

It's found that the VNC server may work when the screen is changed, e.g. reboot the OS, or enter the OS.
When OS is in idle, usually it will not work.

Quoted 3 lines

Also, try these two patches. I added one more patch.
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433

Tried the patches, still the same issue as above.

#16 lore raw

Hi Jae,

It works. Found one issue which is explained below.

I have applied below 3 patches:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19393

I can confirm that:
1. KVM video appears fine in webui
2. All keyboard characters works fine.
3. Both mouse buttons works fine (with patch)

Now the issue is when host OS ( RHEL 7.5 in my case ) goes to power saving mode (black screen) after some N mins (depends on OS settings), looks like it does not accept the KVM HID events, it doesn't wake up from power saving mode through KVM. I see below error when this happens.

Mar 21 09:31:43  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:43  kernel: aspeed-video 1e700000.video: Timed out; first mode
detect
Mar 21 09:31:44  kernel: aspeed-video 1e700000.video: Timed out when
stopping streaming
Mar 21 09:31:45  kernel: aspeed-video 1e700000.video: Timed out; first mode
detect
Mar 21 09:31:46  kernel: aspeed-video 1e700000.video: Timed out; first mode
detect
Mar 21 09:31:47  env[1408]: 21/03/2019 09:31:47 Sending
rfbEncodingNewFBSize for resize to (640x480)
Mar 21 09:31:48  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:49  kernel: aspeed-video 1e700000.video: Timed out; first mode
detect
Mar 21 09:31:50  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:51  kernel: aspeed-video 1e700000.video: Timed out; first mode
detect
Mar 21 09:31:52  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:53  kernel: aspeed-video 1e700000.video: Timed out; first mode

detect

On this state, I used physical keyboard on the host OS to bring it back from power saving mode and I see KVM also comes back. I have also tried setting the power saving timeout to "never". It works fine.

To recreate this issue
1. Set power saving timeout to 1 min. (In RHEL 7.5, Settings > Power > Blank Screen)
2. Leave the host OS to idle for 1 min.
3. You will see screen goes black after 1 min and KVM stops responding.

Hope you understood the issue. Thanks for your support.

Regards,
Thaj

Quoted 83 lines · Jae Hyun Yoo

Hi Lei,

Are you using a web browser from Windows OS? It seems that there is an issue of Windows websockify/novnc. Use a browser from OS other than Windows til the issue is resolved.

Did you test it using Tight VNC viewer after enabling external connection? The reason why I suggested that is to identify whether this issue comes from KVM web proxy or from KVM server.

Need more information of your test environment.
Server platform:
BMC SoC:
Host OS:

Regards,
Jae

Quoted 5 lines · Lei YU
Quoted 2 lines

Check journal (journalctl -f) while you're testing KVM if there is any suspicious error. In normal opration, you should see logs like below:

I see the same logs as Thaj, only below logs show when the issue occurs.

Quoted 10 lines
Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Unable to
establish connection with systemd socket
Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
VNC connections on TCP port 5900
Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
VNC connections on TCP6 port 5900
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Got connection
from client 127.0.0.1
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06   other clients:
Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Normal socket

connection

Quoted 4 lines

And sometimes it works, then I see logs like:

Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Normal socket

connection

Quoted 42 lines
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Client Protocol
Version 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Protocol
version sent 3.8, using 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
rfbProcessClientSecurityType: executing handler for type 1
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
rfbProcessClientSecurityType: returning securityResult for client rfb
version >= 3.8
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Pixel format
for client 192.168.10.11:
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   32 bpp, depth
24, little endian
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   true colour:
max r 255 g 255 b 255, shift r 16 g 8 b 0
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using
compression level 1 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using image
quality level 6 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using JPEG
subsampling 0, Q79 for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
X-style cursor updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
full-color cursor updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling cursor
position updates for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
LastRect protocol extension for client 192.168.10.11
Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using tight

encoding for client 192.168.10.11

It's found that the VNC server may work when the screen is changed, e.g. reboot the OS, or enter the OS.
When OS is in idle, usually it will not work.

Quoted 3 lines

Also, try these two patches. I added one more patch.
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433

Tried the patches, still the same issue as above.

#17 lore raw

Hi Thaju,

It's really good to hear that you can see the working KVM in your machine. Thanks for sharing the result.

I saw the video waking up issue on Ubuntu 18 too. Need debugging on it. I'll share if I find something.

Regards,
Jae

Quoted 134 lines · Thaju

Hi Jae,

It works. Found one issue which is explained below.

I have applied below 3 patches:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19393

I can confirm that:
1. KVM video appears fine in webui
2. All keyboard characters works fine.
3. Both mouse buttons works fine (with patch)

Now the issue is when host OS ( RHEL 7.5 in my case ) goes to power saving mode (black screen) after some N mins (depends on OS settings), looks like it does not accept the KVM HID events, it doesn't wake up from power saving mode through KVM. I see below error when this happens.

Mar 21 09:31:43  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:43  kernel: aspeed-video 1e700000.video: Timed out; first mode detect
Mar 21 09:31:44  kernel: aspeed-video 1e700000.video: Timed out when stopping streaming
Mar 21 09:31:45  kernel: aspeed-video 1e700000.video: Timed out; first mode detect
Mar 21 09:31:46  kernel: aspeed-video 1e700000.video: Timed out; first mode detect
Mar 21 09:31:47  env[1408]: 21/03/2019 09:31:47 Sending rfbEncodingNewFBSize for resize to (640x480)
Mar 21 09:31:48  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:49  kernel: aspeed-video 1e700000.video: Timed out; first mode detect
Mar 21 09:31:50  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:51  kernel: aspeed-video 1e700000.video: Timed out; first mode detect
Mar 21 09:31:52  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:53  kernel: aspeed-video 1e700000.video: Timed out; first mode detect

On this state, I used physical keyboard on the host OS to bring it back from power saving mode and I see KVM also comes back. I have also tried setting the power saving timeout to "never". It works fine.

To recreate this issue
1. Set power saving timeout to 1 min. (In RHEL 7.5, Settings > Power > Blank Screen)
2. Leave the host OS to idle for 1 min.
3. You will see screen goes black after 1 min and KVM stops responding.

Hope you understood the issue. Thanks for your support.

Regards,
Thaj

On Thu, Mar 21, 2019 at 9:49 PM Jae Hyun Yoo <[email protected] <mailto:[email protected]>> wrote:

    Hi Lei,

    Are you using a web browser from Windows OS? It seems that there is an
    issue of Windows websockify/novnc. Use a browser from OS other than
    Windows til the issue is resolved.

    Did you test it using Tight VNC viewer after enabling external
    connection? The reason why I suggested that is to identify whether
    this issue comes from KVM web proxy or from KVM server.

    Need more information of your test environment.
    Server platform:
    BMC SoC:
    Host OS:

    Regards,
    Jae

    On 3/21/2019 12:02 AM, Lei YU wrote:
     >> Check journal (journalctl -f) while you're testing KVM if there
    is any
     >> suspicious error. In normal opration, you should see logs like
    below:
     >
     > I see the same logs as Thaj, only below logs show when the issue
    occurs.
     >
     > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Unable to
     > establish connection with systemd socket
     > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
     > VNC connections on TCP port 5900
     > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
     > VNC connections on TCP6 port 5900
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Got connection
     > from client 127.0.0.1
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06   other
    clients:
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Normal
    socket connection
     >
     > And sometimes it works, then I see logs like:
     >
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Normal
    socket connection
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Client
    Protocol
     > Version 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Protocol
     > version sent 3.8, using 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
     > rfbProcessClientSecurityType: executing handler for type 1
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
     > rfbProcessClientSecurityType: returning securityResult for client rfb
     > version >= 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Pixel format
     > for client 192.168.10.11 <http://192.168.10.11>:
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   32 bpp,
    depth
     > 24, little endian
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   true colour:
     > max r 255 g 255 b 255, shift r 16 g 8 b 0
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using
     > compression level 1 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using image
     > quality level 6 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using JPEG
     > subsampling 0, Q79 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > X-style cursor updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > full-color cursor updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
    cursor
     > position updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > LastRect protocol extension for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using tight
     > encoding for client 192.168.10.11
     >
     > It's found that the VNC server may work when the screen is
    changed, e.g.
     > reboot the OS, or enter the OS.
     > When OS is in idle, usually it will not work.
     >
     >> Also, try these two patches. I added one more patch.
     >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
     >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
     >
     > Tried the patches, still the same issue as above.
     >
#21 lore raw
In reply to Jae Hyun Yoo
Quoted 3 lines

Are you using a web browser from Windows OS? It seems that there is an issue of Windows websockify/novnc. Use a browser from OS other than Windows til the issue is resolved.

It's Ubuntu 18.04's Firefox and Chrome.

Quoted 3 lines

Did you test it using Tight VNC viewer after enabling external connection? The reason why I suggested that is to identify whether this issue comes from KVM web proxy or from KVM server.

Yes, TightVNC behaves the same as the KVM in browser.

Quoted 4 lines

Need more information of your test environment.
Server platform:
BMC SoC:
Host OS:

Server platform: OpenPOWER Romulus
BMC SoC: AST2500 with OpenBMC master + patches needed for KVM Host OS: Ubuntu 18.04 server

There is an update from my testing:
1. If the host is in Peitiboot (this is an OpenPOWER's thing), or in OS without X, KVM sometimes works, sometimes does not work, as the log shown above;
2. If the host is in OS with X (e.g. lightdm), then KVM always works well.

#22 lore raw
Quoted 14 lines · Lei YU
Quoted 3 lines

Are you using a web browser from Windows OS? It seems that there is an issue of Windows websockify/novnc. Use a browser from OS other than Windows til the issue is resolved.

It's Ubuntu 18.04's Firefox and Chrome.

Quoted 3 lines

Did you test it using Tight VNC viewer after enabling external connection? The reason why I suggested that is to identify whether this issue comes from KVM web proxy or from KVM server.

Yes, TightVNC behaves the same as the KVM in browser.

Quoted 4 lines

Need more information of your test environment.
Server platform:
BMC SoC:
Host OS:

Quoted 8 lines

Server platform: OpenPOWER Romulus
BMC SoC: AST2500 with OpenBMC master + patches needed for KVM Host OS: Ubuntu 18.04 server

There is an update from my testing:
1. If the host is in Peitiboot (this is an OpenPOWER's thing), or in OS without X, KVM sometimes works, sometimes does not work, as the log shown above;

I have seen this too; I think it's a VGA or VGA driver problem more than KVM. Not sure we will be able to find a fix.

Thanks,

Eddie

Quoted 2 lines

2. If the host is in OS with X (e.g. lightdm), then KVM always works well.

author
#23 lore raw

ikvm refresh failed on host screen 1680x1050 resolution

In reply to xiuzhi

Hi Jae, Ed,
The ikvm screen cann't refresh on ikvm webui screen when setting the host screen resolution to 1680*1050 . I must click the "Refresh" button to get the new host screen. Other resolutions works very well.
Please check it.
Best,
xiuzhi

#24 lore raw

Re: ikvm refresh failed on host screen 1680x1050 resolution

Quoted 4 lines · xiuzhi

Hi Jae, Ed,
    The ikvm screen cann't refresh on ikvm webui screen when setting
the host screen resolution to 1680*1050 . I must click the "Refresh" button to get the new host screen.  Other resolutions works very well.
  Please check it.

Hi Xiuzhi,

I can't see the resolution option in Ubuntu 18.03. What host OS did you use for the test? Let me clarify one more. Do you mean the "Refresh" button in KVM web page, or reload button of the web browser?

Thanks,
Jae

Quoted 2 lines

Best,
xiuzhi

author
#25 lore raw

Re: ikvm refresh failed on host screen 1680x1050 resolution

Hi Jae,
My host OS is RedHat7.
Both "Refresh" button on ikvm web Page and "View--Reload This Page" or reload button on chrome browser menu works. Xiuzhi

------------------ Original ------------------

From:  "jae.hyun.yoo";<[email protected]>;
Date:  Apr 5, 2019
To:  "xiuzhi"<[email protected]>; "openbmc"<[email protected]>; "ed"<[email protected]>; "Eddie James"<[email protected]>; 

Subject: Re: ikvm refresh failed on host screen 1680x1050 resolution

Quoted 5 lines · xiuzhi

Hi Jae, Ed,
The ikvm screen cann't refresh on ikvm webui screen when setting
the host screen resolution to 1680*1050 . I must click the "Refresh" button to get the new host screen. Other resolutions works very well.
Please check it.

Hi Xiuzhi,

I can't see the resolution option in Ubuntu 18.03. What host OS did you use for the test? Let me clarify one more. Do you mean the "Refresh" button in KVM web page, or reload button of the web browser?

Thanks,
Jae

Quoted 2 lines

Best,
xiuzhi

author
#18 lore raw
In reply to Jae Hyun Yoo

Hi Jae, Ed,
Are there any updates for the video waking up issue?
Best,
Xiuzhi

------------------ Original ------------------
From: "jae.hyun.yoo";<[email protected]>;
Send time: Friday, Mar 22, 2019 5:31 AM
To: "Thaju"<[email protected]>;
Cc: "xiuzhi"<[email protected]>; "openbmc"<[email protected]>; "ed"<[email protected]>; "Eddie James"<[email protected]>;
Subject: Re: Does obmc-ikvm work with bmcweb?

Hi Thaju,

It's really good to hear that you can see the working KVM in your machine. Thanks for sharing the result.

I saw the video waking up issue on Ubuntu 18 too. Need debugging on it. I'll share if I find something.

Regards,
Jae

Quoted 149 lines · Thaju

Hi Jae,

It works. Found one issue which is explained below.

I have applied below 3 patches:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19393

I can confirm that:
1. KVM video appears fine in webui
2. All keyboard characters works fine.
3. Both mouse buttons works fine (with patch)

Now the issue is when host OS ( RHEL 7.5 in my case ) goes to power saving mode (black screen) after some N mins (depends on OS settings), looks like it does not accept the KVM HID events, it doesn't wake up from power saving mode through KVM. I see below error when this happens.

Mar 21 09:31:43  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:43  kernel: aspeed-video 1e700000.video: Timed out; first 
mode detect
Mar 21 09:31:44  kernel: aspeed-video 1e700000.video: Timed out when 
stopping streaming
Mar 21 09:31:45  kernel: aspeed-video 1e700000.video: Timed out; first 
mode detect
Mar 21 09:31:46  kernel: aspeed-video 1e700000.video: Timed out; first 
mode detect
Mar 21 09:31:47  env[1408]: 21/03/2019 09:31:47 Sending 
rfbEncodingNewFBSize for resize to (640x480)
Mar 21 09:31:48  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:49  kernel: aspeed-video 1e700000.video: Timed out; first 
mode detect
Mar 21 09:31:50  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:51  kernel: aspeed-video 1e700000.video: Timed out; first 
mode detect
Mar 21 09:31:52  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:53  kernel: aspeed-video 1e700000.video: Timed out; first 

mode detect

On this state, I used physical keyboard on the host OS to bring it back from power saving mode and I see KVM also comes back. I have also tried setting the power saving timeout to "never". It works fine.

To recreate this issue
1. Set power saving timeout to 1 min. (In RHEL 7.5, Settings > Power > Blank Screen)
2. Leave the host OS to idle for 1 min.
3. You will see screen goes black after 1 min and KVM stops responding.

Hope you understood the issue. Thanks for your support.

Regards,
Thaj

On Thu, Mar 21, 2019 at 9:49 PM Jae Hyun Yoo <[email protected] <mailto:[email protected]>> wrote:

    Hi Lei,

    Are you using a web browser from Windows OS? It seems that there is an
    issue of Windows websockify/novnc. Use a browser from OS other than
    Windows til the issue is resolved.

    Did you test it using Tight VNC viewer after enabling external
    connection? The reason why I suggested that is to identify whether
    this issue comes from KVM web proxy or from KVM server.

    Need more information of your test environment.
    Server platform:
    BMC SoC:
    Host OS:

    Regards,
    Jae

    On 3/21/2019 12:02 AM, Lei YU wrote:
     >> Check journal (journalctl -f) while you're testing KVM if there
    is any
     >> suspicious error. In normal opration, you should see logs like
    below:
     >
     > I see the same logs as Thaj, only below logs show when the issue
    occurs.
     >
     > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Unable to
     > establish connection with systemd socket
     > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
     > VNC connections on TCP port 5900
     > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Listening for
     > VNC connections on TCP6 port 5900
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Got connection
     > from client 127.0.0.1
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06   other
    clients:
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Normal
    socket connection
     >
     > And sometimes it works, then I see logs like:
     >
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Normal
    socket connection
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Client
    Protocol
     > Version 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Protocol
     > version sent 3.8, using 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
     > rfbProcessClientSecurityType: executing handler for type 1
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
     > rfbProcessClientSecurityType: returning securityResult for client rfb
     > version >= 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Pixel format
     > for client 192.168.10.11 <http://192.168.10.11>:
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   32 bpp,
    depth
     > 24, little endian
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   true colour:
     > max r 255 g 255 b 255, shift r 16 g 8 b 0
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using
     > compression level 1 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using image
     > quality level 6 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using JPEG
     > subsampling 0, Q79 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > X-style cursor updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > full-color cursor updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
    cursor
     > position updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > LastRect protocol extension for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using tight
     > encoding for client 192.168.10.11
     >
     > It's found that the VNC server may work when the screen is
    changed, e.g.
     > reboot the OS, or enter the OS.
     > When OS is in idle, usually it will not work.
     >
     >> Also, try these two patches. I added one more patch.
     >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
     >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
     >
     > Tried the patches, still the same issue as above.
     >
#19 lore raw

Hi Xiuzhi,

Please try this patch:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/20593

Thanks,
Jae

Quoted 182 lines · xiuzhi

Hi Jae, Ed,
   Are there any updates for the  video waking up issue? Best,
Xiuzhi

------------------ Original ------------------
*From: * "jae.hyun.yoo";<[email protected]>;
*Send time:* Friday, Mar 22, 2019 5:31 AM
*To:* "Thaju"<[email protected]>;
*Cc:* "xiuzhi"<[email protected]>; "openbmc"<[email protected]>; "ed"<[email protected]>; "Eddie James"<[email protected]>; *Subject: * Re: Does obmc-ikvm work with bmcweb?

Hi Thaju,

It's really good to hear that you can see the working KVM in your machine. Thanks for sharing the result.

I saw the video waking up issue on Ubuntu 18 too. Need debugging on it. I'll share if I find something.

Regards,
Jae

Quoted 57 lines · Thaju

Hi Jae,

It works. Found one issue which is explained below.

I have applied below 3 patches:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19393

I can confirm that:
1. KVM video appears fine in webui
2. All keyboard characters works fine.
3. Both mouse buttons works fine (with patch)

Now the issue is when host OS ( RHEL 7.5 in my case ) goes to power saving mode (black screen) after some N mins (depends on OS settings), looks like it does not accept the KVM HID events, it doesn't wake up from power saving mode through KVM. I see below error when this happens.

Mar 21 09:31:43  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:43  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:44  kernel: aspeed-video 1e700000.video: Timed out when
stopping streaming
Mar 21 09:31:45  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:46  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:47  env[1408]: 21/03/2019 09:31:47 Sending
rfbEncodingNewFBSize for resize to (640x480)
Mar 21 09:31:48  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:49  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:50  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:51  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:52  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:53  kernel: aspeed-video 1e700000.video: Timed out; first

mode detect

On this state, I used physical keyboard on the host OS to bring it back from power saving mode and I see KVM also comes back. I have also tried setting the power saving timeout to "never". It works fine.

To recreate this issue
1. Set power saving timeout to 1 min. (In RHEL 7.5, Settings > Power > Blank Screen)
2. Leave the host OS to idle for 1 min.
3. You will see screen goes black after 1 min and KVM stops responding.

Hope you understood the issue. Thanks for your support.

Regards,
Thaj

On Thu, Mar 21, 2019 at 9:49 PM Jae Hyun Yoo <[email protected] <mailto:[email protected]>>

Quoted 4 lines · wrote:

     Hi Lei,

     Are you using a web browser from Windows OS? It seems that there

is an

Quoted 15 lines

     issue of Windows websockify/novnc. Use a browser from OS other than Windows til the issue is resolved.

     Did you test it using Tight VNC viewer after enabling external connection? The reason why I suggested that is to identify whether this issue comes from KVM web proxy or from KVM server.

     Need more information of your test environment. Server platform:
     BMC SoC:
     Host OS:

     Regards,
     Jae

Quoted 12 lines

     On 3/21/2019 12:02 AM, Lei YU wrote:
      >> Check journal (journalctl -f) while you're testing KVM if there is any
      >> suspicious error. In normal opration, you should see logs like below:
      >
      > I see the same logs as Thaj, only below logs show when the issue occurs.
      >
      > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Unable to > establish connection with systemd socket
      > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02

Listening for

Quoted 2 lines

      > VNC connections on TCP port 5900
      > Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02

Listening for

Quoted 2 lines

      > VNC connections on TCP6 port 5900
      > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Got

connection

Quoted 11 lines

      > from client 127.0.0.1
      > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06   other clients:
      > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Normal socket connection
      >
      > And sometimes it works, then I see logs like:
      >
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Normal socket connection
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Client

Quoted 8 lines

     Protocol
      > Version 3.8
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Protocol > version sent 3.8, using 3.8
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 > rfbProcessClientSecurityType: executing handler for type 1 > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 > rfbProcessClientSecurityType: returning securityResult for

client rfb

Quoted 2 lines

      > version >= 3.8
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Pixel

format

Quoted 5 lines

      > for client 192.168.10.11 <http://192.168.10.11>:
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   32 bpp, depth
      > 24, little endian
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   true

colour:

Quoted 31 lines

      > max r 255 g 255 b 255, shift r 16 g 8 b 0
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using > compression level 1 for client 192.168.10.11
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using image > quality level 6 for client 192.168.10.11
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using JPEG > subsampling 0, Q79 for client 192.168.10.11
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling > X-style cursor updates for client 192.168.10.11
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling > full-color cursor updates for client 192.168.10.11
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling cursor
      > position updates for client 192.168.10.11
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling > LastRect protocol extension for client 192.168.10.11
      > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using tight > encoding for client 192.168.10.11
      >
      > It's found that the VNC server may work when the screen is changed, e.g.
      > reboot the OS, or enter the OS.
      > When OS is in idle, usually it will not work.
      >
      >> Also, try these two patches. I added one more patch. >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
      >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
      >
      > Tried the patches, still the same issue as above.
      >

author
#20 lore raw

Hi Jae,
The patch works very well. Thanks.
Best,
Xiuzhi
Hi Xiuzhi,

Please try this patch:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/20593

Thanks,
Jae

Quoted 183 lines · xiuzhi

Hi Jae, Ed,
Are there any updates for the video waking up issue?
Best,
Xiuzhi

------------------ Original ------------------
*From: * "jae.hyun.yoo";<[email protected]>; *Send time:* Friday, Mar 22, 2019 5:31 AM
*To:* "Thaju"<[email protected]>;
*Cc:* "xiuzhi"<[email protected]>; "openbmc"<[email protected]>; "ed"<[email protected]>; "Eddie James"<[email protected]>; *Subject: * Re: Does obmc-ikvm work with bmcweb?

Hi Thaju,

It's really good to hear that you can see the working KVM in your machine. Thanks for sharing the result.

I saw the video waking up issue on Ubuntu 18 too. Need debugging on it. I'll share if I find something.

Regards,
Jae

Quoted 57 lines · Thaju

Hi Jae,

It works. Found one issue which is explained below.

I have applied below 3 patches:
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19393

I can confirm that:
1. KVM video appears fine in webui
2. All keyboard characters works fine.
3. Both mouse buttons works fine (with patch)

Now the issue is when host OS ( RHEL 7.5 in my case ) goes to power saving mode (black screen) after some N mins (depends on OS settings), looks like it does not accept the KVM HID events, it doesn't wake up from power saving mode through KVM. I see below error when this happens.

Mar 21 09:31:43  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:43  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:44  kernel: aspeed-video 1e700000.video: Timed out when
stopping streaming
Mar 21 09:31:45  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:46  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:47  env[1408]: 21/03/2019 09:31:47 Sending
rfbEncodingNewFBSize for resize to (640x480)
Mar 21 09:31:48  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:49  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:50  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:51  kernel: aspeed-video 1e700000.video: Timed out; first
mode detect
Mar 21 09:31:52  obmc-ikvm[1408]: Failed to query timings
Mar 21 09:31:53  kernel: aspeed-video 1e700000.video: Timed out; first

mode detect

On this state, I used physical keyboard on the host OS to bring it back from power saving mode and I see KVM also comes back. I have also tried setting the power saving timeout to "never". It works fine.

To recreate this issue
1. Set power saving timeout to 1 min. (In RHEL 7.5, Settings > Power > Blank Screen)
2. Leave the host OS to idle for 1 min.
3. You will see screen goes black after 1 min and KVM stops responding.

Hope you understood the issue. Thanks for your support.

Regards,
Thaj

On Thu, Mar 21, 2019 at 9:49 PM Jae Hyun Yoo <[email protected] <mailto:[email protected]>>

Quoted 4 lines · wrote:

Hi Lei,

Are you using a web browser from Windows OS? It seems that there

is an

Quoted 15 lines

issue of Windows websockify/novnc. Use a browser from OS other than Windows til the issue is resolved.

Did you test it using Tight VNC viewer after enabling external connection? The reason why I suggested that is to identify whether this issue comes from KVM web proxy or from KVM server.

Need more information of your test environment. Server platform:
BMC SoC:
Host OS:

Regards,
Jae

Quoted 12 lines

On 3/21/2019 12:02 AM, Lei YU wrote:
>> Check journal (journalctl -f) while you're testing KVM if there
is any
>> suspicious error. In normal opration, you should see logs like
below:
>
> I see the same logs as Thaj, only below logs show when the issue
occurs.
>
> Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02 Unable to > establish connection with systemd socket
> Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02

Listening for

Quoted 2 lines

> VNC connections on TCP port 5900
> Mar 21 05:26:02 romulus env[1564]: 21/03/2019 05:26:02

Listening for

Quoted 2 lines

> VNC connections on TCP6 port 5900
> Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Got

connection

Quoted 11 lines
     > from client 127.0.0.1
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06   other
    clients:
     > Mar 21 05:26:06 romulus env[1564]: 21/03/2019 05:26:06 Normal
    socket connection
     >
     > And sometimes it works, then I see logs like:
     >
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Normal
    socket connection
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Client
Quoted 8 lines
    Protocol
     > Version 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Protocol
     > version sent 3.8, using 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
     > rfbProcessClientSecurityType: executing handler for type 1
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12
     > rfbProcessClientSecurityType: returning securityResult for 

client rfb

Quoted 2 lines
     > version >= 3.8
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Pixel 

format

Quoted 5 lines
     > for client 192.168.10.11 <http://192.168.10.11>:
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   32 bpp,
    depth
     > 24, little endian
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12   true 

colour:

Quoted 31 lines
     > max r 255 g 255 b 255, shift r 16 g 8 b 0
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using
     > compression level 1 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using image
     > quality level 6 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using JPEG
     > subsampling 0, Q79 for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > X-style cursor updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > full-color cursor updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
    cursor
     > position updates for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Enabling
     > LastRect protocol extension for client 192.168.10.11
     > Mar 21 06:54:12 romulus env[1957]: 21/03/2019 06:54:12 Using tight
     > encoding for client 192.168.10.11
     >
     > It's found that the VNC server may work when the screen is
    changed, e.g.
     > reboot the OS, or enter the OS.
     > When OS is in idle, usually it will not work.
     >
     >> Also, try these two patches. I added one more patch.
     >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19392
     >> https://gerrit.openbmc-project.xyz/c/openbmc/obmc-ikvm/+/19433
     >
     > Tried the patches, still the same issue as above.
     >