[PATCH] ipmi:pci: Blacklist a Realtek "IPMI" device

[PATCH] ipmi:pci: Blacklist a Realtek "IPMI" device

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 1935 bytes --]

From: Corey Minyard <cminyard@mvista.com>
Date: Thu, 15 Feb 2018 16:58:26 -0600

[ upstream commit bc48fa1b9d3b04106055b27078da824cd209865a ]

Realtek has some sort of "Virtual" IPMI device on the PCI bus as a
KCS controller, but whatever it is, it's not one.  Ignore it if seen.

Reported-by: Chris Chiu <chiu@endlessm.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Tested-by: Daniel Drake <drake@endlessm.com>

[ Commit 13d0b35c (ipmi_si: Move PCI setup to another file) from Linux
  4.15-rc1 has not been back ported, so the PCI code is still in
  `drivers/char/ipmi/ipmi_si_intf.c`, requiring to apply the commit
  manually.

  This fixes a 100 s boot delay on the HP EliteDesk 705 G4 MT with Linux
  4.14.94. ]

Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
---
 drivers/char/ipmi/ipmi_si_intf.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c
index c04aa11f0e21..6d18f8090cea 100644
--- a/drivers/char/ipmi/ipmi_si_intf.c
+++ b/drivers/char/ipmi/ipmi_si_intf.c
@@ -2447,6 +2447,15 @@ static int ipmi_pci_probe_regspacing(struct smi_info *info)
 	return DEFAULT_REGSPACING;
 }
 
+static struct pci_device_id ipmi_pci_blacklist[] = {
+        /*
+         * This is a "Virtual IPMI device", whatever that is.  It appears
+         * as a KCS device by the class, but it is not one.
+         */
+        { PCI_VDEVICE(REALTEK, 0x816c) },
+        { 0, }
+};
+
 static int ipmi_pci_probe(struct pci_dev *pdev,
 				    const struct pci_device_id *ent)
 {
@@ -2454,6 +2463,9 @@ static int ipmi_pci_probe(struct pci_dev *pdev,
 	int class_type = pdev->class & PCI_ERMC_CLASSCODE_TYPE_MASK;
 	struct smi_info *info;
 
+        if (pci_match_id(ipmi_pci_blacklist, pdev))
+                return -ENODEV;
+
 	info = smi_info_alloc();
 	if (!info)
 		return -ENOMEM;
-- 
2.17.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 5174 bytes --]

Re: [PATCH] ipmi:pci: Blacklist a Realtek "IPMI" device

[Greg KH]

On Wed, Jan 23, 2019 at 05:27:51PM +0100, Paul Menzel wrote:
> From: Corey Minyard <cminyard@mvista.com>
> Date: Thu, 15 Feb 2018 16:58:26 -0600
> 
> [ upstream commit bc48fa1b9d3b04106055b27078da824cd209865a ]
> 
> Realtek has some sort of "Virtual" IPMI device on the PCI bus as a
> KCS controller, but whatever it is, it's not one.  Ignore it if seen.
> 
> Reported-by: Chris Chiu <chiu@endlessm.com>
> Signed-off-by: Corey Minyard <cminyard@mvista.com>
> Tested-by: Daniel Drake <drake@endlessm.com>
> 
> [ Commit 13d0b35c (ipmi_si: Move PCI setup to another file) from Linux
>   4.15-rc1 has not been back ported, so the PCI code is still in
>   `drivers/char/ipmi/ipmi_si_intf.c`, requiring to apply the commit
>   manually.
> 
>   This fixes a 100 s boot delay on the HP EliteDesk 705 G4 MT with Linux
>   4.14.94. ]
> 
> Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
> ---
>  drivers/char/ipmi/ipmi_si_intf.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c
> index c04aa11f0e21..6d18f8090cea 100644
> --- a/drivers/char/ipmi/ipmi_si_intf.c
> +++ b/drivers/char/ipmi/ipmi_si_intf.c
> @@ -2447,6 +2447,15 @@ static int ipmi_pci_probe_regspacing(struct smi_info *info)
>  	return DEFAULT_REGSPACING;
>  }
>  
> +static struct pci_device_id ipmi_pci_blacklist[] = {
> +        /*
> +         * This is a "Virtual IPMI device", whatever that is.  It appears
> +         * as a KCS device by the class, but it is not one.
> +         */
> +        { PCI_VDEVICE(REALTEK, 0x816c) },
> +        { 0, }
> +};
> +
>  static int ipmi_pci_probe(struct pci_dev *pdev,
>  				    const struct pci_device_id *ent)
>  {
> @@ -2454,6 +2463,9 @@ static int ipmi_pci_probe(struct pci_dev *pdev,
>  	int class_type = pdev->class & PCI_ERMC_CLASSCODE_TYPE_MASK;
>  	struct smi_info *info;
>  
> +        if (pci_match_id(ipmi_pci_blacklist, pdev))
> +                return -ENODEV;
> +

Your patch does not have tabs where it should have tabs :(

Can you fix this up and resend?

thanks,

greg k-h

[PATCH v2] ipmi:pci: Blacklist a Realtek "IPMI" device

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 1918 bytes --]

From: Corey Minyard <cminyard@mvista.com>
Date: Thu, 15 Feb 2018 16:58:26 -0600

[ upstream commit bc48fa1b9d3b04106055b27078da824cd209865a ]

Realtek has some sort of "Virtual" IPMI device on the PCI bus as a
KCS controller, but whatever it is, it's not one.  Ignore it if seen.

Reported-by: Chris Chiu <chiu@endlessm.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Tested-by: Daniel Drake <drake@endlessm.com>

[ Commit 13d0b35c (ipmi_si: Move PCI setup to another file) from Linux
  4.15-rc1 has not been back ported, so the PCI code is still in
  `drivers/char/ipmi/ipmi_si_intf.c`, requiring to apply the commit
  manually.

  This fixes a 100 s boot delay on the HP EliteDesk 705 G4 MT with Linux
  4.14.94. ]

Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
---

v2: Use tabs. Sorry for messing that up.

 drivers/char/ipmi/ipmi_si_intf.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c
index c04aa11f0e21..a106cf7b5ee0 100644
--- a/drivers/char/ipmi/ipmi_si_intf.c
+++ b/drivers/char/ipmi/ipmi_si_intf.c
@@ -2447,6 +2447,15 @@ static int ipmi_pci_probe_regspacing(struct smi_info *info)
 	return DEFAULT_REGSPACING;
 }
 
+static struct pci_device_id ipmi_pci_blacklist[] = {
+	/*
+	 * This is a "Virtual IPMI device", whatever that is.  It appears
+	 * as a KCS device by the class, but it is not one.
+	 */
+	{ PCI_VDEVICE(REALTEK, 0x816c) },
+	{ 0, }
+};
+
 static int ipmi_pci_probe(struct pci_dev *pdev,
 				    const struct pci_device_id *ent)
 {
@@ -2454,6 +2463,9 @@ static int ipmi_pci_probe(struct pci_dev *pdev,
 	int class_type = pdev->class & PCI_ERMC_CLASSCODE_TYPE_MASK;
 	struct smi_info *info;
 
+	if (pci_match_id(ipmi_pci_blacklist, pdev))
+		return -ENODEV;
+
 	info = smi_info_alloc();
 	if (!info)
 		return -ENOMEM;
-- 
2.17.1


[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 5174 bytes --]

Re: [PATCH v2] ipmi:pci: Blacklist a Realtek "IPMI" device

[Greg KH]

On Wed, Jan 23, 2019 at 05:44:57PM +0100, Paul Menzel wrote:
> From: Corey Minyard <cminyard@mvista.com>
> Date: Thu, 15 Feb 2018 16:58:26 -0600
> 
> [ upstream commit bc48fa1b9d3b04106055b27078da824cd209865a ]
> 
> Realtek has some sort of "Virtual" IPMI device on the PCI bus as a
> KCS controller, but whatever it is, it's not one.  Ignore it if seen.
> 
> Reported-by: Chris Chiu <chiu@endlessm.com>
> Signed-off-by: Corey Minyard <cminyard@mvista.com>
> Tested-by: Daniel Drake <drake@endlessm.com>
> 
> [ Commit 13d0b35c (ipmi_si: Move PCI setup to another file) from Linux
>   4.15-rc1 has not been back ported, so the PCI code is still in
>   `drivers/char/ipmi/ipmi_si_intf.c`, requiring to apply the commit
>   manually.
> 
>   This fixes a 100 s boot delay on the HP EliteDesk 705 G4 MT with Linux
>   4.14.94. ]
> 
> Signed-off-by: Paul Menzel <pmenzel@molgen.mpg.de>
> ---
> 
> v2: Use tabs. Sorry for messing that up.

That worked, thanks.

greg k-h