[PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Li Ma]

From: Heiner Kallweit <hkallweit1@gmail.com>

[Backport: commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e]
This backport to avoid the bug caused by r8169.

This effectively reverts 4b5f82f6aaef. On a number of systems ASPM L1
causes tx timeouts with RTL8168h, see referenced bug report.

Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
Cc: stable@vger.kernel.org
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/ethernet/realtek/r8169_main.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
index 45147a1016be..27efd07f09ef 100644
--- a/drivers/net/ethernet/realtek/r8169_main.c
+++ b/drivers/net/ethernet/realtek/r8169_main.c
@@ -5224,13 +5224,9 @@ static int rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	/* Disable ASPM L1 as that cause random device stop working
 	 * problems as well as full system hangs for some PCIe devices users.
-	 * Chips from RTL8168h partially have issues with L1.2, but seem
-	 * to work fine with L1 and L1.1.
 	 */
 	if (rtl_aspm_is_safe(tp))
 		rc = 0;
-	else if (tp->mac_version >= RTL_GIGA_MAC_VER_46)
-		rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1_2);
 	else
 		rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1);
 	tp->aspm_manageable = !rc;
-- 
2.25.1

Re: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Mario Limonciello]

On 11/7/2023 02:52, Li Ma wrote:
> From: Heiner Kallweit <hkallweit1@gmail.com>
> 
> [Backport: commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e]
> This backport to avoid the bug caused by r8169.
> 
> This effectively reverts 4b5f82f6aaef. On a number of systems ASPM L1
> causes tx timeouts with RTL8168h, see referenced bug report.
> 
> Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
> Cc: stable@vger.kernel.org
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> Signed-off-by: David S. Miller <davem@davemloft.net>
> ---
>   drivers/net/ethernet/realtek/r8169_main.c | 4 ----
>   1 file changed, 4 deletions(-)
> 
> diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
> index 45147a1016be..27efd07f09ef 100644
> --- a/drivers/net/ethernet/realtek/r8169_main.c
> +++ b/drivers/net/ethernet/realtek/r8169_main.c
> @@ -5224,13 +5224,9 @@ static int rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
>   
>   	/* Disable ASPM L1 as that cause random device stop working
>   	 * problems as well as full system hangs for some PCIe devices users.
> -	 * Chips from RTL8168h partially have issues with L1.2, but seem
> -	 * to work fine with L1 and L1.1.
>   	 */
>   	if (rtl_aspm_is_safe(tp))
>   		rc = 0;
> -	else if (tp->mac_version >= RTL_GIGA_MAC_VER_46)
> -		rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1_2);
>   	else
>   		rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1);
>   	tp->aspm_manageable = !rc;

This is a backport from 6.6-rc1, I suppose you're sending the backport 
out because missing it is causing problems testing either 
amd-staging-drm-next or drm-next, right?

If the problems are amd-staging-drm-next I think you can:
# git cherry-pick -x 90ca51e8c654699b672ba61aeaa418dfb3252e5e
and commit there.  Alex will just skip it when he builds the next PR or 
rebases to a newer release.

If the problem is on drm-next, we'll need to wait for drm-next to move 
up rather than cherry-picking it there.

RE: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Ma, Li]

[AMD Official Use Only - General]

Hi Mario,

Yes, this problem is on amd-staging-drm-next and I checked the adeucher/amd-staging-drm-next, this patch is not in there.

Best Regards,
Ma,Li

-----Original Message-----
From: Limonciello, Mario <Mario.Limonciello@amd.com>
Sent: Wednesday, November 8, 2023 2:17 AM
To: Ma, Li <Li.Ma@amd.com>; amd-gfx@lists.freedesktop.org
Cc: Deucher, Alexander <Alexander.Deucher@amd.com>; Zhang, Yifan <Yifan1.Zhang@amd.com>; Heiner Kallweit <hkallweit1@gmail.com>; stable@vger.kernel.org; David S . Miller <davem@davemloft.net>
Subject: Re: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

On 11/7/2023 02:52, Li Ma wrote:
> From: Heiner Kallweit <hkallweit1@gmail.com>
>
> [Backport: commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e]
> This backport to avoid the bug caused by r8169.
>
> This effectively reverts 4b5f82f6aaef. On a number of systems ASPM L1
> causes tx timeouts with RTL8168h, see referenced bug report.
>
> Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
> Cc: stable@vger.kernel.org
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> Signed-off-by: David S. Miller <davem@davemloft.net>
> ---
>   drivers/net/ethernet/realtek/r8169_main.c | 4 ----
>   1 file changed, 4 deletions(-)
>
> diff --git a/drivers/net/ethernet/realtek/r8169_main.c
> b/drivers/net/ethernet/realtek/r8169_main.c
> index 45147a1016be..27efd07f09ef 100644
> --- a/drivers/net/ethernet/realtek/r8169_main.c
> +++ b/drivers/net/ethernet/realtek/r8169_main.c
> @@ -5224,13 +5224,9 @@ static int rtl_init_one(struct pci_dev *pdev,
> const struct pci_device_id *ent)
>
>       /* Disable ASPM L1 as that cause random device stop working
>        * problems as well as full system hangs for some PCIe devices users.
> -      * Chips from RTL8168h partially have issues with L1.2, but seem
> -      * to work fine with L1 and L1.1.
>        */
>       if (rtl_aspm_is_safe(tp))
>               rc = 0;
> -     else if (tp->mac_version >= RTL_GIGA_MAC_VER_46)
> -             rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1_2);
>       else
>               rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1);
>       tp->aspm_manageable = !rc;

This is a backport from 6.6-rc1, I suppose you're sending the backport out because missing it is causing problems testing either amd-staging-drm-next or drm-next, right?

If the problems are amd-staging-drm-next I think you can:
# git cherry-pick -x 90ca51e8c654699b672ba61aeaa418dfb3252e5e
and commit there.  Alex will just skip it when he builds the next PR or rebases to a newer release.

If the problem is on drm-next, we'll need to wait for drm-next to move up rather than cherry-picking it there.

[PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Li Ma]

From: Heiner Kallweit <hkallweit1@gmail.com>

This effectively reverts 4b5f82f6aaef. On a number of systems ASPM L1
causes tx timeouts with RTL8168h, see referenced bug report.

Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
Cc: stable@vger.kernel.org
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e)
---
 drivers/net/ethernet/realtek/r8169_main.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
index 45147a1016be..27efd07f09ef 100644
--- a/drivers/net/ethernet/realtek/r8169_main.c
+++ b/drivers/net/ethernet/realtek/r8169_main.c
@@ -5224,13 +5224,9 @@ static int rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	/* Disable ASPM L1 as that cause random device stop working
 	 * problems as well as full system hangs for some PCIe devices users.
-	 * Chips from RTL8168h partially have issues with L1.2, but seem
-	 * to work fine with L1 and L1.1.
 	 */
 	if (rtl_aspm_is_safe(tp))
 		rc = 0;
-	else if (tp->mac_version >= RTL_GIGA_MAC_VER_46)
-		rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1_2);
 	else
 		rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L1);
 	tp->aspm_manageable = !rc;
-- 
2.25.1

Re: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Greg KH]

On Wed, Nov 08, 2023 at 11:34:00AM +0800, Li Ma wrote:
> From: Heiner Kallweit <hkallweit1@gmail.com>
> 
> This effectively reverts 4b5f82f6aaef. On a number of systems ASPM L1
> causes tx timeouts with RTL8168h, see referenced bug report.
> 
> Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
> Cc: stable@vger.kernel.org
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> Signed-off-by: David S. Miller <davem@davemloft.net>
> (cherry picked from commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e)
> ---
>  drivers/net/ethernet/realtek/r8169_main.c | 4 ----
>  1 file changed, 4 deletions(-)

Is this a proposed stable tree patch?  If so, what kernel(s) are you
wanting it applied to?

thanks,

greg k-h

Re: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Heiner Kallweit]

On 08.11.2023 08:05, Greg KH wrote:
> On Wed, Nov 08, 2023 at 11:34:00AM +0800, Li Ma wrote:
>> From: Heiner Kallweit <hkallweit1@gmail.com>
>>
>> This effectively reverts 4b5f82f6aaef. On a number of systems ASPM L1
>> causes tx timeouts with RTL8168h, see referenced bug report.
>>
>> Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
>> Cc: stable@vger.kernel.org
>> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
>> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
>> Signed-off-by: David S. Miller <davem@davemloft.net>
>> (cherry picked from commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e)
>> ---
>>  drivers/net/ethernet/realtek/r8169_main.c | 4 ----
>>  1 file changed, 4 deletions(-)
> 
> Is this a proposed stable tree patch?  If so, what kernel(s) are you
> wanting it applied to?
> 
This should have been sent neither to you nor stable@vger.kernel.org.
This patch has been applied to stable already, the mail is something
AMD-internal it seems.

> thanks,
> 
> greg k-h

Heiner

Re: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Greg KH]

On Wed, Nov 08, 2023 at 08:40:48AM +0100, Heiner Kallweit wrote:
> On 08.11.2023 08:05, Greg KH wrote:
> > On Wed, Nov 08, 2023 at 11:34:00AM +0800, Li Ma wrote:
> >> From: Heiner Kallweit <hkallweit1@gmail.com>
> >>
> >> This effectively reverts 4b5f82f6aaef. On a number of systems ASPM L1
> >> causes tx timeouts with RTL8168h, see referenced bug report.
> >>
> >> Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
> >> Cc: stable@vger.kernel.org
> >> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
> >> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> >> Signed-off-by: David S. Miller <davem@davemloft.net>
> >> (cherry picked from commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e)
> >> ---
> >>  drivers/net/ethernet/realtek/r8169_main.c | 4 ----
> >>  1 file changed, 4 deletions(-)
> > 
> > Is this a proposed stable tree patch?  If so, what kernel(s) are you
> > wanting it applied to?
> > 
> This should have been sent neither to you nor stable@vger.kernel.org.
> This patch has been applied to stable already, the mail is something
> AMD-internal it seems.

Then someone needs to seriously fix their scripts as it is very
confusing :(

RE: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

[Ma, Li]

[AMD Official Use Only - General]

Hi Greg KH & Heiner Kallweit,

Sorry for the trouble, this was sent to AMD internel. Because I want to backport this patch to avoid a problem caused by r8169.
I forgot to clean up the sent object.
Sorry again for the trouble caused by this misunderstanding.

Best Regads,
Li

-----Original Message-----
From: Greg KH <gregkh@linuxfoundation.org>
Sent: Wednesday, November 8, 2023 3:53 PM
To: Heiner Kallweit <hkallweit1@gmail.com>
Cc: amd-gfx@lists.freedesktop.org; Deucher, Alexander <Alexander.Deucher@amd.com>; Limonciello, Mario <Mario.Limonciello@amd.com>; Zhang, Yifan <Yifan1.Zhang@amd.com>; stable@vger.kernel.org; David S . Miller <davem@davemloft.net>; Ma, Li <Li.Ma@amd.com>
Subject: Re: [PATCH] r8169: fix ASPM-related issues on a number of systems with NIC version from RTL8168h

Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.


On Wed, Nov 08, 2023 at 08:40:48AM +0100, Heiner Kallweit wrote:
> On 08.11.2023 08:05, Greg KH wrote:
> > On Wed, Nov 08, 2023 at 11:34:00AM +0800, Li Ma wrote:
> >> From: Heiner Kallweit <hkallweit1@gmail.com>
> >>
> >> This effectively reverts 4b5f82f6aaef. On a number of systems ASPM
> >> L1 causes tx timeouts with RTL8168h, see referenced bug report.
> >>
> >> Fixes: 4b5f82f6aaef ("r8169: enable ASPM L1/L1.1 from RTL8168h")
> >> Cc: stable@vger.kernel.org
> >> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217814
> >> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> >> Signed-off-by: David S. Miller <davem@davemloft.net> (cherry picked
> >> from commit 90ca51e8c654699b672ba61aeaa418dfb3252e5e)
> >> ---
> >>  drivers/net/ethernet/realtek/r8169_main.c | 4 ----
> >>  1 file changed, 4 deletions(-)
> >
> > Is this a proposed stable tree patch?  If so, what kernel(s) are you
> > wanting it applied to?
> >
> This should have been sent neither to you nor stable@vger.kernel.org.
> This patch has been applied to stable already, the mail is something
> AMD-internal it seems.

Then someone needs to seriously fix their scripts as it is very confusing :(