From: Heiner Kallweit <hkallweit1@gmail.com>
To: Kai-Heng Feng <kai.heng.feng@canonical.com>
Cc: nic_swsd <nic_swsd@realtek.com>,
"David S. Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>,
"open list:8169 10/100/1000 GIGABIT ETHERNET DRIVER"
<netdev@vger.kernel.org>,
open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2 2/2] r8169: Enable ASPM for selected NICs
Date: Sat, 14 Aug 2021 13:23:09 +0200 [thread overview]
Message-ID: <0c62c140-94d7-96ff-a837-0f9f9086d120@gmail.com> (raw)
In-Reply-To: <CAAd53p6KGTVWp5PgZux5t2mdTXK29XdyB5Ke5YbR_-UacR03Sg@mail.gmail.com>
On 13.08.2021 12:11, Kai-Heng Feng wrote:
> On Fri, Aug 13, 2021 at 3:39 AM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>
>> On 12.08.2021 17:53, Kai-Heng Feng wrote:
>>> The latest vendor driver enables ASPM for more recent r8168 NICs, do the
>>> same here to match the behavior.
>>>
>>> In addition, pci_disable_link_state() is only used for RTL8168D/8111D in
>>> vendor driver, also match that.
>>>
>>> Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
>>> ---
>>> v2:
>>> - No change
>>>
>>> drivers/net/ethernet/realtek/r8169_main.c | 34 +++++++++++++++++------
>>> 1 file changed, 26 insertions(+), 8 deletions(-)
>>>
>>> diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
>>> index 7ab2e841dc69..caa29e72a21a 100644
>>> --- a/drivers/net/ethernet/realtek/r8169_main.c
>>> +++ b/drivers/net/ethernet/realtek/r8169_main.c
>>> @@ -623,7 +623,7 @@ struct rtl8169_private {
>>> } wk;
>>>
>>> unsigned supports_gmii:1;
>>> - unsigned aspm_manageable:1;
>>> + unsigned aspm_supported:1;
>>> unsigned aspm_enabled:1;
>>> struct delayed_work aspm_toggle;
>>> struct mutex aspm_mutex;
>>> @@ -2667,8 +2667,11 @@ static void rtl_pcie_state_l2l3_disable(struct rtl8169_private *tp)
>>>
>>> static void rtl_hw_aspm_clkreq_enable(struct rtl8169_private *tp, bool enable)
>>> {
>>> + if (!tp->aspm_supported)
>>> + return;
>>> +
>>> /* Don't enable ASPM in the chip if OS can't control ASPM */
>>> - if (enable && tp->aspm_manageable) {
>>> + if (enable) {
>>> RTL_W8(tp, Config5, RTL_R8(tp, Config5) | ASPM_en);
>>> RTL_W8(tp, Config2, RTL_R8(tp, Config2) | ClkReqEn);
>>> } else {
>>> @@ -5284,6 +5287,21 @@ static void rtl_init_mac_address(struct rtl8169_private *tp)
>>> rtl_rar_set(tp, mac_addr);
>>> }
>>>
>>> +static int rtl_hw_aspm_supported(struct rtl8169_private *tp)
>>> +{
>>> + switch (tp->mac_version) {
>>> + case RTL_GIGA_MAC_VER_32 ... RTL_GIGA_MAC_VER_36:
>>> + case RTL_GIGA_MAC_VER_38:
>>> + case RTL_GIGA_MAC_VER_40 ... RTL_GIGA_MAC_VER_42:
>>> + case RTL_GIGA_MAC_VER_44 ... RTL_GIGA_MAC_VER_46:
>>> + case RTL_GIGA_MAC_VER_49 ... RTL_GIGA_MAC_VER_63:
>>
>> This shouldn't be needed because ASPM support is announced the
>> standard PCI way. Max a blacklist should be needed if there are
>> chip versions that announce ASPM support whilst in reality they
>> do not support it (or support is completely broken).
>
> So can we also remove aspm_manageable since blacklist will be used?
>
That's independent. What I mean is replace the whitelist with auto-
detected ASPM support and blacklist just the ones that are where
ASPM is completely unusable.
Retrieving the info about ASPM support may need a smll PCI core
extension. We need something similar to pcie_aspm_enabled(),
just exposing link->aspm_support. link->aspm_enabled may change
at runtime (sysfs link attributes).
> Kai-Heng
>
>>
>>> + return 1;
>>> +
>>> + default:
>>> + return 0;
>>> + }
>>> +}
>>> +
>>> static int rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
>>> {
>>> struct rtl8169_private *tp;
>>> @@ -5315,12 +5333,12 @@ static int rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
>>> if (rc)
>>> return rc;
>>>
>>> - /* Disable ASPM completely as that cause random device stop working
>>> - * problems as well as full system hangs for some PCIe devices users.
>>> - */
>>> - rc = pci_disable_link_state(pdev, PCIE_LINK_STATE_L0S |
>>> - PCIE_LINK_STATE_L1);
>>> - tp->aspm_manageable = !rc;
>>> + if (tp->mac_version == RTL_GIGA_MAC_VER_25)
>>> + pci_disable_link_state(pdev, PCIE_LINK_STATE_L0S |
>>> + PCIE_LINK_STATE_L1 |
>>> + PCIE_LINK_STATE_CLKPM);
>>> +
>>> + tp->aspm_supported = rtl_hw_aspm_supported(tp);
>>>
>>> /* enable device (incl. PCI PM wakeup and hotplug setup) */
>>> rc = pcim_enable_device(pdev);
>>>
>>
next prev parent reply other threads:[~2021-08-14 11:34 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-12 15:53 [PATCH v2 1/2] r8169: Implement dynamic ASPM mechanism Kai-Heng Feng
2021-08-12 15:53 ` [PATCH v2 2/2] r8169: Enable ASPM for selected NICs Kai-Heng Feng
2021-08-12 19:38 ` Heiner Kallweit
2021-08-13 10:11 ` Kai-Heng Feng
2021-08-14 11:23 ` Heiner Kallweit [this message]
2021-08-12 19:34 ` [PATCH v2 1/2] r8169: Implement dynamic ASPM mechanism Heiner Kallweit
2021-08-13 9:46 ` Kai-Heng Feng
2021-08-14 11:33 ` Heiner Kallweit
2021-08-13 6:48 ` Heiner Kallweit
2021-08-13 9:54 ` Kai-Heng Feng
2021-08-14 11:31 ` Heiner Kallweit
2021-08-19 3:11 ` Kai-Heng Feng
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=0c62c140-94d7-96ff-a837-0f9f9086d120@gmail.com \
--to=hkallweit1@gmail.com \
--cc=davem@davemloft.net \
--cc=kai.heng.feng@canonical.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=nic_swsd@realtek.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).