linux-kernel - Re: [PATCH net-next 1/1] r8169: enable RTL8125 ASPM L1.2

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 26 Jan 2022 11:42:35 +0100
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Hau <hau@...ltek.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Cc:     nic_swsd <nic_swsd@...ltek.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH net-next 1/1] r8169: enable RTL8125 ASPM L1.2

On 26.01.2022 10:02, Hau wrote:
> 
> 
>> On 24.01.2022 19:19, Chunhao Lin wrote:
>>> This patch will enable RTL8125 ASPM L1.2 on the platforms that have
>>> tested RTL8125 with ASPM L1.2 enabled.
>>> Register mac ocp 0xc0b2 will help to identify if RTL8125 has been
>>> tested on L1.2 enabled platform. If it is, this register will be set to 0xf.
>>> If not, this register will be default value 0.
>>>
>>> Signed-off-by: Chunhao Lin <hau@...ltek.com>
>>> ---
>>>  drivers/net/ethernet/realtek/r8169_main.c | 99
>>> ++++++++++++++++++-----
>>>  1 file changed, 79 insertions(+), 20 deletions(-)
>>>
>>> diff --git a/drivers/net/ethernet/realtek/r8169_main.c
>>> b/drivers/net/ethernet/realtek/r8169_main.c
>>> index 19e2621e0645..b1e013969d4c 100644
>>> --- a/drivers/net/ethernet/realtek/r8169_main.c
>>> +++ b/drivers/net/ethernet/realtek/r8169_main.c
>>> @@ -2238,21 +2238,6 @@ static void rtl_wol_enable_rx(struct
>> rtl8169_private *tp)
>>>  			AcceptBroadcast | AcceptMulticast |
>> AcceptMyPhys);  }
>>>
>>> -static void rtl_prepare_power_down(struct rtl8169_private *tp) -{
>>> -	if (tp->dash_type != RTL_DASH_NONE)
>>> -		return;
>>> -
>>> -	if (tp->mac_version == RTL_GIGA_MAC_VER_32 ||
>>> -	    tp->mac_version == RTL_GIGA_MAC_VER_33)
>>> -		rtl_ephy_write(tp, 0x19, 0xff64);
>>> -
>>> -	if (device_may_wakeup(tp_to_dev(tp))) {
>>> -		phy_speed_down(tp->phydev, false);
>>> -		rtl_wol_enable_rx(tp);
>>> -	}
>>> -}
>>> -
>>>  static void rtl_init_rxcfg(struct rtl8169_private *tp)  {
>>>  	switch (tp->mac_version) {
>>> @@ -2650,6 +2635,34 @@ static void rtl_pcie_state_l2l3_disable(struct
>> rtl8169_private *tp)
>>>  	RTL_W8(tp, Config3, RTL_R8(tp, Config3) & ~Rdy_to_L23);  }
>>>
>>> +static void rtl_disable_exit_l1(struct rtl8169_private *tp) {
>>> +	/* Bits control which events trigger ASPM L1 exit:
>>> +	 * Bit 12: rxdv
>>> +	 * Bit 11: ltr_msg
>>> +	 * Bit 10: txdma_poll
>>> +	 * Bit  9: xadm
>>> +	 * Bit  8: pktavi
>>> +	 * Bit  7: txpla
>>> +	 */
>>> +	switch (tp->mac_version) {
>>> +	case RTL_GIGA_MAC_VER_34 ... RTL_GIGA_MAC_VER_36:
>>> +		rtl_eri_clear_bits(tp, 0xd4, 0x1f00);
>>> +		break;
>>> +	case RTL_GIGA_MAC_VER_37 ... RTL_GIGA_MAC_VER_38:
>>> +		rtl_eri_clear_bits(tp, 0xd4, 0x0c00);
>>> +		break;
>>> +	case RTL_GIGA_MAC_VER_40 ... RTL_GIGA_MAC_VER_53:
>>> +		rtl_eri_clear_bits(tp, 0xd4, 0x1f80);
>>> +		break;
>>> +	case RTL_GIGA_MAC_VER_60 ... RTL_GIGA_MAC_VER_63:
>>> +		r8168_mac_ocp_modify(tp, 0xc0ac, 0x1f80, 0);
>>> +		break;
>>> +	default:
>>> +		break;
>>> +	}
>>> +}
>>> +
>>>  static void rtl_enable_exit_l1(struct rtl8169_private *tp)  {
>>>  	/* Bits control which events trigger ASPM L1 exit:
>>> @@ -2692,6 +2705,33 @@ static void rtl_hw_aspm_clkreq_enable(struct
>> rtl8169_private *tp, bool enable)
>>>  	udelay(10);
>>>  }
>>>
>>> +static void rtl_hw_aspm_l12_enable(struct rtl8169_private *tp, bool
>>> +enable) {
>>> +	/* Don't enable L1.2 in the chip if OS can't control ASPM */
>>> +	if (enable && tp->aspm_manageable) {
>>> +		r8168_mac_ocp_modify(tp, 0xe094, 0xff00, 0);
>>> +		r8168_mac_ocp_modify(tp, 0xe092, 0x00ff, BIT(2));
>>> +	} else {
>>> +		r8168_mac_ocp_modify(tp, 0xe092, 0x00ff, 0);
>>> +	}
>>> +}
>>> +
>>
>> Register E094 bits 0..15 are cleared when enabling, but not touched on
>> disabling. I this correct?
>    Register E094 bits 8...15 is a timer counter that is used to control when to disable ephy tx/rx.
>    Set it to 0 means disable ephy tx/rx immediately when certain condition meet. 
>    It has no meaning when register E092 bit 2 is set to 0.
> 
Thanks for the explanation.

>> And for basically the same purpose we have the following function.
>> "don't enable L1.2 in the chip" is not covered by ASPM_en in Config5?
>    Register E092 is like  ASPM_en in Config5. But it controls L1 substate (L1.1/L1.2) enable status.
> 
How is this handled for the RTL8168 chip versions supporting L1 sub-states (RTL8168h)?
Is there a similar register or does Config5 ASPM_en control also the L1 substates
on these chip versions?

>>
>> static void rtl_hw_aspm_clkreq_enable(struct rtl8169_private *tp, bool
>> enable) {
>> 	/* Don't enable ASPM in the chip if OS can't control ASPM */
>> 	if (enable && tp->aspm_manageable) {
>> 		RTL_W8(tp, Config5, RTL_R8(tp, Config5) | ASPM_en);
>> 		RTL_W8(tp, Config2, RTL_R8(tp, Config2) | ClkReqEn);
>> 	} else {
>> 		RTL_W8(tp, Config2, RTL_R8(tp, Config2) & ~ClkReqEn);
>> 		RTL_W8(tp, Config5, RTL_R8(tp, Config5) & ~ASPM_en);
>> 	}
>>
>> 	udelay(10);
>> }
>>
>>
...