From patchwork Mon Nov 18 09:54:29 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tony Chuang X-Patchwork-Id: 11249239 X-Patchwork-Delegate: kvalo@adurom.com Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 0C6291393 for ; Mon, 18 Nov 2019 09:54:48 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id E90D820692 for ; Mon, 18 Nov 2019 09:54:47 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726518AbfKRJyo (ORCPT ); Mon, 18 Nov 2019 04:54:44 -0500 Received: from rtits2.realtek.com ([211.75.126.72]:40035 "EHLO rtits2.realtek.com.tw" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726460AbfKRJyn (ORCPT ); Mon, 18 Nov 2019 04:54:43 -0500 Authenticated-By: X-SpamFilter-By: BOX Solutions SpamTrap 5.62 with qID xAI9sbHT018251, This message is accepted by code: ctloc85258 Received: from mail.realtek.com (RTITCASV01.realtek.com.tw[172.21.6.18]) by rtits2.realtek.com.tw (8.15.2/2.57/5.78) with ESMTPS id xAI9sbHT018251 (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Mon, 18 Nov 2019 17:54:37 +0800 Received: from localhost.localdomain (172.21.68.126) by RTITCASV01.realtek.com.tw (172.21.6.18) with Microsoft SMTP Server id 14.3.468.0; Mon, 18 Nov 2019 17:54:36 +0800 From: To: CC: , Subject: [PATCH v2 1/4] rtw88: pci: use macros to access PCI DBI/MDIO registers Date: Mon, 18 Nov 2019 17:54:29 +0800 Message-ID: <20191118095432.4507-2-yhchuang@realtek.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20191118095432.4507-1-yhchuang@realtek.com> References: <20191118095432.4507-1-yhchuang@realtek.com> MIME-Version: 1.0 X-Originating-IP: [172.21.68.126] Sender: linux-wireless-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-wireless@vger.kernel.org From: Yan-Hsuan Chuang Add some register and bit macros to access DBI/MDIO register. This should not change the logic. Signed-off-by: Yan-Hsuan Chuang --- v1 -> v2 * no change drivers/net/wireless/realtek/rtw88/pci.c | 21 ++++++++++----------- drivers/net/wireless/realtek/rtw88/pci.h | 10 ++++++++++ 2 files changed, 20 insertions(+), 11 deletions(-) diff --git a/drivers/net/wireless/realtek/rtw88/pci.c b/drivers/net/wireless/realtek/rtw88/pci.c index 17b9cdf9cb05..b158ef8ded17 100644 --- a/drivers/net/wireless/realtek/rtw88/pci.c +++ b/drivers/net/wireless/realtek/rtw88/pci.c @@ -1060,14 +1060,15 @@ static void rtw_pci_io_unmapping(struct rtw_dev *rtwdev, static void rtw_dbi_write8(struct rtw_dev *rtwdev, u16 addr, u8 data) { u16 write_addr; - u16 remainder = addr & 0x3; + u16 remainder = addr & ~(BITS_DBI_WREN | BITS_DBI_ADDR_MASK); u8 flag; - u8 cnt = 20; + u8 cnt = RTW_PCI_WR_RETRY_CNT; - write_addr = ((addr & 0x0ffc) | (BIT(0) << (remainder + 12))); + write_addr = addr & BITS_DBI_ADDR_MASK; + write_addr |= u16_encode_bits(BIT(remainder), BITS_DBI_WREN); rtw_write8(rtwdev, REG_DBI_WDATA_V1 + remainder, data); rtw_write16(rtwdev, REG_DBI_FLAG_V1, write_addr); - rtw_write8(rtwdev, REG_DBI_FLAG_V1 + 2, 0x01); + rtw_write8(rtwdev, REG_DBI_FLAG_V1 + 2, BIT_DBI_WFLAG >> 16); flag = rtw_read8(rtwdev, REG_DBI_FLAG_V1 + 2); while (flag && (cnt != 0)) { @@ -1083,19 +1084,17 @@ static void rtw_mdio_write(struct rtw_dev *rtwdev, u8 addr, u16 data, bool g1) { u8 page; u8 wflag; - u8 cnt; + u8 cnt = RTW_PCI_WR_RETRY_CNT; rtw_write16(rtwdev, REG_MDIO_V1, data); - page = addr < 0x20 ? 0 : 1; - page += g1 ? 0 : 2; - rtw_write8(rtwdev, REG_PCIE_MIX_CFG, addr & 0x1f); + page = addr < RTW_PCI_MDIO_PG_SZ ? 0 : 1; + page += g1 ? RTW_PCI_MDIO_PG_OFFS_G1 : RTW_PCI_MDIO_PG_OFFS_G2; + rtw_write8(rtwdev, REG_PCIE_MIX_CFG, addr & BITS_MDIO_ADDR_MASK); rtw_write8(rtwdev, REG_PCIE_MIX_CFG + 3, page); - rtw_write32_mask(rtwdev, REG_PCIE_MIX_CFG, BIT_MDIO_WFLAG_V1, 1); - wflag = rtw_read32_mask(rtwdev, REG_PCIE_MIX_CFG, BIT_MDIO_WFLAG_V1); - cnt = 20; + wflag = rtw_read32_mask(rtwdev, REG_PCIE_MIX_CFG, BIT_MDIO_WFLAG_V1); while (wflag && (cnt != 0)) { udelay(10); wflag = rtw_read32_mask(rtwdev, REG_PCIE_MIX_CFG, diff --git a/drivers/net/wireless/realtek/rtw88/pci.h b/drivers/net/wireless/realtek/rtw88/pci.h index 87824a4caba9..50aff49738d4 100644 --- a/drivers/net/wireless/realtek/rtw88/pci.h +++ b/drivers/net/wireless/realtek/rtw88/pci.h @@ -21,9 +21,19 @@ #define BIT_RX_TAG_EN BIT(15) #define REG_DBI_WDATA_V1 0x03E8 #define REG_DBI_FLAG_V1 0x03F0 +#define BIT_DBI_RFLAG BIT(17) +#define BIT_DBI_WFLAG BIT(16) +#define BITS_DBI_WREN GENMASK(15, 12) +#define BITS_DBI_ADDR_MASK GENMASK(11, 2) + #define REG_MDIO_V1 0x03F4 #define REG_PCIE_MIX_CFG 0x03F8 +#define BITS_MDIO_ADDR_MASK GENMASK(4, 0) #define BIT_MDIO_WFLAG_V1 BIT(5) +#define RTW_PCI_MDIO_PG_SZ BIT(5) +#define RTW_PCI_MDIO_PG_OFFS_G1 0 +#define RTW_PCI_MDIO_PG_OFFS_G2 2 +#define RTW_PCI_WR_RETRY_CNT 20 #define BIT_PCI_BCNQ_FLAG BIT(4) #define RTK_PCI_TXBD_DESA_BCNQ 0x308 From patchwork Mon Nov 18 09:54:30 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tony Chuang X-Patchwork-Id: 11249233 X-Patchwork-Delegate: kvalo@adurom.com Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id A432C1593 for ; Mon, 18 Nov 2019 09:54:45 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 8C76E20692 for ; Mon, 18 Nov 2019 09:54:45 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726539AbfKRJyo (ORCPT ); Mon, 18 Nov 2019 04:54:44 -0500 Received: from rtits2.realtek.com ([211.75.126.72]:40038 "EHLO rtits2.realtek.com.tw" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726490AbfKRJyo (ORCPT ); Mon, 18 Nov 2019 04:54:44 -0500 Authenticated-By: X-SpamFilter-By: BOX Solutions SpamTrap 5.62 with qID xAI9scV0018255, This message is accepted by code: ctloc85258 Received: from mail.realtek.com (RTITCASV01.realtek.com.tw[172.21.6.18]) by rtits2.realtek.com.tw (8.15.2/2.57/5.78) with ESMTPS id xAI9scV0018255 (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Mon, 18 Nov 2019 17:54:38 +0800 Received: from localhost.localdomain (172.21.68.126) by RTITCASV01.realtek.com.tw (172.21.6.18) with Microsoft SMTP Server id 14.3.468.0; Mon, 18 Nov 2019 17:54:37 +0800 From: To: CC: , Subject: [PATCH v2 2/4] rtw88: pci: use for loop instead of while loop for DBI/MDIO Date: Mon, 18 Nov 2019 17:54:30 +0800 Message-ID: <20191118095432.4507-3-yhchuang@realtek.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20191118095432.4507-1-yhchuang@realtek.com> References: <20191118095432.4507-1-yhchuang@realtek.com> MIME-Version: 1.0 X-Originating-IP: [172.21.68.126] Sender: linux-wireless-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-wireless@vger.kernel.org From: Yan-Hsuan Chuang Use a for loop to polling DBI/MDIO read/write flags to avoid infinite loop happens Signed-off-by: Yan-Hsuan Chuang Reviewed-by: Chris Chiu --- v1 -> v2 * no change drivers/net/wireless/realtek/rtw88/pci.c | 26 +++++++++++++----------- 1 file changed, 14 insertions(+), 12 deletions(-) diff --git a/drivers/net/wireless/realtek/rtw88/pci.c b/drivers/net/wireless/realtek/rtw88/pci.c index b158ef8ded17..6d1aa6f41e84 100644 --- a/drivers/net/wireless/realtek/rtw88/pci.c +++ b/drivers/net/wireless/realtek/rtw88/pci.c @@ -1062,7 +1062,7 @@ static void rtw_dbi_write8(struct rtw_dev *rtwdev, u16 addr, u8 data) u16 write_addr; u16 remainder = addr & ~(BITS_DBI_WREN | BITS_DBI_ADDR_MASK); u8 flag; - u8 cnt = RTW_PCI_WR_RETRY_CNT; + u8 cnt; write_addr = addr & BITS_DBI_ADDR_MASK; write_addr |= u16_encode_bits(BIT(remainder), BITS_DBI_WREN); @@ -1070,21 +1070,22 @@ static void rtw_dbi_write8(struct rtw_dev *rtwdev, u16 addr, u8 data) rtw_write16(rtwdev, REG_DBI_FLAG_V1, write_addr); rtw_write8(rtwdev, REG_DBI_FLAG_V1 + 2, BIT_DBI_WFLAG >> 16); - flag = rtw_read8(rtwdev, REG_DBI_FLAG_V1 + 2); - while (flag && (cnt != 0)) { - udelay(10); + for (cnt = 0; cnt < RTW_PCI_WR_RETRY_CNT; cnt++) { flag = rtw_read8(rtwdev, REG_DBI_FLAG_V1 + 2); - cnt--; + if (flag == 0) + return; + + udelay(10); } - WARN(flag, "DBI write fail\n"); + WARN(flag, "failed to write to DBI register, addr=0x%04x\n", addr); } static void rtw_mdio_write(struct rtw_dev *rtwdev, u8 addr, u16 data, bool g1) { u8 page; u8 wflag; - u8 cnt = RTW_PCI_WR_RETRY_CNT; + u8 cnt; rtw_write16(rtwdev, REG_MDIO_V1, data); @@ -1094,15 +1095,16 @@ static void rtw_mdio_write(struct rtw_dev *rtwdev, u8 addr, u16 data, bool g1) rtw_write8(rtwdev, REG_PCIE_MIX_CFG + 3, page); rtw_write32_mask(rtwdev, REG_PCIE_MIX_CFG, BIT_MDIO_WFLAG_V1, 1); - wflag = rtw_read32_mask(rtwdev, REG_PCIE_MIX_CFG, BIT_MDIO_WFLAG_V1); - while (wflag && (cnt != 0)) { - udelay(10); + for (cnt = 0; cnt < RTW_PCI_WR_RETRY_CNT; cnt++) { wflag = rtw_read32_mask(rtwdev, REG_PCIE_MIX_CFG, BIT_MDIO_WFLAG_V1); - cnt--; + if (wflag == 0) + return; + + udelay(10); } - WARN(wflag, "MDIO write fail\n"); + WARN(wflag, "failed to write to MDIO register, addr=0x%02x\n", addr); } static void rtw_pci_phy_cfg(struct rtw_dev *rtwdev) From patchwork Mon Nov 18 09:54:31 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tony Chuang X-Patchwork-Id: 11249231 X-Patchwork-Delegate: kvalo@adurom.com Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id DBD1F1393 for ; Mon, 18 Nov 2019 09:54:44 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id C23D52071C for ; Mon, 18 Nov 2019 09:54:44 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726563AbfKRJyo (ORCPT ); Mon, 18 Nov 2019 04:54:44 -0500 Received: from rtits2.realtek.com ([211.75.126.72]:40037 "EHLO rtits2.realtek.com.tw" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726488AbfKRJyo (ORCPT ); Mon, 18 Nov 2019 04:54:44 -0500 Authenticated-By: X-SpamFilter-By: BOX Solutions SpamTrap 5.62 with qID xAI9scDr018260, This message is accepted by code: ctloc85258 Received: from mail.realtek.com (RTITCASV01.realtek.com.tw[172.21.6.18]) by rtits2.realtek.com.tw (8.15.2/2.57/5.78) with ESMTPS id xAI9scDr018260 (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Mon, 18 Nov 2019 17:54:38 +0800 Received: from localhost.localdomain (172.21.68.126) by RTITCASV01.realtek.com.tw (172.21.6.18) with Microsoft SMTP Server id 14.3.468.0; Mon, 18 Nov 2019 17:54:38 +0800 From: To: CC: , Subject: [PATCH v2 3/4] rtw88: pci: enable CLKREQ function if host supports it Date: Mon, 18 Nov 2019 17:54:31 +0800 Message-ID: <20191118095432.4507-4-yhchuang@realtek.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20191118095432.4507-1-yhchuang@realtek.com> References: <20191118095432.4507-1-yhchuang@realtek.com> MIME-Version: 1.0 X-Originating-IP: [172.21.68.126] Sender: linux-wireless-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-wireless@vger.kernel.org From: Yan-Hsuan Chuang By Realtek's design, there are two HW modules associated for CLKREQ, one is responsible to follow the PCIE host settings, and another is to actually working on it. But the module that is actually working on it is default disabled, and driver should enable that module if host and device have successfully sync'ed with each other. The module is default disabled because sometimes the host does not support it, and if there is any incorrect settings (ex. CLKREQ# is not Bi-Direction), device can be lost and disconnected to the host. So driver should first check after host and device are sync'ed, and the host does support the function and set it in configuration space, then driver can turn on the HW module to working on it. Signed-off-by: Yan-Hsuan Chuang Reviewed-by: Chris Chiu --- v1 -> v2 * slightly modify code flow of rtw_dbi_read8() to look cleaner drivers/net/wireless/realtek/rtw88/pci.c | 80 ++++++++++++++++++++++++ drivers/net/wireless/realtek/rtw88/pci.h | 5 ++ 2 files changed, 85 insertions(+) diff --git a/drivers/net/wireless/realtek/rtw88/pci.c b/drivers/net/wireless/realtek/rtw88/pci.c index 6d1aa6f41e84..6e99aad39487 100644 --- a/drivers/net/wireless/realtek/rtw88/pci.c +++ b/drivers/net/wireless/realtek/rtw88/pci.c @@ -1081,6 +1081,30 @@ static void rtw_dbi_write8(struct rtw_dev *rtwdev, u16 addr, u8 data) WARN(flag, "failed to write to DBI register, addr=0x%04x\n", addr); } +static int rtw_dbi_read8(struct rtw_dev *rtwdev, u16 addr, u8 *value) +{ + u16 read_addr = addr & BITS_DBI_ADDR_MASK; + u8 flag; + u8 cnt; + + rtw_write16(rtwdev, REG_DBI_FLAG_V1, read_addr); + rtw_write8(rtwdev, REG_DBI_FLAG_V1 + 2, BIT_DBI_RFLAG >> 16); + + for (cnt = 0; cnt < RTW_PCI_WR_RETRY_CNT; cnt++) { + flag = rtw_read8(rtwdev, REG_DBI_FLAG_V1 + 2); + if (flag == 0) { + read_addr = REG_DBI_RDATA_V1 + (addr & 3); + *value = rtw_read8(rtwdev, read_addr); + return 0; + } + + udelay(10); + } + + WARN(1, "failed to read DBI register, addr=0x%04x\n", addr); + return -EIO; +} + static void rtw_mdio_write(struct rtw_dev *rtwdev, u8 addr, u16 data, bool g1) { u8 page; @@ -1107,6 +1131,60 @@ static void rtw_mdio_write(struct rtw_dev *rtwdev, u8 addr, u16 data, bool g1) WARN(wflag, "failed to write to MDIO register, addr=0x%02x\n", addr); } +static void rtw_pci_clkreq_set(struct rtw_dev *rtwdev, bool enable) +{ + u8 value; + int ret; + + ret = rtw_dbi_read8(rtwdev, RTK_PCIE_LINK_CFG, &value); + if (ret) { + rtw_err(rtwdev, "failed to read CLKREQ_L1, ret=%d", ret); + return; + } + + if (enable) + value |= BIT_CLKREQ_SW_EN; + else + value &= ~BIT_CLKREQ_SW_EN; + + rtw_dbi_write8(rtwdev, RTK_PCIE_LINK_CFG, value); +} + +static void rtw_pci_link_cfg(struct rtw_dev *rtwdev) +{ + struct rtw_pci *rtwpci = (struct rtw_pci *)rtwdev->priv; + struct pci_dev *pdev = rtwpci->pdev; + u16 link_ctrl; + int ret; + + /* Though there is standard PCIE configuration space to set the + * link control register, but by Realtek's design, driver should + * check if host supports CLKREQ/ASPM to enable the HW module. + * + * These functions are implemented by two HW modules associated, + * one is responsible to access PCIE configuration space to + * follow the host settings, and another is in charge of doing + * CLKREQ/ASPM mechanisms, it is default disabled. Because sometimes + * the host does not support it, and due to some reasons or wrong + * settings (ex. CLKREQ# not Bi-Direction), it could lead to device + * loss if HW misbehaves on the link. + * + * Hence it's designed that driver should first check the PCIE + * configuration space is sync'ed and enabled, then driver can turn + * on the other module that is actually working on the mechanism. + */ + ret = pcie_capability_read_word(pdev, PCI_EXP_LNKCTL, &link_ctrl); + if (ret) { + rtw_err(rtwdev, "failed to read PCI cap, ret=%d\n", ret); + return; + } + + if (link_ctrl & PCI_EXP_LNKCTL_CLKREQ_EN) + rtw_pci_clkreq_set(rtwdev, true); + + rtwpci->link_ctrl = link_ctrl; +} + static void rtw_pci_phy_cfg(struct rtw_dev *rtwdev) { struct rtw_chip_info *chip = rtwdev->chip; @@ -1145,6 +1223,8 @@ static void rtw_pci_phy_cfg(struct rtw_dev *rtwdev) else rtw_dbi_write8(rtwdev, offset, value); } + + rtw_pci_link_cfg(rtwdev); } static int rtw_pci_claim(struct rtw_dev *rtwdev, struct pci_dev *pdev) diff --git a/drivers/net/wireless/realtek/rtw88/pci.h b/drivers/net/wireless/realtek/rtw88/pci.h index 50aff49738d4..90efb73c607e 100644 --- a/drivers/net/wireless/realtek/rtw88/pci.h +++ b/drivers/net/wireless/realtek/rtw88/pci.h @@ -20,6 +20,7 @@ #define BIT_RST_TRXDMA_INTF BIT(20) #define BIT_RX_TAG_EN BIT(15) #define REG_DBI_WDATA_V1 0x03E8 +#define REG_DBI_RDATA_V1 0x03EC #define REG_DBI_FLAG_V1 0x03F0 #define BIT_DBI_RFLAG BIT(17) #define BIT_DBI_WFLAG BIT(16) @@ -35,6 +36,9 @@ #define RTW_PCI_MDIO_PG_OFFS_G2 2 #define RTW_PCI_WR_RETRY_CNT 20 +#define RTK_PCIE_LINK_CFG 0x0719 +#define BIT_CLKREQ_SW_EN BIT(4) + #define BIT_PCI_BCNQ_FLAG BIT(4) #define RTK_PCI_TXBD_DESA_BCNQ 0x308 #define RTK_PCI_TXBD_DESA_H2CQ 0x1320 @@ -200,6 +204,7 @@ struct rtw_pci { u16 rx_tag; struct rtw_pci_tx_ring tx_rings[RTK_MAX_TX_QUEUE_NUM]; struct rtw_pci_rx_ring rx_rings[RTK_MAX_RX_QUEUE_NUM]; + u16 link_ctrl; void __iomem *mmap; }; From patchwork Mon Nov 18 09:54:32 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tony Chuang X-Patchwork-Id: 11249237 X-Patchwork-Delegate: kvalo@adurom.com Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id F290D930 for ; Mon, 18 Nov 2019 09:54:46 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id DC0342071C for ; Mon, 18 Nov 2019 09:54:46 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726600AbfKRJyp (ORCPT ); Mon, 18 Nov 2019 04:54:45 -0500 Received: from rtits2.realtek.com ([211.75.126.72]:40041 "EHLO rtits2.realtek.com.tw" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726488AbfKRJyp (ORCPT ); Mon, 18 Nov 2019 04:54:45 -0500 Authenticated-By: X-SpamFilter-By: BOX Solutions SpamTrap 5.62 with qID xAI9sdC6018268, This message is accepted by code: ctloc85258 Received: from mail.realtek.com (RTITCASV01.realtek.com.tw[172.21.6.18]) by rtits2.realtek.com.tw (8.15.2/2.57/5.78) with ESMTPS id xAI9sdC6018268 (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Mon, 18 Nov 2019 17:54:39 +0800 Received: from localhost.localdomain (172.21.68.126) by RTITCASV01.realtek.com.tw (172.21.6.18) with Microsoft SMTP Server id 14.3.468.0; Mon, 18 Nov 2019 17:54:39 +0800 From: To: CC: , Subject: [PATCH v2 4/4] rtw88: allows to enable/disable HCI link PS mechanism Date: Mon, 18 Nov 2019 17:54:32 +0800 Message-ID: <20191118095432.4507-5-yhchuang@realtek.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20191118095432.4507-1-yhchuang@realtek.com> References: <20191118095432.4507-1-yhchuang@realtek.com> MIME-Version: 1.0 X-Originating-IP: [172.21.68.126] Sender: linux-wireless-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-wireless@vger.kernel.org From: Yan-Hsuan Chuang Different interfaces have its own link-related power save mechanism. Such as PCI can enter L1 state based on the traffic on the link, and sometimes driver needs to enable/disable it to avoid some issues, like throughput degrade when PCI trying to enter L1 state even if driver is having heavy traffic. For now, rtw88 only supports PCIE chips, and they just need to disable ASPM L1 when driver is not in power save mode, such as IPS and LPS. Signed-off-by: Yan-Hsuan Chuang --- v1 -> v2 * no change drivers/net/wireless/realtek/rtw88/hci.h | 6 ++++ drivers/net/wireless/realtek/rtw88/pci.c | 38 ++++++++++++++++++++++++ drivers/net/wireless/realtek/rtw88/pci.h | 1 + drivers/net/wireless/realtek/rtw88/ps.c | 6 ++++ 4 files changed, 51 insertions(+) diff --git a/drivers/net/wireless/realtek/rtw88/hci.h b/drivers/net/wireless/realtek/rtw88/hci.h index 4afbf0d163d1..3d91aea942c3 100644 --- a/drivers/net/wireless/realtek/rtw88/hci.h +++ b/drivers/net/wireless/realtek/rtw88/hci.h @@ -14,6 +14,7 @@ struct rtw_hci_ops { int (*start)(struct rtw_dev *rtwdev); void (*stop)(struct rtw_dev *rtwdev); void (*deep_ps)(struct rtw_dev *rtwdev, bool enter); + void (*link_ps)(struct rtw_dev *rtwdev, bool enter); int (*write_data_rsvd_page)(struct rtw_dev *rtwdev, u8 *buf, u32 size); int (*write_data_h2c)(struct rtw_dev *rtwdev, u8 *buf, u32 size); @@ -53,6 +54,11 @@ static inline void rtw_hci_deep_ps(struct rtw_dev *rtwdev, bool enter) rtwdev->hci.ops->deep_ps(rtwdev, enter); } +static inline void rtw_hci_link_ps(struct rtw_dev *rtwdev, bool enter) +{ + rtwdev->hci.ops->link_ps(rtwdev, enter); +} + static inline int rtw_hci_write_data_rsvd_page(struct rtw_dev *rtwdev, u8 *buf, u32 size) { diff --git a/drivers/net/wireless/realtek/rtw88/pci.c b/drivers/net/wireless/realtek/rtw88/pci.c index 6e99aad39487..a58e8276a41a 100644 --- a/drivers/net/wireless/realtek/rtw88/pci.c +++ b/drivers/net/wireless/realtek/rtw88/pci.c @@ -1150,6 +1150,43 @@ static void rtw_pci_clkreq_set(struct rtw_dev *rtwdev, bool enable) rtw_dbi_write8(rtwdev, RTK_PCIE_LINK_CFG, value); } +static void rtw_pci_aspm_set(struct rtw_dev *rtwdev, bool enable) +{ + u8 value; + int ret; + + ret = rtw_dbi_read8(rtwdev, RTK_PCIE_LINK_CFG, &value); + if (ret) { + rtw_err(rtwdev, "failed to read ASPM, ret=%d", ret); + return; + } + + if (enable) + value |= BIT_L1_SW_EN; + else + value &= ~BIT_L1_SW_EN; + + rtw_dbi_write8(rtwdev, RTK_PCIE_LINK_CFG, value); +} + +static void rtw_pci_link_ps(struct rtw_dev *rtwdev, bool enter) +{ + struct rtw_pci *rtwpci = (struct rtw_pci *)rtwdev->priv; + + /* Like CLKREQ, ASPM is also implemented by two HW modules, and can + * only be enabled when host supports it. + * + * And ASPM mechanism should be enabled when driver/firmware enters + * power save mode, without having heavy traffic. Because we've + * experienced some inter-operability issues that the link tends + * to enter L1 state on the fly even when driver is having high + * throughput. This is probably because the ASPM behavior slightly + * varies from different SOC. + */ + if (rtwpci->link_ctrl & PCI_EXP_LNKCTL_ASPM_L1) + rtw_pci_aspm_set(rtwdev, enter); +} + static void rtw_pci_link_cfg(struct rtw_dev *rtwdev) { struct rtw_pci *rtwpci = (struct rtw_pci *)rtwdev->priv; @@ -1292,6 +1329,7 @@ static struct rtw_hci_ops rtw_pci_ops = { .start = rtw_pci_start, .stop = rtw_pci_stop, .deep_ps = rtw_pci_deep_ps, + .link_ps = rtw_pci_link_ps, .read8 = rtw_pci_read8, .read16 = rtw_pci_read16, diff --git a/drivers/net/wireless/realtek/rtw88/pci.h b/drivers/net/wireless/realtek/rtw88/pci.h index 90efb73c607e..49bf29a92152 100644 --- a/drivers/net/wireless/realtek/rtw88/pci.h +++ b/drivers/net/wireless/realtek/rtw88/pci.h @@ -38,6 +38,7 @@ #define RTK_PCIE_LINK_CFG 0x0719 #define BIT_CLKREQ_SW_EN BIT(4) +#define BIT_L1_SW_EN BIT(3) #define BIT_PCI_BCNQ_FLAG BIT(4) #define RTK_PCI_TXBD_DESA_BCNQ 0x308 diff --git a/drivers/net/wireless/realtek/rtw88/ps.c b/drivers/net/wireless/realtek/rtw88/ps.c index 2226e3e7d7f8..913e6f47130f 100644 --- a/drivers/net/wireless/realtek/rtw88/ps.c +++ b/drivers/net/wireless/realtek/rtw88/ps.c @@ -31,6 +31,7 @@ int rtw_enter_ips(struct rtw_dev *rtwdev) rtw_coex_ips_notify(rtwdev, COEX_IPS_ENTER); rtw_core_stop(rtwdev); + rtw_hci_link_ps(rtwdev, true); return 0; } @@ -49,6 +50,8 @@ int rtw_leave_ips(struct rtw_dev *rtwdev) { int ret; + rtw_hci_link_ps(rtwdev, false); + ret = rtw_ips_pwr_up(rtwdev); if (ret) { rtw_err(rtwdev, "failed to leave ips state\n"); @@ -150,6 +153,7 @@ static void rtw_leave_lps_core(struct rtw_dev *rtwdev) conf->rlbm = 0; conf->smart_ps = 0; + rtw_hci_link_ps(rtwdev, false); rtw_fw_set_pwr_mode(rtwdev); rtw_fw_leave_lps_state_check(rtwdev); @@ -187,6 +191,8 @@ static void rtw_enter_lps_core(struct rtw_dev *rtwdev) rtw_coex_lps_notify(rtwdev, COEX_LPS_ENABLE); rtw_fw_set_pwr_mode(rtwdev); + rtw_hci_link_ps(rtwdev, true); + set_bit(RTW_FLAG_LEISURE_PS, rtwdev->flags); }