diff options
author | Yunhong Jiang <yunhong.jiang@intel.com> | 2015-08-04 12:17:53 -0700 |
---|---|---|
committer | Yunhong Jiang <yunhong.jiang@intel.com> | 2015-08-04 15:44:42 -0700 |
commit | 9ca8dbcc65cfc63d6f5ef3312a33184e1d726e00 (patch) | |
tree | 1c9cafbcd35f783a87880a10f85d1a060db1a563 /kernel/drivers/pci/pcie/aspm.c | |
parent | 98260f3884f4a202f9ca5eabed40b1354c489b29 (diff) |
Add the rt linux 4.1.3-rt3 as base
Import the rt linux 4.1.3-rt3 as OPNFV kvm base.
It's from git://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-4.1.y-rt and
the base is:
commit 0917f823c59692d751951bf5ea699a2d1e2f26a2
Author: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Date: Sat Jul 25 12:13:34 2015 +0200
Prepare v4.1.3-rt3
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
We lose all the git history this way and it's not good. We
should apply another opnfv project repo in future.
Change-Id: I87543d81c9df70d99c5001fbdf646b202c19f423
Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com>
Diffstat (limited to 'kernel/drivers/pci/pcie/aspm.c')
-rw-r--r-- | kernel/drivers/pci/pcie/aspm.c | 975 |
1 files changed, 975 insertions, 0 deletions
diff --git a/kernel/drivers/pci/pcie/aspm.c b/kernel/drivers/pci/pcie/aspm.c new file mode 100644 index 000000000..7d4fcdc51 --- /dev/null +++ b/kernel/drivers/pci/pcie/aspm.c @@ -0,0 +1,975 @@ +/* + * File: drivers/pci/pcie/aspm.c + * Enabling PCIe link L0s/L1 state and Clock Power Management + * + * Copyright (C) 2007 Intel + * Copyright (C) Zhang Yanmin (yanmin.zhang@intel.com) + * Copyright (C) Shaohua Li (shaohua.li@intel.com) + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/moduleparam.h> +#include <linux/pci.h> +#include <linux/pci_regs.h> +#include <linux/errno.h> +#include <linux/pm.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/jiffies.h> +#include <linux/delay.h> +#include <linux/pci-aspm.h> +#include "../pci.h" + +#ifdef MODULE_PARAM_PREFIX +#undef MODULE_PARAM_PREFIX +#endif +#define MODULE_PARAM_PREFIX "pcie_aspm." + +/* Note: those are not register definitions */ +#define ASPM_STATE_L0S_UP (1) /* Upstream direction L0s state */ +#define ASPM_STATE_L0S_DW (2) /* Downstream direction L0s state */ +#define ASPM_STATE_L1 (4) /* L1 state */ +#define ASPM_STATE_L0S (ASPM_STATE_L0S_UP | ASPM_STATE_L0S_DW) +#define ASPM_STATE_ALL (ASPM_STATE_L0S | ASPM_STATE_L1) + +struct aspm_latency { + u32 l0s; /* L0s latency (nsec) */ + u32 l1; /* L1 latency (nsec) */ +}; + +struct pcie_link_state { + struct pci_dev *pdev; /* Upstream component of the Link */ + struct pcie_link_state *root; /* pointer to the root port link */ + struct pcie_link_state *parent; /* pointer to the parent Link state */ + struct list_head sibling; /* node in link_list */ + struct list_head children; /* list of child link states */ + struct list_head link; /* node in parent's children list */ + + /* ASPM state */ + u32 aspm_support:3; /* Supported ASPM state */ + u32 aspm_enabled:3; /* Enabled ASPM state */ + u32 aspm_capable:3; /* Capable ASPM state with latency */ + u32 aspm_default:3; /* Default ASPM state by BIOS */ + u32 aspm_disable:3; /* Disabled ASPM state */ + + /* Clock PM state */ + u32 clkpm_capable:1; /* Clock PM capable? */ + u32 clkpm_enabled:1; /* Current Clock PM state */ + u32 clkpm_default:1; /* Default Clock PM state by BIOS */ + + /* Exit latencies */ + struct aspm_latency latency_up; /* Upstream direction exit latency */ + struct aspm_latency latency_dw; /* Downstream direction exit latency */ + /* + * Endpoint acceptable latencies. A pcie downstream port only + * has one slot under it, so at most there are 8 functions. + */ + struct aspm_latency acceptable[8]; +}; + +static int aspm_disabled, aspm_force; +static bool aspm_support_enabled = true; +static DEFINE_MUTEX(aspm_lock); +static LIST_HEAD(link_list); + +#define POLICY_DEFAULT 0 /* BIOS default setting */ +#define POLICY_PERFORMANCE 1 /* high performance */ +#define POLICY_POWERSAVE 2 /* high power saving */ + +#ifdef CONFIG_PCIEASPM_PERFORMANCE +static int aspm_policy = POLICY_PERFORMANCE; +#elif defined CONFIG_PCIEASPM_POWERSAVE +static int aspm_policy = POLICY_POWERSAVE; +#else +static int aspm_policy; +#endif + +static const char *policy_str[] = { + [POLICY_DEFAULT] = "default", + [POLICY_PERFORMANCE] = "performance", + [POLICY_POWERSAVE] = "powersave" +}; + +#define LINK_RETRAIN_TIMEOUT HZ + +static int policy_to_aspm_state(struct pcie_link_state *link) +{ + switch (aspm_policy) { + case POLICY_PERFORMANCE: + /* Disable ASPM and Clock PM */ + return 0; + case POLICY_POWERSAVE: + /* Enable ASPM L0s/L1 */ + return ASPM_STATE_ALL; + case POLICY_DEFAULT: + return link->aspm_default; + } + return 0; +} + +static int policy_to_clkpm_state(struct pcie_link_state *link) +{ + switch (aspm_policy) { + case POLICY_PERFORMANCE: + /* Disable ASPM and Clock PM */ + return 0; + case POLICY_POWERSAVE: + /* Disable Clock PM */ + return 1; + case POLICY_DEFAULT: + return link->clkpm_default; + } + return 0; +} + +static void pcie_set_clkpm_nocheck(struct pcie_link_state *link, int enable) +{ + struct pci_dev *child; + struct pci_bus *linkbus = link->pdev->subordinate; + + list_for_each_entry(child, &linkbus->devices, bus_list) { + if (enable) + pcie_capability_set_word(child, PCI_EXP_LNKCTL, + PCI_EXP_LNKCTL_CLKREQ_EN); + else + pcie_capability_clear_word(child, PCI_EXP_LNKCTL, + PCI_EXP_LNKCTL_CLKREQ_EN); + } + link->clkpm_enabled = !!enable; +} + +static void pcie_set_clkpm(struct pcie_link_state *link, int enable) +{ + /* Don't enable Clock PM if the link is not Clock PM capable */ + if (!link->clkpm_capable && enable) + enable = 0; + /* Need nothing if the specified equals to current state */ + if (link->clkpm_enabled == enable) + return; + pcie_set_clkpm_nocheck(link, enable); +} + +static void pcie_clkpm_cap_init(struct pcie_link_state *link, int blacklist) +{ + int capable = 1, enabled = 1; + u32 reg32; + u16 reg16; + struct pci_dev *child; + struct pci_bus *linkbus = link->pdev->subordinate; + + /* All functions should have the same cap and state, take the worst */ + list_for_each_entry(child, &linkbus->devices, bus_list) { + pcie_capability_read_dword(child, PCI_EXP_LNKCAP, ®32); + if (!(reg32 & PCI_EXP_LNKCAP_CLKPM)) { + capable = 0; + enabled = 0; + break; + } + pcie_capability_read_word(child, PCI_EXP_LNKCTL, ®16); + if (!(reg16 & PCI_EXP_LNKCTL_CLKREQ_EN)) + enabled = 0; + } + link->clkpm_enabled = enabled; + link->clkpm_default = enabled; + link->clkpm_capable = (blacklist) ? 0 : capable; +} + +/* + * pcie_aspm_configure_common_clock: check if the 2 ends of a link + * could use common clock. If they are, configure them to use the + * common clock. That will reduce the ASPM state exit latency. + */ +static void pcie_aspm_configure_common_clock(struct pcie_link_state *link) +{ + int same_clock = 1; + u16 reg16, parent_reg, child_reg[8]; + unsigned long start_jiffies; + struct pci_dev *child, *parent = link->pdev; + struct pci_bus *linkbus = parent->subordinate; + /* + * All functions of a slot should have the same Slot Clock + * Configuration, so just check one function + */ + child = list_entry(linkbus->devices.next, struct pci_dev, bus_list); + BUG_ON(!pci_is_pcie(child)); + + /* Check downstream component if bit Slot Clock Configuration is 1 */ + pcie_capability_read_word(child, PCI_EXP_LNKSTA, ®16); + if (!(reg16 & PCI_EXP_LNKSTA_SLC)) + same_clock = 0; + + /* Check upstream component if bit Slot Clock Configuration is 1 */ + pcie_capability_read_word(parent, PCI_EXP_LNKSTA, ®16); + if (!(reg16 & PCI_EXP_LNKSTA_SLC)) + same_clock = 0; + + /* Configure downstream component, all functions */ + list_for_each_entry(child, &linkbus->devices, bus_list) { + pcie_capability_read_word(child, PCI_EXP_LNKCTL, ®16); + child_reg[PCI_FUNC(child->devfn)] = reg16; + if (same_clock) + reg16 |= PCI_EXP_LNKCTL_CCC; + else + reg16 &= ~PCI_EXP_LNKCTL_CCC; + pcie_capability_write_word(child, PCI_EXP_LNKCTL, reg16); + } + + /* Configure upstream component */ + pcie_capability_read_word(parent, PCI_EXP_LNKCTL, ®16); + parent_reg = reg16; + if (same_clock) + reg16 |= PCI_EXP_LNKCTL_CCC; + else + reg16 &= ~PCI_EXP_LNKCTL_CCC; + pcie_capability_write_word(parent, PCI_EXP_LNKCTL, reg16); + + /* Retrain link */ + reg16 |= PCI_EXP_LNKCTL_RL; + pcie_capability_write_word(parent, PCI_EXP_LNKCTL, reg16); + + /* Wait for link training end. Break out after waiting for timeout */ + start_jiffies = jiffies; + for (;;) { + pcie_capability_read_word(parent, PCI_EXP_LNKSTA, ®16); + if (!(reg16 & PCI_EXP_LNKSTA_LT)) + break; + if (time_after(jiffies, start_jiffies + LINK_RETRAIN_TIMEOUT)) + break; + msleep(1); + } + if (!(reg16 & PCI_EXP_LNKSTA_LT)) + return; + + /* Training failed. Restore common clock configurations */ + dev_err(&parent->dev, "ASPM: Could not configure common clock\n"); + list_for_each_entry(child, &linkbus->devices, bus_list) + pcie_capability_write_word(child, PCI_EXP_LNKCTL, + child_reg[PCI_FUNC(child->devfn)]); + pcie_capability_write_word(parent, PCI_EXP_LNKCTL, parent_reg); +} + +/* Convert L0s latency encoding to ns */ +static u32 calc_l0s_latency(u32 encoding) +{ + if (encoding == 0x7) + return (5 * 1000); /* > 4us */ + return (64 << encoding); +} + +/* Convert L0s acceptable latency encoding to ns */ +static u32 calc_l0s_acceptable(u32 encoding) +{ + if (encoding == 0x7) + return -1U; + return (64 << encoding); +} + +/* Convert L1 latency encoding to ns */ +static u32 calc_l1_latency(u32 encoding) +{ + if (encoding == 0x7) + return (65 * 1000); /* > 64us */ + return (1000 << encoding); +} + +/* Convert L1 acceptable latency encoding to ns */ +static u32 calc_l1_acceptable(u32 encoding) +{ + if (encoding == 0x7) + return -1U; + return (1000 << encoding); +} + +struct aspm_register_info { + u32 support:2; + u32 enabled:2; + u32 latency_encoding_l0s; + u32 latency_encoding_l1; +}; + +static void pcie_get_aspm_reg(struct pci_dev *pdev, + struct aspm_register_info *info) +{ + u16 reg16; + u32 reg32; + + pcie_capability_read_dword(pdev, PCI_EXP_LNKCAP, ®32); + info->support = (reg32 & PCI_EXP_LNKCAP_ASPMS) >> 10; + info->latency_encoding_l0s = (reg32 & PCI_EXP_LNKCAP_L0SEL) >> 12; + info->latency_encoding_l1 = (reg32 & PCI_EXP_LNKCAP_L1EL) >> 15; + pcie_capability_read_word(pdev, PCI_EXP_LNKCTL, ®16); + info->enabled = reg16 & PCI_EXP_LNKCTL_ASPMC; +} + +static void pcie_aspm_check_latency(struct pci_dev *endpoint) +{ + u32 latency, l1_switch_latency = 0; + struct aspm_latency *acceptable; + struct pcie_link_state *link; + + /* Device not in D0 doesn't need latency check */ + if ((endpoint->current_state != PCI_D0) && + (endpoint->current_state != PCI_UNKNOWN)) + return; + + link = endpoint->bus->self->link_state; + acceptable = &link->acceptable[PCI_FUNC(endpoint->devfn)]; + + while (link) { + /* Check upstream direction L0s latency */ + if ((link->aspm_capable & ASPM_STATE_L0S_UP) && + (link->latency_up.l0s > acceptable->l0s)) + link->aspm_capable &= ~ASPM_STATE_L0S_UP; + + /* Check downstream direction L0s latency */ + if ((link->aspm_capable & ASPM_STATE_L0S_DW) && + (link->latency_dw.l0s > acceptable->l0s)) + link->aspm_capable &= ~ASPM_STATE_L0S_DW; + /* + * Check L1 latency. + * Every switch on the path to root complex need 1 + * more microsecond for L1. Spec doesn't mention L0s. + */ + latency = max_t(u32, link->latency_up.l1, link->latency_dw.l1); + if ((link->aspm_capable & ASPM_STATE_L1) && + (latency + l1_switch_latency > acceptable->l1)) + link->aspm_capable &= ~ASPM_STATE_L1; + l1_switch_latency += 1000; + + link = link->parent; + } +} + +static void pcie_aspm_cap_init(struct pcie_link_state *link, int blacklist) +{ + struct pci_dev *child, *parent = link->pdev; + struct pci_bus *linkbus = parent->subordinate; + struct aspm_register_info upreg, dwreg; + + if (blacklist) { + /* Set enabled/disable so that we will disable ASPM later */ + link->aspm_enabled = ASPM_STATE_ALL; + link->aspm_disable = ASPM_STATE_ALL; + return; + } + + /* Configure common clock before checking latencies */ + pcie_aspm_configure_common_clock(link); + + /* Get upstream/downstream components' register state */ + pcie_get_aspm_reg(parent, &upreg); + child = list_entry(linkbus->devices.next, struct pci_dev, bus_list); + pcie_get_aspm_reg(child, &dwreg); + + /* + * Setup L0s state + * + * Note that we must not enable L0s in either direction on a + * given link unless components on both sides of the link each + * support L0s. + */ + if (dwreg.support & upreg.support & PCIE_LINK_STATE_L0S) + link->aspm_support |= ASPM_STATE_L0S; + if (dwreg.enabled & PCIE_LINK_STATE_L0S) + link->aspm_enabled |= ASPM_STATE_L0S_UP; + if (upreg.enabled & PCIE_LINK_STATE_L0S) + link->aspm_enabled |= ASPM_STATE_L0S_DW; + link->latency_up.l0s = calc_l0s_latency(upreg.latency_encoding_l0s); + link->latency_dw.l0s = calc_l0s_latency(dwreg.latency_encoding_l0s); + + /* Setup L1 state */ + if (upreg.support & dwreg.support & PCIE_LINK_STATE_L1) + link->aspm_support |= ASPM_STATE_L1; + if (upreg.enabled & dwreg.enabled & PCIE_LINK_STATE_L1) + link->aspm_enabled |= ASPM_STATE_L1; + link->latency_up.l1 = calc_l1_latency(upreg.latency_encoding_l1); + link->latency_dw.l1 = calc_l1_latency(dwreg.latency_encoding_l1); + + /* Save default state */ + link->aspm_default = link->aspm_enabled; + + /* Setup initial capable state. Will be updated later */ + link->aspm_capable = link->aspm_support; + /* + * If the downstream component has pci bridge function, don't + * do ASPM for now. + */ + list_for_each_entry(child, &linkbus->devices, bus_list) { + if (pci_pcie_type(child) == PCI_EXP_TYPE_PCI_BRIDGE) { + link->aspm_disable = ASPM_STATE_ALL; + break; + } + } + + /* Get and check endpoint acceptable latencies */ + list_for_each_entry(child, &linkbus->devices, bus_list) { + u32 reg32, encoding; + struct aspm_latency *acceptable = + &link->acceptable[PCI_FUNC(child->devfn)]; + + if (pci_pcie_type(child) != PCI_EXP_TYPE_ENDPOINT && + pci_pcie_type(child) != PCI_EXP_TYPE_LEG_END) + continue; + + pcie_capability_read_dword(child, PCI_EXP_DEVCAP, ®32); + /* Calculate endpoint L0s acceptable latency */ + encoding = (reg32 & PCI_EXP_DEVCAP_L0S) >> 6; + acceptable->l0s = calc_l0s_acceptable(encoding); + /* Calculate endpoint L1 acceptable latency */ + encoding = (reg32 & PCI_EXP_DEVCAP_L1) >> 9; + acceptable->l1 = calc_l1_acceptable(encoding); + + pcie_aspm_check_latency(child); + } +} + +static void pcie_config_aspm_dev(struct pci_dev *pdev, u32 val) +{ + pcie_capability_clear_and_set_word(pdev, PCI_EXP_LNKCTL, + PCI_EXP_LNKCTL_ASPMC, val); +} + +static void pcie_config_aspm_link(struct pcie_link_state *link, u32 state) +{ + u32 upstream = 0, dwstream = 0; + struct pci_dev *child, *parent = link->pdev; + struct pci_bus *linkbus = parent->subordinate; + + /* Nothing to do if the link is already in the requested state */ + state &= (link->aspm_capable & ~link->aspm_disable); + if (link->aspm_enabled == state) + return; + /* Convert ASPM state to upstream/downstream ASPM register state */ + if (state & ASPM_STATE_L0S_UP) + dwstream |= PCI_EXP_LNKCTL_ASPM_L0S; + if (state & ASPM_STATE_L0S_DW) + upstream |= PCI_EXP_LNKCTL_ASPM_L0S; + if (state & ASPM_STATE_L1) { + upstream |= PCI_EXP_LNKCTL_ASPM_L1; + dwstream |= PCI_EXP_LNKCTL_ASPM_L1; + } + /* + * Spec 2.0 suggests all functions should be configured the + * same setting for ASPM. Enabling ASPM L1 should be done in + * upstream component first and then downstream, and vice + * versa for disabling ASPM L1. Spec doesn't mention L0S. + */ + if (state & ASPM_STATE_L1) + pcie_config_aspm_dev(parent, upstream); + list_for_each_entry(child, &linkbus->devices, bus_list) + pcie_config_aspm_dev(child, dwstream); + if (!(state & ASPM_STATE_L1)) + pcie_config_aspm_dev(parent, upstream); + + link->aspm_enabled = state; +} + +static void pcie_config_aspm_path(struct pcie_link_state *link) +{ + while (link) { + pcie_config_aspm_link(link, policy_to_aspm_state(link)); + link = link->parent; + } +} + +static void free_link_state(struct pcie_link_state *link) +{ + link->pdev->link_state = NULL; + kfree(link); +} + +static int pcie_aspm_sanity_check(struct pci_dev *pdev) +{ + struct pci_dev *child; + u32 reg32; + + /* + * Some functions in a slot might not all be PCIe functions, + * very strange. Disable ASPM for the whole slot + */ + list_for_each_entry(child, &pdev->subordinate->devices, bus_list) { + if (!pci_is_pcie(child)) + return -EINVAL; + + /* + * If ASPM is disabled then we're not going to change + * the BIOS state. It's safe to continue even if it's a + * pre-1.1 device + */ + + if (aspm_disabled) + continue; + + /* + * Disable ASPM for pre-1.1 PCIe device, we follow MS to use + * RBER bit to determine if a function is 1.1 version device + */ + pcie_capability_read_dword(child, PCI_EXP_DEVCAP, ®32); + if (!(reg32 & PCI_EXP_DEVCAP_RBER) && !aspm_force) { + dev_info(&child->dev, "disabling ASPM on pre-1.1 PCIe device. You can enable it with 'pcie_aspm=force'\n"); + return -EINVAL; + } + } + return 0; +} + +static struct pcie_link_state *alloc_pcie_link_state(struct pci_dev *pdev) +{ + struct pcie_link_state *link; + + link = kzalloc(sizeof(*link), GFP_KERNEL); + if (!link) + return NULL; + INIT_LIST_HEAD(&link->sibling); + INIT_LIST_HEAD(&link->children); + INIT_LIST_HEAD(&link->link); + link->pdev = pdev; + if (pci_pcie_type(pdev) == PCI_EXP_TYPE_DOWNSTREAM) { + struct pcie_link_state *parent; + parent = pdev->bus->parent->self->link_state; + if (!parent) { + kfree(link); + return NULL; + } + link->parent = parent; + list_add(&link->link, &parent->children); + } + /* Setup a pointer to the root port link */ + if (!link->parent) + link->root = link; + else + link->root = link->parent->root; + + list_add(&link->sibling, &link_list); + pdev->link_state = link; + return link; +} + +/* + * pcie_aspm_init_link_state: Initiate PCI express link state. + * It is called after the pcie and its children devices are scanned. + * @pdev: the root port or switch downstream port + */ +void pcie_aspm_init_link_state(struct pci_dev *pdev) +{ + struct pcie_link_state *link; + int blacklist = !!pcie_aspm_sanity_check(pdev); + + if (!aspm_support_enabled) + return; + + if (!pci_is_pcie(pdev) || pdev->link_state) + return; + if (pci_pcie_type(pdev) != PCI_EXP_TYPE_ROOT_PORT && + pci_pcie_type(pdev) != PCI_EXP_TYPE_DOWNSTREAM) + return; + + /* VIA has a strange chipset, root port is under a bridge */ + if (pci_pcie_type(pdev) == PCI_EXP_TYPE_ROOT_PORT && + pdev->bus->self) + return; + + down_read(&pci_bus_sem); + if (list_empty(&pdev->subordinate->devices)) + goto out; + + mutex_lock(&aspm_lock); + link = alloc_pcie_link_state(pdev); + if (!link) + goto unlock; + /* + * Setup initial ASPM state. Note that we need to configure + * upstream links also because capable state of them can be + * update through pcie_aspm_cap_init(). + */ + pcie_aspm_cap_init(link, blacklist); + + /* Setup initial Clock PM state */ + pcie_clkpm_cap_init(link, blacklist); + + /* + * At this stage drivers haven't had an opportunity to change the + * link policy setting. Enabling ASPM on broken hardware can cripple + * it even before the driver has had a chance to disable ASPM, so + * default to a safe level right now. If we're enabling ASPM beyond + * the BIOS's expectation, we'll do so once pci_enable_device() is + * called. + */ + if (aspm_policy != POLICY_POWERSAVE) { + pcie_config_aspm_path(link); + pcie_set_clkpm(link, policy_to_clkpm_state(link)); + } + +unlock: + mutex_unlock(&aspm_lock); +out: + up_read(&pci_bus_sem); +} + +/* Recheck latencies and update aspm_capable for links under the root */ +static void pcie_update_aspm_capable(struct pcie_link_state *root) +{ + struct pcie_link_state *link; + BUG_ON(root->parent); + list_for_each_entry(link, &link_list, sibling) { + if (link->root != root) + continue; + link->aspm_capable = link->aspm_support; + } + list_for_each_entry(link, &link_list, sibling) { + struct pci_dev *child; + struct pci_bus *linkbus = link->pdev->subordinate; + if (link->root != root) + continue; + list_for_each_entry(child, &linkbus->devices, bus_list) { + if ((pci_pcie_type(child) != PCI_EXP_TYPE_ENDPOINT) && + (pci_pcie_type(child) != PCI_EXP_TYPE_LEG_END)) + continue; + pcie_aspm_check_latency(child); + } + } +} + +/* @pdev: the endpoint device */ +void pcie_aspm_exit_link_state(struct pci_dev *pdev) +{ + struct pci_dev *parent = pdev->bus->self; + struct pcie_link_state *link, *root, *parent_link; + + if (!parent || !parent->link_state) + return; + + down_read(&pci_bus_sem); + mutex_lock(&aspm_lock); + /* + * All PCIe functions are in one slot, remove one function will remove + * the whole slot, so just wait until we are the last function left. + */ + if (!list_is_last(&pdev->bus_list, &parent->subordinate->devices)) + goto out; + + link = parent->link_state; + root = link->root; + parent_link = link->parent; + + /* All functions are removed, so just disable ASPM for the link */ + pcie_config_aspm_link(link, 0); + list_del(&link->sibling); + list_del(&link->link); + /* Clock PM is for endpoint device */ + free_link_state(link); + + /* Recheck latencies and configure upstream links */ + if (parent_link) { + pcie_update_aspm_capable(root); + pcie_config_aspm_path(parent_link); + } +out: + mutex_unlock(&aspm_lock); + up_read(&pci_bus_sem); +} + +/* @pdev: the root port or switch downstream port */ +void pcie_aspm_pm_state_change(struct pci_dev *pdev) +{ + struct pcie_link_state *link = pdev->link_state; + + if (aspm_disabled || !pci_is_pcie(pdev) || !link) + return; + if ((pci_pcie_type(pdev) != PCI_EXP_TYPE_ROOT_PORT) && + (pci_pcie_type(pdev) != PCI_EXP_TYPE_DOWNSTREAM)) + return; + /* + * Devices changed PM state, we should recheck if latency + * meets all functions' requirement + */ + down_read(&pci_bus_sem); + mutex_lock(&aspm_lock); + pcie_update_aspm_capable(link->root); + pcie_config_aspm_path(link); + mutex_unlock(&aspm_lock); + up_read(&pci_bus_sem); +} + +void pcie_aspm_powersave_config_link(struct pci_dev *pdev) +{ + struct pcie_link_state *link = pdev->link_state; + + if (aspm_disabled || !pci_is_pcie(pdev) || !link) + return; + + if (aspm_policy != POLICY_POWERSAVE) + return; + + if ((pci_pcie_type(pdev) != PCI_EXP_TYPE_ROOT_PORT) && + (pci_pcie_type(pdev) != PCI_EXP_TYPE_DOWNSTREAM)) + return; + + down_read(&pci_bus_sem); + mutex_lock(&aspm_lock); + pcie_config_aspm_path(link); + pcie_set_clkpm(link, policy_to_clkpm_state(link)); + mutex_unlock(&aspm_lock); + up_read(&pci_bus_sem); +} + +static void __pci_disable_link_state(struct pci_dev *pdev, int state, bool sem, + bool force) +{ + struct pci_dev *parent = pdev->bus->self; + struct pcie_link_state *link; + + if (!pci_is_pcie(pdev)) + return; + + if (pci_pcie_type(pdev) == PCI_EXP_TYPE_ROOT_PORT || + pci_pcie_type(pdev) == PCI_EXP_TYPE_DOWNSTREAM) + parent = pdev; + if (!parent || !parent->link_state) + return; + + /* + * A driver requested that ASPM be disabled on this device, but + * if we don't have permission to manage ASPM (e.g., on ACPI + * systems we have to observe the FADT ACPI_FADT_NO_ASPM bit and + * the _OSC method), we can't honor that request. Windows has + * a similar mechanism using "PciASPMOptOut", which is also + * ignored in this situation. + */ + if (aspm_disabled && !force) { + dev_warn(&pdev->dev, "can't disable ASPM; OS doesn't have ASPM control\n"); + return; + } + + if (sem) + down_read(&pci_bus_sem); + mutex_lock(&aspm_lock); + link = parent->link_state; + if (state & PCIE_LINK_STATE_L0S) + link->aspm_disable |= ASPM_STATE_L0S; + if (state & PCIE_LINK_STATE_L1) + link->aspm_disable |= ASPM_STATE_L1; + pcie_config_aspm_link(link, policy_to_aspm_state(link)); + + if (state & PCIE_LINK_STATE_CLKPM) { + link->clkpm_capable = 0; + pcie_set_clkpm(link, 0); + } + mutex_unlock(&aspm_lock); + if (sem) + up_read(&pci_bus_sem); +} + +void pci_disable_link_state_locked(struct pci_dev *pdev, int state) +{ + __pci_disable_link_state(pdev, state, false, false); +} +EXPORT_SYMBOL(pci_disable_link_state_locked); + +/** + * pci_disable_link_state - Disable device's link state, so the link will + * never enter specific states. Note that if the BIOS didn't grant ASPM + * control to the OS, this does nothing because we can't touch the LNKCTL + * register. + * + * @pdev: PCI device + * @state: ASPM link state to disable + */ +void pci_disable_link_state(struct pci_dev *pdev, int state) +{ + __pci_disable_link_state(pdev, state, true, false); +} +EXPORT_SYMBOL(pci_disable_link_state); + +static int pcie_aspm_set_policy(const char *val, struct kernel_param *kp) +{ + int i; + struct pcie_link_state *link; + + if (aspm_disabled) + return -EPERM; + for (i = 0; i < ARRAY_SIZE(policy_str); i++) + if (!strncmp(val, policy_str[i], strlen(policy_str[i]))) + break; + if (i >= ARRAY_SIZE(policy_str)) + return -EINVAL; + if (i == aspm_policy) + return 0; + + down_read(&pci_bus_sem); + mutex_lock(&aspm_lock); + aspm_policy = i; + list_for_each_entry(link, &link_list, sibling) { + pcie_config_aspm_link(link, policy_to_aspm_state(link)); + pcie_set_clkpm(link, policy_to_clkpm_state(link)); + } + mutex_unlock(&aspm_lock); + up_read(&pci_bus_sem); + return 0; +} + +static int pcie_aspm_get_policy(char *buffer, struct kernel_param *kp) +{ + int i, cnt = 0; + for (i = 0; i < ARRAY_SIZE(policy_str); i++) + if (i == aspm_policy) + cnt += sprintf(buffer + cnt, "[%s] ", policy_str[i]); + else + cnt += sprintf(buffer + cnt, "%s ", policy_str[i]); + return cnt; +} + +module_param_call(policy, pcie_aspm_set_policy, pcie_aspm_get_policy, + NULL, 0644); + +#ifdef CONFIG_PCIEASPM_DEBUG +static ssize_t link_state_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct pci_dev *pci_device = to_pci_dev(dev); + struct pcie_link_state *link_state = pci_device->link_state; + + return sprintf(buf, "%d\n", link_state->aspm_enabled); +} + +static ssize_t link_state_store(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t n) +{ + struct pci_dev *pdev = to_pci_dev(dev); + struct pcie_link_state *link, *root = pdev->link_state->root; + u32 val, state = 0; + + if (kstrtouint(buf, 10, &val)) + return -EINVAL; + + if (aspm_disabled) + return -EPERM; + if (n < 1 || val > 3) + return -EINVAL; + + /* Convert requested state to ASPM state */ + if (val & PCIE_LINK_STATE_L0S) + state |= ASPM_STATE_L0S; + if (val & PCIE_LINK_STATE_L1) + state |= ASPM_STATE_L1; + + down_read(&pci_bus_sem); + mutex_lock(&aspm_lock); + list_for_each_entry(link, &link_list, sibling) { + if (link->root != root) + continue; + pcie_config_aspm_link(link, state); + } + mutex_unlock(&aspm_lock); + up_read(&pci_bus_sem); + return n; +} + +static ssize_t clk_ctl_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct pci_dev *pci_device = to_pci_dev(dev); + struct pcie_link_state *link_state = pci_device->link_state; + + return sprintf(buf, "%d\n", link_state->clkpm_enabled); +} + +static ssize_t clk_ctl_store(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t n) +{ + struct pci_dev *pdev = to_pci_dev(dev); + bool state; + + if (strtobool(buf, &state)) + return -EINVAL; + + down_read(&pci_bus_sem); + mutex_lock(&aspm_lock); + pcie_set_clkpm_nocheck(pdev->link_state, state); + mutex_unlock(&aspm_lock); + up_read(&pci_bus_sem); + + return n; +} + +static DEVICE_ATTR(link_state, 0644, link_state_show, link_state_store); +static DEVICE_ATTR(clk_ctl, 0644, clk_ctl_show, clk_ctl_store); + +static char power_group[] = "power"; +void pcie_aspm_create_sysfs_dev_files(struct pci_dev *pdev) +{ + struct pcie_link_state *link_state = pdev->link_state; + + if (!pci_is_pcie(pdev) || + (pci_pcie_type(pdev) != PCI_EXP_TYPE_ROOT_PORT && + pci_pcie_type(pdev) != PCI_EXP_TYPE_DOWNSTREAM) || !link_state) + return; + + if (link_state->aspm_support) + sysfs_add_file_to_group(&pdev->dev.kobj, + &dev_attr_link_state.attr, power_group); + if (link_state->clkpm_capable) + sysfs_add_file_to_group(&pdev->dev.kobj, + &dev_attr_clk_ctl.attr, power_group); +} + +void pcie_aspm_remove_sysfs_dev_files(struct pci_dev *pdev) +{ + struct pcie_link_state *link_state = pdev->link_state; + + if (!pci_is_pcie(pdev) || + (pci_pcie_type(pdev) != PCI_EXP_TYPE_ROOT_PORT && + pci_pcie_type(pdev) != PCI_EXP_TYPE_DOWNSTREAM) || !link_state) + return; + + if (link_state->aspm_support) + sysfs_remove_file_from_group(&pdev->dev.kobj, + &dev_attr_link_state.attr, power_group); + if (link_state->clkpm_capable) + sysfs_remove_file_from_group(&pdev->dev.kobj, + &dev_attr_clk_ctl.attr, power_group); +} +#endif + +static int __init pcie_aspm_disable(char *str) +{ + if (!strcmp(str, "off")) { + aspm_policy = POLICY_DEFAULT; + aspm_disabled = 1; + aspm_support_enabled = false; + printk(KERN_INFO "PCIe ASPM is disabled\n"); + } else if (!strcmp(str, "force")) { + aspm_force = 1; + printk(KERN_INFO "PCIe ASPM is forcibly enabled\n"); + } + return 1; +} + +__setup("pcie_aspm=", pcie_aspm_disable); + +void pcie_no_aspm(void) +{ + /* + * Disabling ASPM is intended to prevent the kernel from modifying + * existing hardware state, not to clear existing state. To that end: + * (a) set policy to POLICY_DEFAULT in order to avoid changing state + * (b) prevent userspace from changing policy + */ + if (!aspm_force) { + aspm_policy = POLICY_DEFAULT; + aspm_disabled = 1; + } +} + +bool pcie_aspm_support_enabled(void) +{ + return aspm_support_enabled; +} +EXPORT_SYMBOL(pcie_aspm_support_enabled); |