内核中断笔记(MSIx和GSI)。
struct device
{
struct irq_domain *msi_domain; // 设备关联的是在---->irq-gic-v3-its-pci-msi.c中创建，
}
struct irq_domain
{
void *host_data;
struct irq_domain_ops *ops;
}
irq_domain的创建最终通过：
__irq_domain_add函数：
struct irq_domain *__irq_domain_add(struct fwnode_handle *fwnode, int size,
irq_hw_number_t hwirq_max, int direct_max,
const struct irq_domain_ops *ops, //提供ops
void *host_data) //额外的host_data。有不同的意义。层次结构：具体是irq_domain_alloc_irq_data函数建立：
系统创建的irq_domain：总共4个。通过parent成员组成 层次结构。根为ffff8020c0030400。
[ 0.000000] ===__irq_domain_add domain ffff8020c0030400 name irqchip@(____ptrval____) === irq-gic-v3.c
[ 0.000000] ===__irq_domain_add domain ffff8020c0030500 name irqchip@(____ptrval____) === irq-gic-v3-its.c
[ 0.002937] ===__irq_domain_add domain ffff8020c0032100 name irqchip@(____ptrval____) === irq-gic-v3-its-pci-msi
具体创建过程：
第一个：boot核init_IRQ 函数最终调用到irq_gic_v3.c中
host_data: 是struct gic_chip_data gic_data
irq_domain_ops:gic_irq_domain_ops
参考：irq_gic_v3.c ：gic_init_bases函数
[ 0.000000] ===__irq_domain_add domain ffff8020c0030400 name irqchip@(____ptrr
val____) ===
[ 0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.19.0 #100
[ 0.000000] Call trace:
[ 0.000000] dump_backtrace+0x0/0x180
[ 0.000000] show_stack+0x24/0x30
[ 0.000000] dump_stack+0x90/0xb4
[ 0.000000] __irq_domain_add+0x1e4/0x398
[ 0.000000] gic_init_bases+0xd4/0x2dc
[ 0.000000] gic_acpi_init+0x158/0x270
[ 0.000000] acpi_match_madt+0x4c/0x84
[ 0.000000] acpi_table_parse_entries_array+0x140/0x218
[ 0.000000] acpi_table_parse_entries+0x70/0x98
[ 0.000000] acpi_table_parse_madt+0x40/0x50
[ 0.000000] __acpi_probe_device_table+0x88/0xe0
[ 0.000000] irqchip_init+0x38/0x40
[ 0.000000] init_IRQ+0xfc/0x130
[ 0.000000] start_kernel+0x330/0x4b8
irq_domain_update_bus_token(gic_data.domain, DOMAIN_BUS_WIRED);第二个 irq_gic_v3调用了irq_gic_v3_its中的函数创建，作为parent传递。
host_data是msi_domain_info（ops:its_msi_domain_ops）
irq_domain_ops是its_domain_ops .
参考：its_probe_one->its_init_domain函数
[ 0.000000] ===__irq_domain_add domain ffff8020c0030500 name irqchip@(____ptrr
val____) ===
[ 0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.19.0 #100
[ 0.000000] Call trace:
[ 0.000000] dump_backtrace+0x0/0x180
[ 0.000000] show_stack+0x24/0x30
[ 0.000000] dump_stack+0x90/0xb4
[ 0.000000] __irq_domain_add+0x1e4/0x398
[ 0.000000] its_probe_one+0x8cc/0x9bc
[ 0.000000] gic_acpi_parse_madt_its+0x104/0x144
[ 0.000000] acpi_table_parse_entries_array+0x140/0x218
[ 0.000000] acpi_table_parse_entries+0x70/0x98
[ 0.000000] acpi_table_parse_madt+0x40/0x50
[ 0.000000] its_init+0x1bc/0x490
[ 0.000000] gic_init_bases+0x1d4/0x2dc
[ 0.000000] gic_acpi_init+0x158/0x270
[ 0.000000] acpi_match_madt+0x4c/0x84
[ 0.000000] acpi_table_parse_entries_array+0x140/0x218
[ 0.000000] acpi_table_parse_entries+0x70/0x98
[ 0.000000] acpi_table_parse_madt+0x40/0x50
[ 0.000000] __acpi_probe_device_table+0x88/0xe0
[ 0.000000] irqchip_init+0x38/0x40
irq_domain_update_bus_token(inner_domain, DOMAIN_BUS_NEXUS);第三个 最后的do_one_initcall 调用its_pci_msi_init中函数创建，在msi_create_irq_domain函数下创建。
host_data: msi_domain_info （ ops:its_pci_msi_ops）
irq_domain_ops是 kernel/irq/msi.c提供的msi_domain_ops。
irq_domain_ops在kernel/irq/msi.c和drivers/pci/msi.c也有，在不同的每个地方都提供了部分回调。
参考：its_pci_msi_init_one->pci_msi_create_irq_domain->msi_create_irq_domain .....
[ 0.002937] ===__irq_domain_add domain ffff8020c0032100 name irqchip@(____ptrr
val____) ===
[ 0.002939] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.19.0 #100
[ 0.002941] Call trace:
[ 0.002944] dump_backtrace+0x0/0x180
[ 0.002946] show_stack+0x24/0x30
[ 0.002948] dump_stack+0x90/0xb4
[ 0.002950] __irq_domain_add+0x1e4/0x398
[ 0.002953] irq_domain_create_hierarchy+0x88/0x90
[ 0.002955] msi_create_irq_domain+0x50/0x160
[ 0.002959] pci_msi_create_irq_domain+0x80/0x160
[ 0.002961] its_pci_msi_init_one+0x94/0xd8
[ 0.002963] its_pci_msi_parse_madt+0x60/0x90
[ 0.002965] acpi_table_parse_entries_array+0x140/0x218
[ 0.002967] acpi_table_parse_entries+0x70/0x98
[ 0.002969] acpi_table_parse_madt+0x40/0x50
[ 0.002971] its_pci_msi_init+0xb8/0xcc
[ 0.002973] do_one_initcall+0x54/0x1d8
[ 0.002976] kernel_init_freeable+0x160/0x328
[ 0.002978] kernel_init+0x18/0x110
[ 0.002980] ret_from_fork+0x10/0x18
irq_domain_update_bus_token(domain, DOMAIN_BUS_PCI_MSI);int __irq_domain_alloc_irqs(struct irq_domain *domain, int irq_base,
unsigned int nr_irqs, int node, void *arg,
bool realloc, const struct cpumask *affinity)
{
int i, ret, virq;
if (domain == NULL) {
domain = irq_default_domain;
if (WARN(!domain, "domain is NULL; cannot allocate IRQ\n"))
return -EINVAL;
}
if (!domain->ops->alloc) {
pr_debug("domain->ops->alloc() is NULL\n");
return -ENOSYS;
}
if (realloc && irq_base >= 0) {
virq = irq_base;
} else {
virq = irq_domain_alloc_descs(irq_base, nr_irqs, 0, node,
affinity); //创建irq_desc
if (virq < 0) {
pr_debug("cannot allocate IRQ(base %d, count %d)\n",
irq_base, nr_irqs);
return virq;
}
}
if (irq_domain_alloc_irq_data(domain, virq, nr_irqs)) { //irq_data 也会创建层次，更加parent。
pr_debug("cannot allocate memory for IRQ%d\n", virq);
ret = -ENOMEM;
goto out_free_desc;
}
mutex_lock(&irq_domain_mutex);
ret = irq_domain_alloc_irqs_hierarchy(domain, virq, nr_irqs, arg); //层次调用，下面日志内容，通过irq_domain_set_hwirq_and_chip 打印。
if (ret < 0) {
mutex_unlock(&irq_domain_mutex);
goto out_free_irq_data;
}
for (i = 0; i < nr_irqs; i++)
irq_domain_insert_irq(virq + i);
mutex_unlock(&irq_domain_mutex);
return virq;
out_free_irq_data:
irq_domain_free_irq_data(virq, nr_irqs);
out_free_desc:
irq_free_descs(virq, nr_irqs);
return ret;
}
irq_domain_set_hwirq_and_chip 函数kernel/irq/irqdomain.c
从父到子创建。
pci设备的MSI中断hwirq和virq的创建：
drivers/pci/msi.c函数 pci_msi_setup_msi_irqs中通过dev_get_msi_domain获取的是irq-gic-v3-its-pci-msi.c下创建的domain。然后根据其中的parent，调用irq_domain_alloc_irqs_parent从父节点的domain创建hwirq和virq的映射。
这个是gic-irq-v3.c中的：
[ 13.130169] ===irq_domain_set_hwirq_and_chip domain ffff8020c0030400 irq_dataa
ffff8020d6c8f600 virq 32 hwirq 8197 ===
[ 13.140852] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.19.0 #100
[ 13.146929] Hardware name: Phytium FT2004/FT2004, BIOS V1006 Apr 7 2020
[ 13.153613] Call trace:
[ 13.156047] dump_backtrace+0x0/0x180
[ 13.159695] show_stack+0x24/0x30
[ 13.162996] dump_stack+0x90/0xb4
[ 13.166297] irq_domain_set_hwirq_and_chip+0x78/0xd8
[ 13.171246] irq_domain_set_info+0x58/0x90
[ 13.175327] gic_irq_domain_alloc+0xdc/0x268
[ 13.179582] irq_domain_alloc_irqs_parent+0x48/0x60
[ 13.184445] its_irq_gic_domain_alloc+0xd4/0xf8
[ 13.188960] its_irq_domain_alloc+0xd0/0x168
[ 13.193215] irq_domain_alloc_irqs_parent+0x48/0x60
[ 13.198078] msi_domain_alloc+0x7c/0x160
[ 13.201985] __irq_domain_alloc_irqs+0x150/0x338
[ 13.206587] msi_domain_alloc_irqs+0xa8/0x308
[ 13.210930] pci_msi_setup_msi_irqs+0x64/0x78
[ 13.215271] __pci_enable_msix+0x320/0x540
[ 13.219353] pci_alloc_irq_vectors_affinity+0x120/0x160
下面是gic-irq-v3-its.c中的alloc函数
[ 13.447145] ===irq_domain_set_hwirq_and_chip domain ffff8020c0030500 irq_dataa
ffff8020d6c8cb00 virq 32 hwirq 8197 ===
[ 13.457825] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.19.0 #100
[ 13.463902] Hardware name: Phytium FT2004/FT2004, BIOS V1006 Apr 7 2020
[ 13.470586] Call trace:
[ 13.473019] dump_backtrace+0x0/0x180
[ 13.476667] show_stack+0x24/0x30
[ 13.479967] dump_stack+0x90/0xb4
[ 13.483267] irq_domain_set_hwirq_and_chip+0x78/0xd8
[ 13.488216] its_irq_domain_alloc+0x58/0x168
[ 13.492471] irq_domain_alloc_irqs_parent+0x48/0x60
[ 13.497334] msi_domain_alloc+0x7c/0x160
[ 13.501242] __irq_domain_alloc_irqs+0x150/0x338
[ 13.505844] msi_domain_alloc_irqs+0xa8/0x308
[ 13.510186] pci_msi_setup_msi_irqs+0x64/0x78
[ 13.514527] __pci_enable_msix+0x320/0x540
[ 13.518609] pci_alloc_irq_vectors_affinity+0x120/0x160
最终这个是irq-gic-v3-its-pci-msi.c中的接口调用：
这个3145732 hwirq在函数drivers/pci/msi.c：pci_msi_domain_calc_hwirq中计算。__irq_domain_alloc_irqs下通过ops->set_desc。调到pci_msi_domain_set_desc。到pci_msi_domain_calc_hwirq
[ 13.595814] ===irq_domain_set_hwirq_and_chip domain ffff8020c0032100 irq_dataa
ffff8020c1b70c28 virq 32 hwirq 3145732 ===
[ 13.606754] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.19.0 #100
[ 13.612830] Hardware name: Phytium FT2004/FT2004, BIOS V1006 Apr 7 2020
[ 13.619514] Call trace:
[ 13.621947] dump_backtrace+0x0/0x180
[ 13.625595] show_stack+0x24/0x30
[ 13.628896] dump_stack+0x90/0xb4
[ 13.632196] irq_domain_set_hwirq_and_chip+0x78/0xd8
[ 13.637145] msi_domain_ops_init+0x3c/0x80
[ 13.641227] msi_domain_alloc+0xb8/0x160
[ 13.645135] __irq_domain_alloc_irqs+0x150/0x338
[ 13.649737] msi_domain_alloc_irqs+0xa8/0x308
[ 13.654079] pci_msi_setup_msi_irqs+0x64/0x78
[ 13.658421] __pci_enable_msix+0x320/0x540
[ 13.662502] pci_alloc_irq_vectors_affinity+0x120/0x160
pci驱动初始化irq中断：pci_alloc_irq_vectors_affinity 跟踪：
pci_alloc_irq_vectors_affinity (drivers/pci/msi.c)
__pci_enable_msi_range
msi_capability_init 暂分析msi部分
pci_msi_setup_msi_irqs (drivers/pci/msi.c)
msi_domain_alloc_irqs (kernel/irq/msi.c)
ops->domain_alloc_irqs (kernel/irq/msi.c) //msi_domain_ops->__msi_domain_alloc_irqs
函数：
__msi_domain_alloc_irqs (kernel/irq/msi.c)
msi_domain_prepare_irqs (一个pci/x 设备分配一个its_devie ：调用its_create_device 分配its_device)
for_each_msi_entry循环：
ops->set_desc (kernel/irq/msi.c) （pci_msi_domain_calc_hwirq）计算出的全局唯一ID ，作为最上面的irq_data的hwirq
__irq_domain_alloc_irqs //kernel/irq/irqdomain.c
irq_domain_alloc_descs (kernel/irq/irqdomain.c) ---分配多个irq_desc
irq_domain_alloc_irq_data (kernel/irq/irqdomain.c) ---建立irq_doamin。irq_data的层次结构
irq_domain_alloc_irqs_hierarchy (kernel/irq/irqdomain.c)
irq_set_msi_desc_off //设置msi_entry->irq 。该msi_desc 的irq_base
irq_domain_insert_irq
irq_domain_activate_irq //激活msi/x中断 ,往相关的寄存器写入gic的地址。
函数：
irq_domain_alloc_irqs_hierarchy
domain->ops->alloc(domain, irq_base, nr_irqs, arg) //kernel/irq/msi.c ：irq_domain_ops->alloc（msi_domain_alloc）
msi_domain_alloc
irq_domain_alloc_irqs_parent
irq_domain_alloc_irqs_hierarchy(domain->parent,irq_base,nr_irqsa,arg) ......层级调用 irq-gic-v3-its
domain->ops->alloc //drivers/irqchip/irq-gic-v3-its.c 提供 its_irq_domain_alloc
its_irq_domain_alloc //drivers/irqchip/irq-gic-v3-its.c
its_irq_gic_domain_alloc
irq_domain_alloc_irqs_parent //kernel/irq/irqdomain.c
irq_domain_alloc_irqs_hierarchy(domain->parent, irq_base, nr_irqs, arg); ----------------层级调用到irq-gic-v3
domain->ops->alloc //drivers/irqchip/irq-gic-v3.c 提供gic_irq_domain_alloc
gic_irq_domain_alloc //drivers/irqchip/irq-gic-v3.c
gic_irq_domain_map
irq_domain_set_info //kernel/irq/irqdomain.c
irq_domain_set_hwirq_and_chip 设置最底层的hwirq和irq,irq_chip
irq_domain_set_hwirq_and_chip //设置irq_data 的hwirq和irq ,irq_chip //drivers/irqchip/irq-gic-v3-its.c
msi_domain_ops->msi_init kernel/irq/msi.c 提供msi_domain_ops_init
irq_domain_set_hwirq_and_chip 设置最上层的hwirq和irq ,irq_chip
dmesg日志：
不同irq_domain_ops->alloc：先后关系
[ 11.265652] ===gic_irq_domain_alloc domain ffff8022c0038f00 virq 47 hwirq 8205 nr_irq 1 //首先是最底层的irq-gic-v3.c
[ 11.273733] ===its_irq_domain_alloc domain ffff8022c0039000 virq 47 hwirq 8205 nr_irqs 1 //其次是中级的irq-gic-v3-its.c
[ 11.281899] ===msi_domain_alloc doamin ffff8022c003ad00 virq 47 hwirq 7864320 nr_irqs 1 //最后是irq-gic-v3-its-pci-msi.c 使用 kernel/irq/msi.c 的alloc
可以看出：
每个层级的irq_data的virq是一致的，是同一个虚拟中断号。而hwirq 除了最上层的是pci_msi_domain_calc_hwirq 计算出的全局唯一ID，其它是irq-gic-v3-its.c 中根据gic驱动得出
设置irq_data 中 的hwirq和irq ,irq_chip
irq_domain_set_hwirq_and_chip:
其中irq_chip代表的中断芯片的配置，分别定义了三个
drivers/irqchip/irq-gic-v3.c： gic_chip /"GICv3"
drivers/irqchip/irq-gic-v3-its.c: its_irq_chip /"ITS"
drivers/irqchip/irq-gic-v3-its-pci-msi.c: its_msi_irq_chip /"ITS-MSI"激活msi/msix中断：
irq_domain_activate_irq
__irq_domain_activate_irq
domain->ops->activate //kernel/irq/msi.c 中irq_domain_ops->activate (msi_domain_activate)
msi_domain_activate
irq_chip_write_msi_msg
static int msi_domain_activate(struct irq_domain *domain,
struct irq_data *irq_data, bool early)
{
struct msi_msg msg[2] = { [1] = { }, };
BUG_ON(irq_chip_compose_msi_msg(irq_data, msg));
msi_check_level(irq_data->domain, msg);
irq_chip_write_msi_msg(irq_data, msg);
return 0;
}
static inline void irq_chip_write_msi_msg(struct irq_data *data,
struct msi_msg *msg)
{
data->chip->irq_write_msi_msg(data, msg);
}
__pci_write_msi_msg 函数是具体的写操作、激活msi/msix中断：
irq_domain_activate_irq
__irq_domain_activate_irq
domain->ops->activate //kernel/irq/msi.c 中irq_domain_ops->activate (msi_domain_activate)
msi_domain_activate //kernel/irq/msi.c
irq_chip_compose_msi_msg
irq_chip_write_msi_msg //写msi/x 相关寄存器。写入gic地址。
irq_data->chip->irq_write_msi_msg(data, msg); //此时irq_domain是its_pci_msi创建的，irq_data的 chip数据是 irq-gic-v3-its-pci-msi.c --- irq_chip "ITS-MSI"
激活需要写入msi/x消息：
static void its_irq_compose_msi_msg(struct irq_data *d, struct msi_msg *msg)
{
struct its_device *its_dev = irq_data_get_irq_chip_data(d);
struct its_node *its;
u64 addr;
its = its_dev->its;
addr = its->get_msi_base(its_dev); //调用its_irq_get_msi_base函数
msg->address_lo = lower_32_bits(addr); //gic——its的地址
msg->address_hi = upper_32_bits(addr);
msg->data = its_get_event_id(d); //data
printk("===its_irq_compose_msi_msg its_device %llx address %llx_%llx data %u\n",msg->address_hi,msg->address_lo,msg->data);
iommu_dma_map_msi_msg(d->irq, msg);
}
//data的值实际上就是hwirq和当前设备的起始中断号的差。 如果一个设备只有一个，那么这个差值就是0。
static inline u32 its_get_event_id(struct irq_data *d)
{
struct its_device *its_dev = irq_data_get_irq_chip_data(d);
return d->hwirq - its_dev->event_map.lpi_base;
}
//获取gic——ITS的地址
static u64 its_irq_get_msi_base(struct its_device *its_dev)
{
struct its_node *its = its_dev->its;
return its->phys_base + GITS_TRANSLATER;
}
its_irq_compose_msi_msg：合成msi_msg。
PS:关于补丁中修改的its驱动“msi中断控制器XXXXXXXXXX”的理解：msi_base 没有转换成iova，直接使用。正常使会通过smmu进行转换的。。。
its_irq_compose_msi_msg 中注释了iommu_dma_compose_msi_msg这句话调用。
原函数的意思是需要准备msi_msg。其中有msi_base的地址（its_base+GIC_TRANSLATOR）
static void its_irq_compose_msi_msg(struct irq_data *d, struct msi_msg *msg)
{
struct its_device *its_dev = irq_data_get_irq_chip_data(d);
struct its_node *its;
u64 addr;
its = its_dev->its;
addr = its->get_msi_base(its_dev);
msg->address_lo = lower_32_bits(addr);
msg->address_hi = upper_32_bits(addr);
msg->data = its_get_event_id(d);
iommu_dma_compose_msi_msg(irq_data_get_msi_desc(d), msg); //将msi_base的地址通过 iova地址。
}
void iommu_dma_compose_msi_msg(struct msi_desc *desc,
struct msi_msg *msg)
{
struct device *dev = msi_desc_to_dev(desc);
const struct iommu_domain *domain = iommu_get_domain_for_dev(dev);
const struct iommu_dma_msi_page *msi_page;
msi_page = msi_desc_get_iommu_cookie(desc);
if (!domain || !domain->iova_cookie || WARN_ON(!msi_page))
return;
msg->address_hi = upper_32_bits(msi_page->iova); // 地址是msi_page的iova 这个是什么是否分配的 ？？ 是在its的alloc回调its_irq_domain_alloc->iommu_dma_prepare_msi
msg->address_lo &= cookie_msi_granule(domain->iova_cookie) - 1;
msg->address_lo += lower_32_bits(msi_page->iova);
}
iommu_dma_prepare_msi函数将 分配iommu_dma_msi_page结构体，记录msi_addr和iova。
int iommu_dma_prepare_msi(struct msi_desc *desc, phys_addr_t msi_addr)
{
struct device *dev = msi_desc_to_dev(desc);
struct iommu_domain *domain = iommu_get_domain_for_dev(dev);
struct iommu_dma_msi_page *msi_page;
static DEFINE_MUTEX(msi_prepare_lock); /* see below */
if (!domain || !domain->iova_cookie) {
desc->iommu_cookie = NULL;
return 0;
}
/*
* In fact the whole prepare operation should already be serialised by
* irq_domain_mutex further up the callchain, but that's pretty subtle
* on its own, so consider this locking as failsafe documentation...
*/
mutex_lock(&msi_prepare_lock);
msi_page = iommu_dma_get_msi_page(dev, msi_addr, domain);
mutex_unlock(&msi_prepare_lock);
msi_desc_set_iommu_cookie(desc, msi_page); //将msi_page设置到irq_desc->iommu_cookie
if (!msi_page)
return -ENOMEM;
return 0;
}
分配的8192开始的号，，怎么开始？its_alloc_device_irq 确定msi的硬件中断号
its_device->event_map.lpi_base->lpi_base号怎么来的 ？
分配的8192开始的号，，怎么开始？its_alloc_device_irq 确定msi的硬件中断号
its_device->event_map.lpi_base->lpi_base号怎么来的 ？
gic_its初始化
its_init
its_alloc_lpi_tables
its_lpi_init
free_lpi_range(8192, lpis);
mk_lpi_range //创建lpi_range
list_add(&new->entry, &lpi_range_list);
设备初始化irq
its_create_device
its_lpi_alloc
alloc_lpi_range //为每个its_device 分配dev->event_map.lpi_base。每次在前面已经分配的后面。初始化msi中断号：一个its硬件
static int __init its_alloc_lpi_tables(void)
{
phys_addr_t paddr;
lpi_id_bits = min_t(u32, GICD_TYPER_ID_BITS(gic_rdists->gicd_typer),
ITS_MAX_LPI_NRBITS);
gic_rdists->prop_page = its_allocate_prop_table(GFP_NOWAIT);
if (!gic_rdists->prop_page) {
pr_err("Failed to allocate PROPBASE\n");
return -ENOMEM;
}
paddr = page_to_phys(gic_rdists->prop_page);
pr_info("===GIC: using LPI property table @%pa lpi_id_bits %u \n", &paddr,lpi_id_bits); //这里实测16 。 16个bit位？？？？
return its_lpi_init(lpi_id_bits);
}
static int __init its_lpi_init(u32 id_bits)
{
u32 lpis = (1UL << id_bits) - 8192; //左移16位为65536 。意思是总共65536 - 8192个中断？
u32 numlpis;
int err;
numlpis = 1UL << GICD_TYPER_NUM_LPIS(gic_rdists->gicd_typer);
if (numlpis > 2 && !WARN_ON(numlpis > lpis)) {
lpis = numlpis;
pr_info("ITS: Using hypervisor restricted LPI range [%u]\n",
lpis);
}
/*
* Initializing the allocator is just the same as freeing the
* full range of LPIs.
*/
err = free_lpi_range(8192, lpis); //这里会初始化新的lpi_range 。并加入到lpi_range_list，从8192开始计算。总共65536 - 8192 个
pr_info("===ITS: Allocator initialized for %u LPIs\n", lpis);
return err;
}
static int free_lpi_range(u32 base, u32 nr_lpis)
{
struct lpi_range *new;
int err = 0;
mutex_lock(&lpi_range_lock);
new = mk_lpi_range(base, nr_lpis);
if (!new) {
err = -ENOMEM;
goto out;
}
list_add(&new->entry, &lpi_range_list);
list_sort(NULL, &lpi_range_list, lpi_range_cmp);
merge_lpi_ranges();
out:
mutex_unlock(&lpi_range_lock);
return err;
}
分配中断号：
static int alloc_lpi_range(u32 nr_lpis, u32 *base)
{
struct lpi_range *range, *tmp;
int err = -ENOSPC;
mutex_lock(&lpi_range_lock);
list_for_each_entry_safe(range, tmp, &lpi_range_list, entry) {
if (range->span >= nr_lpis) {
*base = range->base_id;
range->base_id += nr_lpis;
range->span -= nr_lpis;
if (range->span == 0) {
list_del(&range->entry);
kfree(range);
}
err = 0;
break;
}
}
mutex_unlock(&lpi_range_lock);
pr_info("==ITS: alloc %u:%u\n", *base, nr_lpis);
dump_stack();
return err;
}
dmesg日志：
[ 6.293629] ==ITS: alloc 8192:1
[ 6.615759] ==ITS: alloc 8193:1
[ 6.905502] ==ITS: alloc 8194:1
[ 7.280835] ==ITS: alloc 8195:1
[ 7.582356] ==ITS: alloc 8196:1
[ 7.902356] ==ITS: alloc 8197:1
[ 8.198276] ==ITS: alloc 8198:1
[ 8.489553] ==ITS: alloc 8199:1
[ 8.787372] ==ITS: alloc 8200:1
[ 11.019360] ==ITS: alloc 8201:4
[ 12.709670] ==ITS: alloc 8205:1
[ 16.354248] ==ITS: alloc 8206:1
pci设备注册irq相关的handle：
前面初始化的时会返回多个连续virq。
static bool pci_endpoint_test_request_irq(struct pci_endpoint_test *test)
{
int i;
int err;
struct pci_dev *pdev = test->pdev;
struct device *dev = &pdev->dev;
for (i = 0; i < test->num_irqs; i++) {
err = devm_request_irq(dev, pci_irq_vector(pdev, i), //第i个virq
pci_endpoint_test_irqhandler,
IRQF_SHARED, DRV_MODULE_NAME, test);
if (err)
goto fail;
}
return true;
}
获取第nr个virq
int pci_irq_vector(struct pci_dev *dev, unsigned int nr)
{
if (dev->msix_enabled) { //msix 有多个，一个irq对应一个msi_desc。
struct msi_desc *entry;
int i = 0;
for_each_pci_msi_entry(entry, dev) {
if (i == nr)
return entry->irq;
i++;
}
WARN_ON_ONCE(1);
return -EINVAL;
}
if (dev->msi_enabled) { //msi只有一个msi_desc 。里面记录了irq_base和nvec_used
struct msi_desc *entry = first_pci_msi_entry(dev);
if (WARN_ON_ONCE(nr >= entry->nvec_used))
return -EINVAL;
} else {
if (WARN_ON_ONCE(nr > 0))
return -EINVAL;
}
return dev->irq + nr; //从base 开始增加 nr 个。
}pci_irq_vector 返回第n个irq号。用来申请irq_request。
1：根据irq找到irq_data
根据irq找到irq_desc->irq_data
两种方式：
irq_get_irq_data: 获取virq关联的第一层irq_data。
irq_domain_get_irq_data：根据irq和domain 找到irq_data。 首先获取virq关联的第一层irq_data ，然后遍历irq_data的parent，比较doamin是否相对。
中断触发后调用gic_handle_irq函数：drivers/irqchip/irq-gic-v3.c
static asmlinkage void __exception_irq_entry gic_handle_irq(struct pt_regs *regs)
{
u32 irqnr;
do {
irqnr = gic_read_iar();
if (likely(irqnr > 15 && irqnr < 1020) || irqnr >= 8192) {
int err;
if (static_branch_likely(&supports_deactivate_key))
gic_write_eoir(irqnr);
else
isb();
err = handle_domain_irq(gic_data.domain, irqnr, regs); //使用的是gic_data.domain。也就是irq_gic_v3.c创建的doamin。 其会根据hwirq找到virq。并执行handle
if (err) {
WARN_ONCE(true, "Unexpected interrupt received!\n");
if (static_branch_likely(&supports_deactivate_key)) {
if (irqnr < 8192)
gic_write_dir(irqnr);
} else {
gic_write_eoir(irqnr);
}
}
continue;
}
。。。。
}/GSIGSI ：Global System Interrupt，是ACPI spec规定的全局中断表。. 它为多IOAPIC情况下确定了系统唯一的一个中断号。. 例如IOAPIC1有24个IRQ，IOAPIC2也有24个IRQ，则IOAPIC2 的GSI是从24开始，GSI = 24 + IRQ（IOAPIC2）。. SCI ：System Control Interrupt，系统控制中断，是ACPI定义的，专用于ACPI电源管理的一个IRQ。. 它在Intel平台上常常与南桥的电源管理模块一起，当外部EC等发生Event后会引发SCI。. Windows的SCI ISR程序就是著名的acpi.sys。.
非PCI的 中断hwirq和virq
[ 0.000000] ===irq_domain_set_hwirq_and_chip domain ffff8020c0030400 irq_dataa
ffff8020c0042828 virq 3 hwirq 27 ===
[ 0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.19.0 #100
[ 0.000000] Call trace:
[ 0.000000] dump_backtrace+0x0/0x180
[ 0.000000] show_stack+0x24/0x30
[ 0.000000] dump_stack+0x90/0xb4
[ 0.000000] irq_domain_set_hwirq_and_chip+0x78/0xd8
[ 0.000000] irq_domain_set_info+0x58/0x90
[ 0.000000] gic_irq_domain_alloc+0x170/0x268
[ 0.000000] __irq_domain_alloc_irqs+0x150/0x338
[ 0.000000] irq_create_fwspec_mapping+0x118/0x318
[ 0.000000] acpi_register_gsi+0x6c/0xa8
[ 0.000000] map_gt_gsi+0x30/0x3c
[ 0.000000] acpi_gtdt_map_ppi+0x4c/0x78
[ 0.000000] arch_timer_acpi_init+0xa4/0x280
[ 0.000000] acpi_table_parse+0xa4/0xdc
=================================================================
[ 0.000000] ===irq_domain_set_hwirq_and_chip domain ffff8020c0030400 irq_data ffff8020c0042428 virq 1 hwirq 25 ===
[ 0.000000] ===irq_domain_set_hwirq_and_chip domain ffff8020c0030400 irq_data ffff8020c0042628 virq 2 hwirq 30 ===
[ 0.000000] ===irq_domain_set_hwirq_and_chip domain ffff8020c0030400 irq_data ffff8020c0042828 virq 3 hwirq 27 ===
.........
irq_create_fwspec_mapping :
unsigned int irq_create_fwspec_mapping(struct irq_fwspec *fwspec)
{
struct irq_domain *domain;
struct irq_data *irq_data;
irq_hw_number_t hwirq;
unsigned int type = IRQ_TYPE_NONE;
int virq;
if (fwspec->fwnode) {
domain = irq_find_matching_fwspec(fwspec, DOMAIN_BUS_WIRED); //irq-gic-v3.c irq_domain
if (!domain)
domain = irq_find_matching_fwspec(fwspec, DOMAIN_BUS_ANY);
} else {
domain = irq_default_domain;
}
if (!domain) {
pr_warn("no irq domain found for %s !\n",
of_node_full_name(to_of_node(fwspec->fwnode)));
return 0;
}
if (irq_domain_translate(domain, fwspec, &hwirq, &type))
return 0;
if (irq_domain_is_hierarchy(domain)) {
virq = irq_domain_alloc_irqs(domain, 1, NUMA_NO_NODE, fwspec); //__irq_domain_alloc_irqs->分配virq。调用irq_domain_alloc_irqs_hierarchy:调用gic_irq_v3下的irq_domain->ops->alloc .....
if (virq <= 0)
return 0;
} else {
/* Create mapping */
virq = irq_create_mapping(domain, hwirq);
if (!virq)
return virq;
}
}
static int irq_domain_translate(struct irq_domain *d,
struct irq_fwspec *fwspec,
irq_hw_number_t *hwirq, unsigned int *type)
{
#ifdef CONFIG_IRQ_DOMAIN_HIERARCHY
if (d->ops->translate)
return d->ops->translate(d, fwspec, hwirq, type); //gic_irq_domain_translate
#endif
if (d->ops->xlate)
return d->ops->xlate(d, to_of_node(fwspec->fwnode),
fwspec->param, fwspec->param_count,
hwirq, type);
/* If domain has no translation, then we assume interrupt line */
*hwirq = fwspec->param[0];
return 0;
}
static int gic_irq_domain_translate(struct irq_domain *d,
struct irq_fwspec *fwspec,
unsigned long *hwirq,
unsigned int *type)
{
。。。
if (is_fwnode_irqchip(fwspec->fwnode)) {
if(fwspec->param_count != 2)
return -EINVAL;
*hwirq = fwspec->param[0]; //fwspec中记录了hwirq
*type = fwspec->param[1];
WARN_ON(*type == IRQ_TYPE_NONE);
return 0;
}
。。。。
}
irq_create_fwspec_mapping后续调用：
irq_create_fwspec_mapping
irq_domain_alloc_irqs
__irq_domain_alloc_irqs //后续流程参考msi
比如acpi创建platform过程中
[ 5.094866] irq_create_fwspec_mapping+0x60/0x310
[ 5.099555] acpi_register_gsi+0x64/0xa4
[ 5.103464] acpi_dev_resource_interrupt+0x170/0x1c0
[ 5.108414] acpi_dev_process_resource+0xe0/0x160
[ 5.113104] acpi_walk_resource_buffer+0x68/0xc8
[ 5.117706] acpi_walk_resources+0xb0/0xf8
[ 5.121788] acpi_dev_get_resources+0xcc/0x130
[ 5.126217] acpi_create_platform_device+0xac/0x2b0
[ 5.131080] acpi_bus_attach+0x274/0x284
[ 5.134989] acpi_bus_attach+0x80/0x284
[ 5.138810] acpi_bus_attach+0x80/0x284
[ 5.142632] acpi_bus_scan+0x50/0xb0
[ 5.146193] acpi_scan_init+0x120/0x284
[ 5.150015] acpi_init+0x2c8/0x33c
[ 5.153402] do_one_initcall+0x50/0x2f0
[ 5.157224] kernel_init_freeable+0x340/0x3cc
[ 5.161567] kernel_init+0x20/0x140
[ 5.165042] ret_from_fork+0x10/0x18platform_get_irq 函数：
drivers/base/platform.c
int platform_get_irq(struct platform_device *dev, unsigned int num)
{
#ifdef CONFIG_SPARC
/* sparc does not have irqs represented as IORESOURCE_IRQ resources */
if (!dev || num >= dev->archdata.num_irqs)
return -ENXIO;
return dev->archdata.irqs[num];
#else
struct resource *r;
if (IS_ENABLED(CONFIG_OF_IRQ) && dev->dev.of_node) {
int ret;
ret = of_irq_get(dev->dev.of_node, num);
if (ret > 0 || ret == -EPROBE_DEFER)
return ret;
}
r = platform_get_resource(dev, IORESOURCE_IRQ, num); //中断是一种资源IORESOURCE_IRQ，参考acpi创建platform_device的文章
if (has_acpi_companion(&dev->dev)) {
if (r && r->flags & IORESOURCE_DISABLED) {
int ret;
ret = acpi_irq_get(ACPI_HANDLE(&dev->dev), num, r);
if (ret)
return ret;
}
}
/*
* The resources may pass trigger flags to the irqs that need
* to be set up. It so happens that the trigger flags for
* IORESOURCE_BITS correspond 1-to-1 to the IRQF_TRIGGER*
* settings.
*/
if (r && r->flags & IORESOURCE_BITS) {
struct irq_data *irqd;
irqd = irq_get_irq_data(r->start);
if (!irqd)
return -ENXIO;
irqd_set_trigger_type(irqd, r->flags & IORESOURCE_BITS);
}
return r ? r->start : -ENXIO;
#endif
}