IMX6Q sabreSD PCIE + TW6865问题求助

devhuan · 发表于 2015-1-14 10:17:30

本帖最后由 devhuan 于 2015-1-14 10:33 编辑

大家好，我们在调试飞思卡尔的4in1编码器TW6865时遇到了难以解决的问题，希望大家给个意见和帮助；

在通过内核中System Type->Freescale MXC Implementations下的PCI Express选项

[*] PCI Express support
[ ] PCI Express EP mode in the IMX6 RC/EP interconnection system
[ ] PCI Express RC mode in the IMX6 RC/EP interconnection system
在内核中通过补丁pcie_cap_patch.patch 向kernel中添加TW6865相关的驱动和配置；并在内核的选项选上tw68系列的encoder的编码器
<*> Multimedia support --->
[*] Video capture adapters --->
Encoders, decoders, sensors and other helper chips --->
<*> pcie TW68 video encoder

复制代码

启动之后，我们的log中可以看到pcie可以link up，也就是说link上pcie；

PMU: registered new PMU device of type 0

Static Power Management for Freescale i.MX6

wait mode is enabled for i.MX6

cpaddr = c0880000 suspend_iram_base=c091c000

PM driver module loaded

iMX6 PCIe PCIe RC mode imx_pcie_pltfm_probe entering.

PCIE: imx_pcie_pltfm_probe start link up.

IMX PCIe port: link up.

PCI: bus0: Fast back to back transfers disabled

liutest start fixup imx6dq pcie tw6869

pci 0000:01:00.0: Setting PCI class for tw6868 PCIe device

liutest END fixup imx6dq pcie tw6869

PCI: bus1: Fast back to back transfers disabled

pci 0000:00:00.0: BAR 0: assigned [mem 0x01000000-0x010fffff 64bit pref]

pci 0000:00:00.0: BAR 0: set to [mem 0x01000000-0x010fffff 64bit pref] (PCI address [0x1000000-0x10fffff])

pci 0000:00:00.0: BAR 9: assigned [mem 0x01100000-0x011fffff pref]

pci 0000:00:00.0: BAR 6: assigned [mem 0x01200000-0x0120ffff pref]

pci 0000:01:00.0: BAR 0: assigned [mem 0x01100000-0x01100fff pref]

pci 0000:01:00.0: BAR 0: set to [mem 0x01100000-0x01100fff pref] (PCI address [0x1100000-0x1100fff])

pci 0000:00:00.0: PCI bridge to [bus 01-01]

pci 0000:00:00.0: bridge window [io disabled]

pci 0000:00:00.0: bridge window [mem disabled]

pci 0000:00:00.0: bridge window [mem 0x01100000-0x011fffff pref]

liutest PCI: r->start =0x1000000 r->end =0x10fffff

IMX usb wakeup probe

add wake up source irq 75

IMX usb wakeup probe

从这段log中可以看到pcie在探测插槽上的设备时是找到了设备的，并且成功连接，但是我们在这之后，CPU却在运行到如下log的地方停住了；

.......

ARC USBOTG Device Controller driver (1 August 2005)

mousedev: PS/2 mouse device common for all mice

i2c-core: driver [isl29023] using legacy suspend method

i2c-core: driver [isl29023] using legacy resume method

snvs_rtc snvs_rtc.0: rtc core: registered snvs_rtc as rtc0

i2c /dev entries driver

Linux video capture interface: v2.00

TW6868_: v4l2 driver version 2.0.1 loaded

PCI register init called //【死在了这里，没有继续往下启动了】

跟踪内核中的运行到函数__pci_enable_device_flags(struct pci_dev *dev,resource_size_t flags)时在里面调用函数pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);时没有返回一直停留在这个函数里面了，但是CPU没有”死掉“，就是停在里面了，因为在我注释掉这个函数之后，程序可以正确的跑到printk(KERN_INFO "pos 4 : after PCI register init called\n");函数;

File:pci.c

static int __pci_enable_device_flags(struct pci_dev *dev,resource_size_t flags)
{
int err;
int i, bars = 0;
/*
* Power state could be unknown at this point, either due to a fresh
* boot or a device removal call. So get the current power state
* so that things like MSI message writing will behave as expected
* (e.g. if the device really is in D0 at enable time).
*/
if (dev->pm_cap) {
u16 pmcsr;
// add by huan.gong @ 2015-01-13
printk(KERN_INFO "pos 3 : after PCI register init called\n");
pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
dev->current_state = (pmcsr & PCI_PM_CTRL_STATE_MASK);
}
// add by huan.gong @ 2015-01-13
printk(KERN_INFO "pos 4 : after PCI register init called\n");
if (atomic_add_return(1, &dev->enable_cnt) > 1)
{
// add by huan.gong @ 2015-01-13
printk(KERN_INFO "pos 5 : after PCI register init called\n");
return 0; /* already enabled */
}

复制代码

之后我又跟踪了函数pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);内联函数

static inline int pci_read_config_word(struct pci_dev *dev, int where, u16 *val)
{
return pci_bus_read_config_word(dev->bus, dev->devfn, where, val);
}
EXPORT_SYMBOL(pci_bus_read_config_word);

复制代码

为了提高性能这个函数最终的实现时宏

#define PCI_OP_READ(size,type,len) \
int pci_bus_read_config_##size \
(struct pci_bus *bus, unsigned int devfn, int pos, type *value) \
{ \
int res; \
unsigned long flags; \
u32 data = 0; \
if (PCI_##size##_BAD) return PCIBIOS_BAD_REGISTER_NUMBER; \
printk(KERN_INFO "::::lock irq\n");\
raw_spin_lock_irqsave(&pci_lock, flags); \
printk(KERN_INFO "::::read before\n");\
res = bus->ops->read(bus, devfn, pos, len, &data); \
printk(KERN_INFO "::::read after\n");\
*value = (type)data; \
raw_spin_unlock_irqrestore(&pci_lock, flags); \
printk(KERN_INFO "::::unlock irq\n");\
return res; \
}

复制代码

#define PCI_OP_READ(size,type,len) \

int pci_bus_read_config_##size \

(struct pci_bus *bus, unsigned int devfn, int pos, type *value) \

{ \

int res; \

unsigned long flags; \

u32 data = 0; \

if (PCI_##size##_BAD) return PCIBIOS_BAD_REGISTER_NUMBER; \

printk(KERN_INFO "::::lock irq\n");\

raw_spin_lock_irqsave(&pci_lock, flags); \

printk(KERN_INFO "::::read before\n");\

res = bus->ops->read(bus, devfn, pos, len, &data); \

printk(KERN_INFO "::::read after\n");\

*value = (type)data; \

raw_spin_unlock_irqrestore(&pci_lock, flags); \

printk(KERN_INFO "::::unlock irq\n");\

return res; \

}

最终，cpu是挂在了第39行的read函数里面，而这个是一个函数指针，是在驱动开发的过程中通过.read = XXXX()来指定的，如果结合我们前面分析的那部分应该是read(dev->bus, dev->devfn,dev->pm_cap + PCI_PM_CTRL,&pmcsr)其中dev就是pcie设备；也就是说在读取pcie的PCI_PM_CTRL控制字（16bit）时进去就没能出来，而原本读出来是要存入pmcsr的。

通过代码不难猜测出，是在去读取pcie电源相关配置字的时候除了问题。其实在这之前，我们可以是启动到内核加载DVFS module时挂掉的，后来卸载了patch之后，重新打上这个补丁，就变成这样了，而且用没有修改过的最原始的kernel来测试还是在PCI register init called..之后不动了.../欲哭无泪の

通过测试我们还看到了其他的log，在“挂起”之前，也有调用这个函数，只不过传进来的位置和要读取的长度不一样而已。

这可能是目前遇到的最难处理的问题，也不知道如何下手去处理，我们前期功能验证和测试的时候用的是成都嵌智捷科技的板子，后来我们的工程师开发的定制的板子，嵌智捷的板子上我们没有测试过PCIE，希望有相关经验的来帮忙分析一些，不胜感激.....提前来个万分感谢....

我知道答案目前已有6人回答

阅读全文