summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--Documentation/zh_CN/io_ordering.txt67
1 files changed, 67 insertions, 0 deletions
diff --git a/Documentation/zh_CN/io_ordering.txt b/Documentation/zh_CN/io_ordering.txt
new file mode 100644
index 000000000000..e592daf4e014
--- /dev/null
+++ b/Documentation/zh_CN/io_ordering.txt
@@ -0,0 +1,67 @@
+Chinese translated version of Documentation/io_orderings.txt
+
+If you have any comment or update to the content, please contact the
+original document maintainer directly. However, if you have a problem
+communicating in English you can also ask the Chinese maintainer for
+help. Contact the Chinese maintainer if this translation is outdated
+or if there is a problem with the translation.
+
+Chinese maintainer: Lin Yongting <linyongting@gmail.com>
+---------------------------------------------------------------------
+Documentation/io_ordering.txt 的中文翻译
+
+如果想评论或更新本文的内容,请直接联系原文档的维护者。如果你使用英文
+交流有困难的话,也可以向中文版维护者求助。如果本翻译更新不及时或者翻
+译存在问题,请联系中文版维护者。
+
+中文版维护者: 林永听 Lin Yongting <linyongting@gmail.com>
+中文版翻译者: 林永听 Lin Yongting <linyongting@gmail.com>
+中文版校译者: 林永听 Lin Yongting <linyongting@gmail.com>
+
+
+以下为正文
+---------------------------------------------------------------------
+
+在某些平台上,所谓的内存映射I/O是弱顺序。在这些平台上,驱动开发者有责任
+保证I/O内存映射地址的写操作按程序图意的顺序达到设备。通常读取一个“安全”
+设备寄存器或桥寄存器,触发IO芯片清刷未处理的写操作到达设备后才处理读操作,
+而达到保证目的。驱动程序通常在spinlock保护的临界区退出之前使用这种技术。
+这也可以保证后面的写操作只在前面的写操作之后到达设备(这非常类似于内存
+屏障操作,mb(),不过仅适用于I/O)。
+
+假设一个设备驱动程的具体例子:
+
+ ...
+CPU A: spin_lock_irqsave(&dev_lock, flags)
+CPU A: val = readl(my_status);
+CPU A: ...
+CPU A: writel(newval, ring_ptr);
+CPU A: spin_unlock_irqrestore(&dev_lock, flags)
+ ...
+CPU B: spin_lock_irqsave(&dev_lock, flags)
+CPU B: val = readl(my_status);
+CPU B: ...
+CPU B: writel(newval2, ring_ptr);
+CPU B: spin_unlock_irqrestore(&dev_lock, flags)
+ ...
+
+上述例子中,设备可能会先接收到newval2的值,然后接收到newval的值,问题就
+发生了。不过很容易通过下面方法来修复:
+
+ ...
+CPU A: spin_lock_irqsave(&dev_lock, flags)
+CPU A: val = readl(my_status);
+CPU A: ...
+CPU A: writel(newval, ring_ptr);
+CPU A: (void)readl(safe_register); /* 配置寄存器?*/
+CPU A: spin_unlock_irqrestore(&dev_lock, flags)
+ ...
+CPU B: spin_lock_irqsave(&dev_lock, flags)
+CPU B: val = readl(my_status);
+CPU B: ...
+CPU B: writel(newval2, ring_ptr);
+CPU B: (void)readl(safe_register); /* 配置寄存器?*/
+CPU B: spin_unlock_irqrestore(&dev_lock, flags)
+
+在解决方案中,读取safe_register寄存器,触发IO芯片清刷未处理的写操作,
+再处理后面的读操作,防止引发数据不一致问题。