Windows 64 位和 32 位之间的共享内存接口
我需要在 Windows 7（64 位）中编写代码来执行具有共享内存接口 (SMI) 的 32 位程序。更准确地说，我正在编写的程序写入 SMI，而 32 位程序则从此 SMI 读取。
我遇到的第一个问题是我无法访问32位程序的源代码，这个问题无法解决。第二个问题是SMI存储写入信息的地址。使用以下代码将该指向存储为基指针：
gpSharedBlock->m_pData[uiDataPointer] = (char __based(gpSharedBlock)*)pData;
如果 pData 是指向我们正在写入的数据的指针，则 gpSharedBlock->m_pData[i] 指向存储的第 i 个元素。
也许从这里你已经注意到了这个问题； W32 中的指针为 4 个字节，而 W64 中的指针为 8 个字节。那么，由于存储的值是一个64位的指针，所以32位程序最终读取到的值并不是想要的值。
我的问题是：有没有办法将 64 位地址转换为 32 位地址，以便正在运行的程序读取正确的信息？
我读过有关 WOW64 的内容，并且我认为 W32 程序正在其下运行，但我不知道如何利用它。有什么想法吗？
I need to write code in Windows 7 (64 bits) that executes a 32-bits program that has a Shared Memory Interface (SMI). More precisely, the program I am coding writes into the SMI and the 32-bits program reads from this SMI.
The first problem that I have is that I don't have access to the source code of the 32-bit program, problem that can't be solved. The second problem is that the SMI stores the address of the information that is written. This pointed is stored as a based pointer using the following code:
gpSharedBlock->m_pData[uiDataPointer] = (char __based(gpSharedBlock)*)pData;
Were pData is a pointer to the data we are writing, and gpSharedBlock->m_pData[i] points to the i^th element stored.
Probably from here you have already noticed the problem; a pointer in W32 is 4 bytes while a pointer in W64 is 8 bytes. Then, since the value stored is a 64 bit pointer, the value finally read by the 32-bits program is not the desired one.
My question is: is there a way to do a translation of the 64-bit address to a 32-bit address such that the program that is running reads the correct information?
I have read about WOW64, and I suppose that the W32 program is running under it, but I don't know how to take advantage of that. Any ideas?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
__based指针是相对于另一个指针的数字偏移量。它实际上是一个在运行时解释的虚拟指针。64 位中的指针为 8 个字节，因此为了与 32 位程序兼容，您必须在 64 位代码中声明 SharedBlock 类型的指针成员以使用 4 位整数而不是指针，例如：
pData是gpSharedBlock上的__based，因此pData的值是相对于gpSharedBlock。使用该事实来确定数据块相对于gpSharedBlock内存块的实际字节偏移量，然后将该偏移值作为整数存储到m_pData[]中。这就是 SMI 内存块实际上所期望的——一个偏移量，而不是一个真正的指针。__based关键字只是一种使用指针处理偏移量的奇特方法，无需在代码中手动进行偏移量计算。原始代码实际上与以下代码相同，不需要
__based关键字：A
__basedpointer is a numeric offset from another pointer. It is effectively a virtual pointer interpretted at runtime.A pointer is 8 bytes in 64-bit, so to be compatible with the 32-bit program, you will have to declare the pointer members of the SharedBlock type in your 64-bit code to use 4-bit integers instead of pointers, eg:
pDatais__basedongpSharedBlock, so the value ofpDatais a relative offset from the value ofgpSharedBlock. Use that fact to determine the actual byte offset of your data block relative to thegpSharedBlockmemory block, and then store that offset value intom_pData[]as an integer. That is what the SMI memory block is actually expecting anyway - an offset, not a real pointer. The__basedkeyword is just a fancy way of handling offsets using pointers without doing the offset calculations manually in code.The original code is effectively the same as the following without needing the
__basedkeyword: