Re: [ros-dev] [ros-diffs] [cgutman] 52125: [AFD] - Fix IoModifyAccess vs IoReadAccess confusion

[Timo Kreuzer]

I can't get the logic behind this change. Can you explain?

Am 07.06.2011 00:55, schrieb [e-mail]:
> Author: cgutman
> Date: Mon Jun  6 22:55:17 2011
> New Revision: 52125
>
> URL: http://svn.reactos.org/svn/reactos?rev=52125&view=rev
> Log:
> [AFD]
> Fix IoModifyAccess vs IoReadAccess confusion
>
> Modified:
>      trunk/reactos/drivers/network/afd/afd/lock.c
>
> Modified: trunk/reactos/drivers/network/afd/afd/lock.c
> URL: http://svn.reactos.org/svn/reactos/trunk/reactos/drivers/network/afd/afd/loc...
> ==============================================================================
> --- trunk/reactos/drivers/network/afd/afd/lock.c [iso-8859-1] (original)
> +++ trunk/reactos/drivers/network/afd/afd/lock.c [iso-8859-1] Mon Jun  6 22:55:17 2011
> @@ -123,7 +123,7 @@
>    	AFD_DbgPrint(MID_TRACE,("Probe and lock pages\n"));
>    	_SEH2_TRY {
>    	    MmProbeAndLockPages( MapBuf[i].Mdl, KernelMode,
> 		         Write ? IoModifyAccess : IoReadAccess );
> 		         Write ? IoReadAccess : IoModifyAccess );
> } _SEH2_EXCEPT(EXCEPTION_EXECUTE_HANDLER) {
>    LockFailed = TRUE;
> } _SEH2_END;

[Alex Ionescu]

Based on a cursory glance, it looks like AFD sets "Write" to TRUE when calling LockBuffers for reading data (read.c), and FALSE when calling LockBuffers for writing data (write.c).

So this change probably matches up AFD's definition of "Write" (ie: I'm using this buffer to WRITE TO THE SOCKET) with the I/O managers (which means I AM READING YOUR USER BUFFER).

--
Best regards,
Alex Ionescu

On 2011-06-07, at 5:07 AM, Timo Kreuzer wrote:
> > IoReadAccess