Re: [edk2] GDB MMIO Access Issue

Subject: Re: [edk2] GDB MMIO Access Issue

From: Andrew Fish <afish@apple.com>

To: Andrei Warkentin <andreiw@motorola.com>

Date: 2011-07-26 16:24:14



On Jul 26, 2011, at 8:07 AM, Andrei Warkentin  wrote:

> On Tue, Jul 26, 2011 at 10:01 AM, Cohen, Eugene  wrote:
>> I've tested that this fixes the single MMIO access case (which our debug scripts need).
>> 
>> It does not fix access to a MMIO range through a memory dump window (at least when using our debugger front end).  I'm not sure what the best way to improve this is -- the GDB remote serial protocol does not differentiate access size from total length of transfer.  One option would be for the debugger to be made aware of RAM vs MMIO and alter its behavior.  It would be nice if this info could be fed from the host so it does not have any more coupling to the target.
> 
> Might not be the best idea, but you could always define an arbitrary
> command inside the gdb stubs to do the right thing as far as MMIO
> behavior, an then trigger it from GDB via the 'remote commandstring'
> syntax.
> 
> Another idea is to be able to register I/O ranges with GDB stubs,
> specifying access type, so the appropriate read/write is used on
> access.
> 
Maybe we could make a large buffer work more like memcpy? That is do UINTN accesses if the address is aligned and byte accesses for the unaligned part?

Thanks

Andrew

------------------------------------------------------------------------------
Magic Quadrant for Content-Aware Data Loss Prevention
Research study explores the data loss prevention market. Includes in-depth
analysis on the changes within the DLP market, and the criteria used to
evaluate the strengths and weaknesses of these DLP solutions.
http://www.accelacomm.com/jaw/sfnl/114/51385063/
_______________________________________________
edk2-devel mailing list
edk2-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/edk2-devel