Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJAsdNj716Ddq+H94Tu52u1Kwz6VPKfTzUQyHV1x7qcbVvFtxA@mail.gmail.com>
Date: Mon, 5 Aug 2013 23:47:11 +0200
From: Dániel Bali <balijanosdaniel@...il.com>
To: john-dev@...ts.openwall.com
Subject: [GSoC] GCN Local Data Storage problems

Hello!

I've been having a problem with LDS for a week and Lukas suggested to post
here so everybody can see the problem and hopefully someone can help.
Here is the thread where I asked about LDS on the AMD forums:
http://devgurus.amd.com/thread/167167

The answer I got wasn't totally clear to me at first but after looking at
the binary for a long time I realized what it means.
Max LDS size and actual LDS size are the numbers that come after the
integer values 0x80000081, and 0x80000082.

I changed these values in the dummy binary that I patch but it didn't help.
The LDS read instruction still loads 0 into the VGPR.

There was another LDS size related value described:
SetCalNote($00002e13,(ldsSizeBytes+255)shr 8 shl 15,$FFF07FFF{and mask});
<- this updates the ldsSize field in the COMPUTE_PGM_RSRC2 structure.

The byte value after this 0x00002e13 is 0x98, which doesn't make any sense
in this context.

Does anyone know what may be wrong?
Should I just keep asking on the forums?

Thanks,
Daniel

Content of type "text/html" skipped

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.