From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 17515 invoked by alias); 26 Feb 2008 16:39:17 -0000 Received: (qmail 17498 invoked by uid 22791); 26 Feb 2008 16:39:16 -0000 X-Spam-Check-By: sourceware.org Received: from dmz.mips-uk.com (HELO dmz.mips-uk.com) (194.74.144.194) by sourceware.org (qpsmtpd/0.31) with ESMTP; Tue, 26 Feb 2008 16:38:58 +0000 Received: from internal-mx1 ([192.168.192.240] helo=ukservices1.mips.com) by dmz.mips-uk.com with esmtp (Exim 3.35 #1 (Debian)) id 1JU2pT-0005lC-00; Tue, 26 Feb 2008 16:38:55 +0000 Received: from perivale.mips.com ([192.168.192.200]) by ukservices1.mips.com with esmtp (Exim 3.36 #1 (Debian)) id 1JU2pP-0001LC-00; Tue, 26 Feb 2008 16:38:51 +0000 Received: from macro (helo=localhost) by perivale.mips.com with local-esmtp (Exim 4.63) (envelope-from ) id 1JU2pP-0006Fm-EM; Tue, 26 Feb 2008 16:38:51 +0000 Date: Tue, 26 Feb 2008 16:47:00 -0000 From: "Maciej W. Rozycki" To: Daniel Jacobowitz cc: gdb-patches@sourceware.org, David Ung , Nigel Stephens , "Maciej W. Rozycki" Subject: Re: MDI: testsuite support In-Reply-To: <20080226161716.GB19531@caradoc.them.org> Message-ID: References: <20080226161716.GB19531@caradoc.them.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-MIPS-Technologies-UK-MailScanner: Found to be clean X-MIPS-Technologies-UK-MailScanner-From: macro@mips.com Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2008-02/txt/msg00394.txt.bz2 On Tue, 26 Feb 2008, Daniel Jacobowitz wrote: > Let's hold off on this patch until the rest of the MDI target is > posted, please. I appreciate your separating it into more manageable > pieces but at the same time they are hard to consider in isolation > and I want to be sure the whole thing goes in. Well, I had promised to send the testsuite bits shortly, so I did it. :-) Which "rest of the MDI target" are you specifically referring to? The only piece I have got left is support for some obscure "set mdi" and "show mdi" commands that should not be subcommands of "set" and "show" in the first place as they operate on the target (which is against the principle of the commands) and the latter also breaks "show" as used to print all the settings. However if you'd like to see this bit on the list as-is, then I am fine with that, but I do not consider it ready for submission. My current feeling is they should be reimplemented as separate commands -- while the affected "show mdi" commands might fit a command like "info mdi", there is apparently no corresponding command for "set mdi", so I suppose a pair of commands like "mdi set" and "mdi show" might sound better. Note I mean these commands are really obscure and are used things like to peek or poke at TLB entries and cache lines "through a back door", so I gather whoever is skilled enough to dare would know what they are doing and get used to whatever the commands are called. In fact hardly anybody uses them as proved by an embarrasing bug I fixed a while ago somewhere in this area. ;-) They are occasionally useful for stuff like debugging system bringup issues, weird cache corruption problems and suchlike. Maciej