[Openocd-development] retire arm920t and arm720t cp15 commands
Would anyone mind if the arm920t and arm720t cp15 commands were retired? The only user of arm920t in the tcl directory is csb337.cfg. If someone could offer a translation of the "arm920t cp15 2 0xc0001078" command to "mcr" syntax + test it. We'd be ready to retire the arm920t/arm720t commands in favor of the generic mrc/mcr commands. -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 ARM11 XScale Cortex JTAG debugger and flash programmer ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] retire arm920t and arm720t cp15 commands
On Mon, 2009-11-23 at 14:19 +0100, Øyvind Harboe wrote: > Would anyone mind if the arm920t and arm720t cp15 commands > were retired? > > The only user of arm920t in the tcl directory is csb337.cfg. > > If someone could offer a translation of the "arm920t cp15 2 0xc0001078" > command to "mcr" syntax + test it. We'd be ready to retire > the arm920t/arm720t commands in favor of the generic mrc/mcr commands. I have a patch to convert 720t command to the new command registration. If you're going to drop it, then I'll skip that part of 920t for now, but are there other bits that may disappear in the next day or so (i.e. before I wrap up and eventually push my series)? --Z ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] retire arm920t and arm720t cp15 commands
On Mon, Nov 23, 2009 at 2:55 PM, Zach Welch wrote: > On Mon, 2009-11-23 at 14:19 +0100, Øyvind Harboe wrote: >> Would anyone mind if the arm920t and arm720t cp15 commands >> were retired? >> >> The only user of arm920t in the tcl directory is csb337.cfg. >> >> If someone could offer a translation of the "arm920t cp15 2 0xc0001078" >> command to "mcr" syntax + test it. We'd be ready to retire >> the arm920t/arm720t commands in favor of the generic mrc/mcr commands. > > I have a patch to convert 720t command to the new command registration. > If you're going to drop it, then I'll skip that part of 920t for now, > but are there other bits that may disappear in the next day or so (i.e. > before I wrap up and eventually push my series)? We *could* just drop the cp15 720t/920t commands and then just deal with any fallout for the mrc/mcr implementations... I have no testing feedback that mcr/mcr works or is complete enough to replace 720t/920t cp15 commands, but I believe so... Essentially there will then be one script that ships with OpenOCD that has a syntax error which will need a rewrite to the mcr syntax. -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 ARM11 XScale Cortex JTAG debugger and flash programmer ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] retire arm920t and arm720t cp15 commands
On Monday 23 November 2009, Øyvind Harboe wrote: > The only user of arm920t in the tcl directory is csb337.cfg. I can change that, assuming the replacement code actually works, but I want to finish the arm11 register rework first ... first patches on that will come later today, I hope. BTW I thnk you broke the arm926 MCR script conversions. You didn't actually say that "cp15" is involved... (Turns out the ARM11 and Cortex-A8 debug modules fit easily into a common framework. First, registers ... which gives a whole lot of stuff. Then, breakpoints and/or watchpoints Then, the world!!!) ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] retire arm920t and arm720t cp15 commands
On Mon, Nov 23, 2009 at 6:10 PM, David Brownell wrote: > On Monday 23 November 2009, Ųyvind Harboe wrote: >> The only user of arm920t in the tcl directory is csb337.cfg. > > I can change that, assuming the replacement code actually works, > but I want to finish the arm11 register rework first ... first > patches on that will come later today, I hope. That would be fantastic. > BTW I thnk you broke the arm926 MCR script conversions. You > didn't actually say that "cp15" is involved... Crap. Thanks! The first argument should be 15. -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 ARM11 XScale Cortex JTAG debugger and flash programmer ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] retire arm920t and arm720t cp15 commands
How's this? -- Øyvind Harboe http://www.zylin.com/zy1000.html ARM7 ARM9 ARM11 XScale Cortex JTAG debugger and flash programmer From 828d006a9d05b24b6dcdf1c552912e04586d6f7d Mon Sep 17 00:00:00 2001 From: =?utf-8?q?=C3=98yvind=20Harboe?= Date: Mon, 23 Nov 2009 18:23:10 +0100 Subject: [PATCH] arm926ejs: fix gaffe when converting from arm926ejs cp15 to mcr MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit the first arg is the register number 15 = cp15. Signed-off-by: Øyvind Harboe --- tcl/board/dm355evm.cfg |2 +- tcl/board/openrd.cfg |2 +- tcl/board/sheevaplug.cfg |2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/tcl/board/dm355evm.cfg b/tcl/board/dm355evm.cfg index 2e298b7..2c8bea8 100644 --- a/tcl/board/dm355evm.cfg +++ b/tcl/board/dm355evm.cfg @@ -182,7 +182,7 @@ proc dm355evm_init {} { # turn on icache - set I bit in cp15 register c1 - mcr 0 0 1 0 0x00051078 + mcr 15 0 0 1 0 0x00051078 } # NAND -- socket has two chipselects, MT29F16G08FAA puts 1GByte on each one. diff --git a/tcl/board/openrd.cfg b/tcl/board/openrd.cfg index 4bc708d..12cc79e 100644 --- a/tcl/board/openrd.cfg +++ b/tcl/board/openrd.cfg @@ -29,7 +29,7 @@ proc openrd_init { } { jtag_reset 0 0 wait_halt - mcr 0 0 1 0 0x00052078 + mcr 15 0 0 1 0 0x00052078 mww 0xD0001400 0x43000C30 # DDR SDRAM Configuration Register mww 0xD0001404 0x37543000 # Dunit Control Low Register diff --git a/tcl/board/sheevaplug.cfg b/tcl/board/sheevaplug.cfg index 8e8396d..9267eb9 100644 --- a/tcl/board/sheevaplug.cfg +++ b/tcl/board/sheevaplug.cfg @@ -29,7 +29,7 @@ proc sheevaplug_init { } { jtag_reset 0 0 wait_halt - mcr 0 0 1 0 0x00052078 + mcr 15 0 0 1 0 0x00052078 mww 0xD0001400 0x43000C30 # DDR SDRAM Configuration Register mww 0xD0001404 0x39543000 # Dunit Control Low Register -- 1.6.3.3 ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] retire arm920t and arm720t cp15 commands
On Monday 23 November 2009, Øyvind Harboe wrote: > How's this? Seems more plausible, and less certain to automagically break working config scripts. :) ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] retire arm920t and arm720t cp15 commands
On Monday 23 November 2009, Øyvind Harboe wrote: > Would anyone mind if the arm920t and arm720t cp15 commands > were retired? > > The only user of arm920t in the tcl directory is csb337.cfg. I did get around to verifying that there's no regression in that script, despite all the underlying mcr/mrc changes. I'd like to hold off on such changes until the churn in arm920t.c is resolved ... e.g. the current mess with cache stuff, and the general confusion between CP15 access modes. Likewise, those command should get the same kind of year-long grace period as other commands: deprecate first, issue runtime warnings to users, then remove them a year after deprecation. Otherwise we'd be in the position of just yanking commands out from under their users, with no real notice. Recall that there can be a lot of out-of-tree users. And that in these cases, migrating to other commands is nontrivial. - Dave ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development