Date: Wed, 20 Jun 2001 10:54:54 +1000 From: Peter Jeremy <peter.jeremy@alcatel.com.au> To: John Baldwin <jhb@FreeBSD.ORG> Cc: Paul Halliday <dp@penix.org>, alpha@FreeBSD.ORG Subject: Re: SRM Help Message-ID: <20010620105454.U95583@gsmx07.alcatel.com.au> In-Reply-To: <XFMail.010619160031.jhb@FreeBSD.org>; from jhb@FreeBSD.ORG on Tue, Jun 19, 2001 at 04:00:31PM -0700 References: <3B2FDB73.AA21BD43@penix.org> <XFMail.010619160031.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2001-Jun-19 16:00:31 -0700, John Baldwin <jhb@FreeBSD.ORG> wrote: > Basically, I want to do the >equivalent of 'e -gpr ra' for a secondary CPU. I've looked through the manuals for both 4100 and ES40's and it's not clear how deposit and examine handle per-cpu resources (registers) when there are multiple CPUs (the ES40 manuals side-step the issue by not documenting the deposit or examine commands). It would seem reasonable that if you stop only 1 CPU then examine refers to that CPU, but this doesn't scale if you have more than 1 CPU stopped. > Hmm, seems -ipr is either a subset or superset of -gpr. The 4100 manuals document the following address spaces for deposit/examine: fpr Floating-point registers F0-F31 gpr General register R0-R31 ipr Internal processor register pt PALtemp register PT0-PT31 pmem physical memory vmem virtual memory Based on my reading of a 21066 (LCA) datasheet (the only one I have handy): fpr, gpr and ipr are disjoint, with pt a subset of ipr. The LCA HW_MFPR and HW_MTPR instructions support PALtemp, load-store unit (LSU) and instruction decode unit (IDU) register sets. My past experimenting on a Multia suggests that the ipr space is discontiguous (and attempting to reference an illegal register triggers an SRM exception), which makes it difficult to identify all the ipr's without knowing them in advance. > (If anyone has real, nice detailed information on SRM commands in >general that would also be nice.) I'll have a look and see what I have soft-copies of. Whilst I'm at it, does anyone have information on the XDELTA debugger? (The thing that produces prompts like "0012FF08! BPT" when you've really stuffed things up). Just being able to reliably get back to SRM from it would be nice... Peter To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010620105454.U95583>