| To: | Pete Popov <ppopov@pacbell.net>, "linux-mips@oss.sgi.com" <linux-mips@oss.sgi.com> |
|---|---|
| Subject: | Re: edata alignment |
| From: | Justin Carlson <carlson@sibyte.com> |
| Date: | Fri, 6 Apr 2001 10:01:25 -0700 |
| In-reply-to: | <3ACD42DD.A9E0A0E7@pacbell.net> |
| Organization: | Sibyte |
| References: | <3ACD42DD.A9E0A0E7@pacbell.net> |
| Reply-to: | carlson@sibyte.com |
| Sender: | owner-linux-mips@oss.sgi.com |
On Thu, 05 Apr 2001, Pete Popov wrote: > In arch/mips/kernel/head.S, there is this code in kernel_entry: > > la t0, _edata > sw zero, (t0) > > What guarantees that edata will be word aligned? I don't see a .ALIGN > directive in the ld.script so is it safe to assume that edata will > always be at least word aligned? I've linked into the kernel a very > large ramdisk, and edata ends up being an odd address, causing a cpu > fault. This was fixed between revs 1.4 and 1.5 in cvs @ oss.sgi.com -Justin |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: Does Linux support RC32332 CPU now?, Quinn Jensen |
|---|---|
| Next by Date: | set_cp0_status (mipsregs.h), Scott A McConnell |
| Previous by Thread: | edata alignment, Pete Popov |
| Next by Thread: | Re: first packages for mipsel, Florian Lohoff |
| Indexes: | [Date] [Thread] [Top] [All Lists] |