From owner-cvs-all Tue Jan 8 12: 3:13 2002 Delivered-To: cvs-all@freebsd.org Received: from salmon.maths.tcd.ie (salmon.maths.tcd.ie [134.226.81.11]) by hub.freebsd.org (Postfix) with SMTP id 2906E37B41C; Tue, 8 Jan 2002 12:03:04 -0800 (PST) Received: from walton.maths.tcd.ie by salmon.maths.tcd.ie with SMTP id ; 8 Jan 2002 20:03:03 +0000 (GMT) To: Wilko Bulte Cc: Sheldon Hearn , John Baldwin , cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/release/sysinstall install.c In-Reply-To: Your message of "Tue, 08 Jan 2002 19:30:27 +0100." <20020108193027.A16263@freebie.xs4all.nl> Date: Tue, 08 Jan 2002 20:03:03 +0000 From: Ian Dowse Message-ID: <200201082003.aa40140@salmon.maths.tcd.ie> Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG In message <20020108193027.A16263@freebie.xs4all.nl>, Wilko Bulte writes: > >What still puzzles me is why it works for some us without problems. >Does anyone have ideas on that one? I think it may relate to the size of the file being loaded by boot1. What size is /boot/loader on the alpha? A 16k data or directory block written to the data buffer `iobuf' will overwrite the 8k indirection buffer `mapbuf', and a 16k indirection block written there will overwrite other globals. So if there are no indirection blocks then maybe it will work (size <= 196608 bytes). There is also an optimisation in readit() where block-aligned data is read directly into the caller's buffer, which may confuse things. In the i386 case it probably works for similar reasons, though the data layout is quite different. I suspect that loading a kernel directly from boot2 will fail reliably on a 16k/2k filesystem. Ian To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message