home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!paladin.american.edu!darwin.sura.net!Sirius.dfn.de!chx400!bernina!almesber
- From: almesber@nessie.cs.id.ethz.ch (Werner Almesberger)
- Subject: BSS initialization (was Re: IMPORTANT [BUG in 0.99] Re: [ANNOUNCE]: linux version 0.99)
- Message-ID: <1992Dec25.002626.23667@bernina.ethz.ch>
- Keywords: kernel linux 0.99, bss initialization
- Sender: news@bernina.ethz.ch (USENET News System)
- Organization: Swiss Federal Institute of Technology (ETH), Zurich, CH
- References: <1992Dec22.054316.15893@sol.UVic.CA> <2B3751A7.A2E8@tct.com> <20367@ksr.com>
- Date: Fri, 25 Dec 1992 00:26:26 GMT
- Lines: 23
-
- In article <20367@ksr.com> zdenek@ksr.com (Zdenek Radouch) writes:
- > Also, why does the init code run in 16-bit mode? I'd assume that the first
- > five instructions in the boot code would put the processor in some
- > reasonable mode.
-
- The reason why so much is done in 16 bit mode is probably because it's
- easier to access the BIOS this way and because there are comparably few
- occasions where 32 bit arithmetic is necessary.
-
- About zeroing BSS: LILO currently does that (unstripped kernels since 0.3,
- traditional kernel images since 0.6 (*)). However, it might be a good idea
- to handle it in setup or in the kernel itself, because Shoelace, the
- floppy boot sector and probably also BOOTLIN don't initialize BSS.
-
- (*) It may miss a few bytes when using traditional kernels, but this
- shouldn't really be a problem. I might fix this inversion 0.8.
-
- - Werner
- --
- _________________________________________________________________________
- / Werner Almesberger, ETH Zuerich, CH almesber@nessie.cs.id.ethz.ch /
- / IFW A44 Tel. +41 1 254 7213 almesberger@rzvax.ethz.ch /
- /_BITNET:_ALMESBER@CZHETH5A__HEPNET/CHADNET:_[20579::]57414::ALMESBERGER_/
-