Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Poul-Henning Kamp
In message <201703281555.v2sftosu005...@pdx.rh.cn85.dnsmgr.net>, "Rodney W. Gri mes" writes: >I think we still have an 8k size limit on boot1 for ffs/(ufs1 or ufs2) Having a former release-engineer & disk-I/O renovator on the UFS2 team took care of that: We tried to be future

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Bruce Evans
On Tue, 28 Mar 2017, Rodney W. Grimes wrote: On Mon, 27 Mar 2017, Julian Elischer wrote: On Tue, 28 Mar 2017, Bruce Evans wrote: [...] they have to fit below 640K and a few multiples of 64K are already used for buffers). The limit on 8K is mainly a historical mistake. A limit of 7.5K

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Toomas Soome
> On 28. märts 2017, at 18:55, Rodney W. Grimes > wrote: > >> On Mon, 27 Mar 2017, Julian Elischer wrote: >> >>> On Tue, 28 Mar 2017, Bruce Evans wrote: >>> >>> [...] >>> they have to fit below 640K and a few multiples of 64K are already used for

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Rodney W. Grimes
> On Mon, 27 Mar 2017, Julian Elischer wrote: > > > On Tue, 28 Mar 2017, Bruce Evans wrote: > > > > [...] > > > >> they have to fit below 640K and a few multiples of 64K are already > >> used for buffers). The limit on 8K is mainly a historical mistake. > >> A limit of 7.5K simplified booting

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Bruce Evans
On Tue, 28 Mar 2017, Warner Losh wrote: [[ sorry for the top post but it's quick ]] It's quicker and almost equivalent to delete what you reply to. Summary, in Bruce's own words the whole optimization step was silly. I agree. on my -current system, clang compiled boot2 was 4 bytes

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Bruce Evans
On Mon, 27 Mar 2017, Julian Elischer wrote: On Tue, 28 Mar 2017, Bruce Evans wrote: [...] they have to fit below 640K and a few multiples of 64K are already used for buffers). The limit on 8K is mainly a historical mistake. A limit of 7.5K simplified booting from 15-sector floppies.

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Warner Losh
[[ sorry for the top post but it's quick ]] Summary, in Bruce's own words > the whole optimization step was silly. I agree. on my -current system, clang compiled boot2 was 4 bytes smaller after ripping it out. gcc was a whopping 7 bytes larger. Since gcc has 156 still free, I think it's best to

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-28 Thread Julian Elischer
On Tue, 28 Mar 2017, Bruce Evans wrote: [...] they have to fit below 640K and a few multiples of 64K are already used for buffers). The limit on 8K is mainly a historical mistake. A limit of 7.5K simplified booting from 15-sector floppies. 18-sector My memory says that the limit of 7.5K

Re: svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-27 Thread Bruce Evans
On Mon, 27 Mar 2017, Warner Losh wrote: Log: Fix build with path names with 'align' or 'nop' in them. clang is now inserting .file directives with the entire path in them. This is fine, except that our sed peephole optimizer removes them if ${SRCTOP} or ${OBJTOP} contains 'align' or 'nop',

svn commit: r316064 - head/sys/boot/i386/boot2

2017-03-27 Thread Warner Losh
Author: imp Date: Mon Mar 27 22:53:36 2017 New Revision: 316064 URL: https://svnweb.freebsd.org/changeset/base/316064 Log: Fix build with path names with 'align' or 'nop' in them. clang is now inserting .file directives with the entire path in them. This is fine, except that our sed