Basically it is not possible to sanely boot a 32-bit kernel with a 64-bit EFI; the opposite is supported, however. Why would you run a 32-bit kernel these days, though? -hpa
> Basically it is not possible to sanely boot a 32-bit kernel with a > 64-bit EFI; the opposite is supported, however. > > Why would you run a 32-bit kernel these days, though? > > -hpa >FWIW... Reports from users contradict such claims. I am not saying it "as a rule" nor as a scientific poll, but there are many more reports about successfully booting a 32-bit kernel in a UEFI X64 system than reports about booting a 64-bit kernel in a UEFI IA32 system (with a 64-bit microprocessor). Reports might be inaccurate, but that's what has been reported. Considering the very poor / lack of support for Syslinux-related packages in Linux distributions using Syslinux 6.xx, I am not sure how to "read" similar / equivalent reports whether using syslinux.efi or using other UEFI bootloaders. Regards, Ady.
> Why would you run a 32-bit kernel these days, though?Because my GOOD programs are on 32bit. After my system/s were all stolen, I bought a laptop [to stay light, to be able to flee if/when this place goes like the rest of Africa]. With a USB to IDE & SATA, I'm able to run some of my old programs. But I can't understand how/why some 32bit proprams run under this 64bit linux and others don't.>From my memory of 8bit uProcs, I think when Intel upgraded to 16bit,the 8bit-instruction-set remained valid? On 4/3/17, Ady Ady via Syslinux <syslinux at zytor.com> wrote:> >> Basically it is not possible to sanely boot a 32-bit kernel with a >> 64-bit EFI; the opposite is supported, however. >> >> Why would you run a 32-bit kernel these days, though? >> >> -hpa >> > > > FWIW... > > Reports from users contradict such claims. I am not saying it "as a > rule" nor as a scientific poll, but there are many more reports about > successfully booting a 32-bit kernel in a UEFI X64 system than reports > about booting a 64-bit kernel in a UEFI IA32 system (with a 64-bit > microprocessor). > > Reports might be inaccurate, but that's what has been reported. > > Considering the very poor / lack of support for Syslinux-related > packages in Linux distributions using Syslinux 6.xx, I am not sure how > to "read" similar / equivalent reports whether using syslinux.efi or > using other UEFI bootloaders. > > Regards, > Ady. > > _______________________________________________ > Syslinux mailing list > Submissions to Syslinux at zytor.com > Unsubscribe or set options at: > http://www.zytor.com/mailman/listinfo/syslinux >
ok - i do understand the point. and i do guess the hint was also
included in ady's answer.
why running a 32-bit kernel? well because of efficency. i am running
ldap, bdb, mysql db, webservers, e-mail server with imap service on a
512mB RAM kernel without swapping of over 10 years. so - why should i
step up to a 4gB environment? it's all about point of view and
capabilities. i am also developing on 8bit microcontrollers. works fine
on a 20cents cpu as long as you know how to use it and no oo and no java
is envolved. yes - different needs.
other reason would be ... it is not possible to do so. as i do
understand your statement.
hard to understand now .... i have got a x64 system, the vendor claims
to be good for running 32-bit windows and 64-bit windows systems on. 2nd
statement from intel saying 64-bit cpu is backward compatible to run
32-bit applications. can i understand an OS as an application?
so where is the key to glue all statements to one single picture? how
looks the korrekt picture - please?
|-syslinux.32 -> linux
32-bit
|- boot32.efi - |
| |-syslinux.64 -> linux
32/64-bit
x64 UEFI - |
| |-syslinux.32-> linux 32-bit
|- boot64.efi- |
|-syslinux.64->
linux32/64-bit
thank you.
Am 03.04.2017 23:12, schrieb H. Peter Anvin via
Syslinux:> Basically it is not possible to sanely boot a 32-bit kernel with a
> 64-bit EFI; the opposite is supported, however.
>
> Why would you run a 32-bit kernel these days, though?
>
> -hpa
>
> _______________________________________________
> Syslinux mailing list
> Submissions to Syslinux at zytor.com
> Unsubscribe or set options at:
> http://www.zytor.com/mailman/listinfo/syslinux
> ok - i do understand the point. and i do guess the hint was also > included in ady's answer. >Hmm... I'm not sure I understand. I do not recall any specific hint of mine in this email thread. "TL;DR": I (still) claim that Peter's statements in this email thread are incorrect at this time. I do not know whether the following clarification (or attempt of it) is really needed / welcomed, but just in case... Regarding the potential need, or choice, for someone to use a 32-bit kernel nowadays, I have no problem, criticism, question, doubt... I could find several reasons for such use-case, still in these days. I took this detail as known data to be considered as part of the initial matter/question. Hence, in my reply to Peter (hpa), I had nothing to say about his question / phrase / comment regarding using a 32-bit kernel (in these days). On the other hand, my reply to Peter was very much focusing on his claim regarding which UEFI bootloader(s) (i.e. x32/x64) is/are capable of booting which kernel (32/64-bits) in which hardware (AFAICT, "x86_64" is the only type of microprocessor that is relevant to this matter in the Syslinux Mailing List). With all due respect, I think Peter's claims are not correct. As far as I can tell, several reports would indicate that my statements about this matter are accurate at this time, and they contradict those posted by Peter in this email thread. When I replied to Peter's email, my only intention was (and still is) to avoid potential misunderstandings from users/readers. If my statements were not clear enough, or if my wording could be potentially (mis)interpreted as hinting about some other matter(s), it wasn't my intention and I apologize for it. But, of course, I could also be wrong. If it turns out that my prior email in this email thread included inaccurate / incorrect statements, or if my current understanding of users' reports happens to be incorrect in any way, I would appreciate a clarification and explanations about the matter. I think a more-technical explanation (perhaps based on actual source code?) would be a better way to clarify the matter so as to avoid misunderstandings in the future. Unfortunately, the chances that the main author of the relevant UEFI code in Syslinux would read this email thread and reply here is very low. Regards, Ady. PS: A: Because it fouls the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing in e-mail?