* [U-Boot] u-boot-x86 @ 2010-08-20 7:42 sk ong 2010-08-21 1:42 ` Graeme Russ 0 siblings, 1 reply; 8+ messages in thread From: sk ong @ 2010-08-20 7:42 UTC (permalink / raw) To: u-boot Hi, I notice there is u-boot for x86 architecture in the denx site. I wonder whether the u-boot-x86 has the legacy BIOS services, such as dispatching PCI/PCIE device option rom such as video, SATA, etc. Does it have int13, int15 etc. services for x86 Linux to call during booting phase. Thanks, SK ^ permalink raw reply [flat|nested] 8+ messages in thread
* [U-Boot] u-boot-x86 2010-08-20 7:42 [U-Boot] u-boot-x86 sk ong @ 2010-08-21 1:42 ` Graeme Russ 2010-08-22 19:21 ` Robert Schwebel [not found] ` <AANLkTinh8372fgDaiukRz-+CBS5jYfSNTjvdbK9gR8Xj@mail.gmail.com> 0 siblings, 2 replies; 8+ messages in thread From: Graeme Russ @ 2010-08-21 1:42 UTC (permalink / raw) To: u-boot Hi SK, On 20/08/10 17:42, sk ong wrote: > Hi, > > I notice there is u-boot for x86 architecture in the denx site. I wonder > whether the u-boot-x86 has the legacy BIOS services, such as dispatching > PCI/PCIE device option rom such as video, SATA, etc. Does it have int13, > int15 etc. services for x86 Linux to call during booting phase. The x86 port has an (extremely) limited BIOS primarily to provide INT12 and INT15 for the purpose of determining memory size during the Linux Real-Mode startup procedure. I don't think that Linux actually needs BIOS other than during the initial Real-Mode bootstrap. In an embedded environment, the board designer knows all the information the BIOS would otherwise provide. My goal is to remove the BIOS functionality from U-Boot and use the Protected-Mode entry point of the Linux kernel and avoid the switching back-and-forth between Real and Protected Modes It sounds like you might be looking at using U-Boot for booting a x86 PC. If this is the case, maybe you should take a look at coreboot (http://www.coreboot.org/) Regards, Graeme ^ permalink raw reply [flat|nested] 8+ messages in thread
* [U-Boot] u-boot-x86 2010-08-21 1:42 ` Graeme Russ @ 2010-08-22 19:21 ` Robert Schwebel [not found] ` <AANLkTinh8372fgDaiukRz-+CBS5jYfSNTjvdbK9gR8Xj@mail.gmail.com> 1 sibling, 0 replies; 8+ messages in thread From: Robert Schwebel @ 2010-08-22 19:21 UTC (permalink / raw) To: u-boot On Sat, Aug 21, 2010 at 11:42:31AM +1000, Graeme Russ wrote: > It sounds like you might be looking at using U-Boot for booting a x86 > PC. If this is the case, maybe you should take a look at coreboot > (http://www.coreboot.org/) For what it's worth: we have a beginning of x86 support in Barebox (formerly known as U-Boot-v2): http://www.barebox.org It works ontop of the BIOS and can boot into a kernel which is in a raw partition on a BIOS-supported boot device. rsc -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 | ^ permalink raw reply [flat|nested] 8+ messages in thread
[parent not found: <AANLkTinh8372fgDaiukRz-+CBS5jYfSNTjvdbK9gR8Xj@mail.gmail.com>]
* [U-Boot] u-boot-x86 [not found] ` <AANLkTinh8372fgDaiukRz-+CBS5jYfSNTjvdbK9gR8Xj@mail.gmail.com> @ 2010-08-25 11:16 ` Graeme Russ 2010-08-25 13:49 ` sk ong 0 siblings, 1 reply; 8+ messages in thread From: Graeme Russ @ 2010-08-25 11:16 UTC (permalink / raw) To: u-boot Hi Sk, Please keep the Cc list intact (specifically U-Boot Users). Although some of these questions are a little off-topic for the list, there is much confusion about U-Boot as an embedded boot loader, CoreBoot as a BIOS replacement which can boot the Linux kernel directly and x86 'Bootloaders' such as LILO, GRUB etc. Having this information in the mailing list archive is important On 25/08/10 17:58, sk ong wrote: > Graeme, > > Thanks for your response. > > I have not accessed gmail for few days, so I see this late, I will take > some time to digest and look at coreboot too. > > I do not fully understand, can I ask some questions? Sure - Questions are always welcome :) > > Board designer of embedded board knows the information, instead of BIOS, > how does u-boot pass this information to protected-mode entry of linux > kernel? The x86 Linux Kernel has two components, a 'Real-Mode' component (which has access to BIOS but very limited access to memory and devices) and a 'Protected-Mode' component (with no access to BIOS but full access to memory and devices). The 'Real-Mode' component of the Linux Kernel 'probes' BIOS (via BIOS interrupt calls) to build a list of fundamental information (primarily memory layout) in memory and invokes the 'Protected Mode' component passing a memory address of the information collected (from BIOS) in Real-Mode My intent is for U-Boot to replace the 'Real-Mode' component of Linux and build the necessary information and directly invoke the 'Protected Mode' component of the Linux Kernel (bypassing Real-Mode entirely). Because an embedded board designer designs all part of the system (RAM, ROM, device memory and interrupt maps etc), they can build this information without needing BIOS. > > Does uboot execute option rom, such as the option rom for PCI/PCIE video > controller, SATA controller, network controller, etc.? No (although you could write drivers to do so). I think U-Boot may already have drivers for common hardware already. Because of the GPL, any device supported by Linux can be ported to U-Boot be taking the existing Linux driver code (does need to be modified specifically for U-Boot, but all the hard work has already been done) > > What are the x86 CPU and board that can run uboot-x86? Currently U-Boot supports the AMD SC520 'Elan' embedded processor. I am about to port U-Boot to the VIA EPIA EN1200/1500 boards > > Is there a porting guide if I want to port uboot-x86 to another x86 board? > Unfortunately not > Can uboot-x86 boot x86 PC Linux like Fedora, Ubuntu, etc.? U-Boot cannot boot a PC. The VIA EPIA platform is marketed as 'Embedded' but it is (kind-of) a PC motherboard. PC's are very diverse - BIOS is intended to create an abstraction between the software and the hardware to some extent. To boot a PC, U-Boot would need to replace the BIOS and all the abstraction that entails which would, essentially, mimic coreboot (i.e. implementing such 'genericness' would be a complete waste of time) > > If uboot-x86 can boot x86 PC Linux, what stop it from booting Windows > (at least non-ACPI kernel WIndows)? U-Boot can boot other OS's. I think some have even booted Windows CE (the embedded version of windows) > > Without in13h in uboot-x86, how MBR/grub/lilo load Linux image from > storage such as SATA, SD, etc? Most embedded systems use Flash memory to store the filesystem. U-Boot has drivers to access file systems on MTD (Flash) devices. U-Boot also has a wide variety of driver to provide access to various storage devices (as well as network adapters, Serial ports, USB, video, PCMCIA etc) > > I am looking for uboot, not only for x86 PC, but also for other x86 > embedded OS. If you are booting a generic, off-the-shelf PC motherboard (Gigabyte, Tyan, Asus etc) then look at coreboot. If you are creating your own hardware design, use U-Boot > > Thanks, > SK > > > On Sat, Aug 21, 2010 at 9:42 AM, Graeme Russ <graeme.russ@gmail.com > <mailto:graeme.russ@gmail.com>> wrote: > > Hi SK, > > On 20/08/10 17:42, sk ong wrote: > > Hi, > > > > I notice there is u-boot for x86 architecture in the denx site. I > wonder > > whether the u-boot-x86 has the legacy BIOS services, such as > dispatching > > PCI/PCIE device option rom such as video, SATA, etc. Does it have > int13, > > int15 etc. services for x86 Linux to call during booting phase. > > The x86 port has an (extremely) limited BIOS primarily to provide > INT12 and > INT15 for the purpose of determining memory size during the Linux > Real-Mode > startup procedure. > > I don't think that Linux actually needs BIOS other than during the > initial > Real-Mode bootstrap. In an embedded environment, the board designer > knows > all the information the BIOS would otherwise provide. My goal is to > remove > the BIOS functionality from U-Boot and use the Protected-Mode entry > point > of the Linux kernel and avoid the switching back-and-forth between > Real and > Protected Modes > > It sounds like you might be looking at using U-Boot for booting a > x86 PC. > If this is the case, maybe you should take a look at coreboot > (http://www.coreboot.org/) > > Regards, > > Graeme > > ^ permalink raw reply [flat|nested] 8+ messages in thread
* [U-Boot] u-boot-x86 2010-08-25 11:16 ` Graeme Russ @ 2010-08-25 13:49 ` sk ong 2010-08-25 14:49 ` Graeme Russ 0 siblings, 1 reply; 8+ messages in thread From: sk ong @ 2010-08-25 13:49 UTC (permalink / raw) To: u-boot Graeme, Thanks for the information. Now I know I need to take a look at coreboot too. Do you know anyone has or plan to do porting for Intel Atom based uboot or coreboot or other boot loader? Thanks, SK On Wed, Aug 25, 2010 at 7:16 PM, Graeme Russ <graeme.russ@gmail.com> wrote: > Hi Sk, > > Please keep the Cc list intact (specifically U-Boot Users). Although some > of these questions are a little off-topic for the list, there is much > confusion about U-Boot as an embedded boot loader, CoreBoot as a BIOS > replacement which can boot the Linux kernel directly and x86 'Bootloaders' > such as LILO, GRUB etc. Having this information in the mailing list archive > is important > > On 25/08/10 17:58, sk ong wrote: > > Graeme, > > > > Thanks for your response. > > > > I have not accessed gmail for few days, so I see this late, I will take > > some time to digest and look at coreboot too. > > > > I do not fully understand, can I ask some questions? > > Sure - Questions are always welcome :) > > > > > Board designer of embedded board knows the information, instead of BIOS, > > how does u-boot pass this information to protected-mode entry of linux > > kernel? > > The x86 Linux Kernel has two components, a 'Real-Mode' component (which has > access to BIOS but very limited access to memory and devices) and a > 'Protected-Mode' component (with no access to BIOS but full access to > memory and devices). > > The 'Real-Mode' component of the Linux Kernel 'probes' BIOS (via BIOS > interrupt calls) to build a list of fundamental information (primarily > memory layout) in memory and invokes the 'Protected Mode' component passing > a memory address of the information collected (from BIOS) in Real-Mode > > My intent is for U-Boot to replace the 'Real-Mode' component of Linux and > build the necessary information and directly invoke the 'Protected Mode' > component of the Linux Kernel (bypassing Real-Mode entirely). Because an > embedded board designer designs all part of the system (RAM, ROM, device > memory and interrupt maps etc), they can build this information without > needing BIOS. > > > > > Does uboot execute option rom, such as the option rom for PCI/PCIE video > > controller, SATA controller, network controller, etc.? > > No (although you could write drivers to do so). I think U-Boot may already > have drivers for common hardware already. Because of the GPL, any device > supported by Linux can be ported to U-Boot be taking the existing Linux > driver code (does need to be modified specifically for U-Boot, but all the > hard work has already been done) > > > > > What are the x86 CPU and board that can run uboot-x86? > > Currently U-Boot supports the AMD SC520 'Elan' embedded processor. I am > about to port U-Boot to the VIA EPIA EN1200/1500 boards > > > > > Is there a porting guide if I want to port uboot-x86 to another x86 > board? > > > > Unfortunately not > > > Can uboot-x86 boot x86 PC Linux like Fedora, Ubuntu, etc.? > > U-Boot cannot boot a PC. The VIA EPIA platform is marketed as 'Embedded' > but it is (kind-of) a PC motherboard. PC's are very diverse - BIOS is > intended to create an abstraction between the software and the hardware to > some extent. To boot a PC, U-Boot would need to replace the BIOS and all > the abstraction that entails which would, essentially, mimic coreboot (i.e. > implementing such 'genericness' would be a complete waste of time) > > > > > If uboot-x86 can boot x86 PC Linux, what stop it from booting Windows > > (at least non-ACPI kernel WIndows)? > > U-Boot can boot other OS's. I think some have even booted Windows CE (the > embedded version of windows) > > > > > Without in13h in uboot-x86, how MBR/grub/lilo load Linux image from > > storage such as SATA, SD, etc? > > Most embedded systems use Flash memory to store the filesystem. U-Boot has > drivers to access file systems on MTD (Flash) devices. U-Boot also has a > wide variety of driver to provide access to various storage devices (as > well as network adapters, Serial ports, USB, video, PCMCIA etc) > > > > > I am looking for uboot, not only for x86 PC, but also for other x86 > > embedded OS. > > If you are booting a generic, off-the-shelf PC motherboard (Gigabyte, Tyan, > Asus etc) then look at coreboot. If you are creating your own hardware > design, use U-Boot > > > > > Thanks, > > SK > > > > > > On Sat, Aug 21, 2010 at 9:42 AM, Graeme Russ <graeme.russ@gmail.com > > <mailto:graeme.russ@gmail.com>> wrote: > > > > Hi SK, > > > > On 20/08/10 17:42, sk ong wrote: > > > Hi, > > > > > > I notice there is u-boot for x86 architecture in the denx site. I > > wonder > > > whether the u-boot-x86 has the legacy BIOS services, such as > > dispatching > > > PCI/PCIE device option rom such as video, SATA, etc. Does it have > > int13, > > > int15 etc. services for x86 Linux to call during booting phase. > > > > The x86 port has an (extremely) limited BIOS primarily to provide > > INT12 and > > INT15 for the purpose of determining memory size during the Linux > > Real-Mode > > startup procedure. > > > > I don't think that Linux actually needs BIOS other than during the > > initial > > Real-Mode bootstrap. In an embedded environment, the board designer > > knows > > all the information the BIOS would otherwise provide. My goal is to > > remove > > the BIOS functionality from U-Boot and use the Protected-Mode entry > > point > > of the Linux kernel and avoid the switching back-and-forth between > > Real and > > Protected Modes > > > > It sounds like you might be looking at using U-Boot for booting a > > x86 PC. > > If this is the case, maybe you should take a look at coreboot > > (http://www.coreboot.org/) > > > > Regards, > > > > Graeme > > > > > > ^ permalink raw reply [flat|nested] 8+ messages in thread
* [U-Boot] u-boot-x86 2010-08-25 13:49 ` sk ong @ 2010-08-25 14:49 ` Graeme Russ 2010-08-25 15:40 ` sk ong 2010-08-27 10:07 ` Detlev Zundel 0 siblings, 2 replies; 8+ messages in thread From: Graeme Russ @ 2010-08-25 14:49 UTC (permalink / raw) To: u-boot Sk, On 25/08/10 23:49, sk ong wrote: > Graeme, > > Thanks for the information. Now I know I need to take a look at coreboot > too. > > Do you know anyone has or plan to do porting for Intel Atom based uboot > or coreboot or other boot loader? Not specifically however I understand that the Atom is an x86/x64 based processor which makes life a little easier. Depending on the use-case, I think U-Boot could be useful. I notice the Atom is targeted at the Netbook market which Linux is particularly strong which implies good hardware driver support (i.e. re-usable driver code) which could (potentially) make the porting process easier. I notice there is a work in progress for coreboot on only one Intel Atom motherboard (the D945GCLF) - Not a lot of support there. Are you designing the hardware yourself or is this for an existing design? I suppose the next big questions relate to your experience with the x86 architecture. It is a real pain to work with with a steep learning curve Regards, Graeme ^ permalink raw reply [flat|nested] 8+ messages in thread
* [U-Boot] u-boot-x86 2010-08-25 14:49 ` Graeme Russ @ 2010-08-25 15:40 ` sk ong 2010-08-27 10:07 ` Detlev Zundel 1 sibling, 0 replies; 8+ messages in thread From: sk ong @ 2010-08-25 15:40 UTC (permalink / raw) To: u-boot Graeme, I am not familiar with device driver for embedded Linux, but I think the Linux device driver for Atom currently still use the BIOS alike support at least in the booting phase. I think there may be a challenge to port the device driver (video, SATA, ethernet, etc.) to use in uboot from one Atom to another different Atom, unless uboot can reuse directly the device option rom that provides the int services and the device driver that uses the int services in the booting phase. I assume coreboot will be able to have such reuse. However I still see a value to uboot (without PC support) for deeply embedded application using certain Atom. Thanks, SK On Wed, Aug 25, 2010 at 10:49 PM, Graeme Russ <graeme.russ@gmail.com> wrote: > Sk, > > On 25/08/10 23:49, sk ong wrote: > > Graeme, > > > > Thanks for the information. Now I know I need to take a look at coreboot > > too. > > > > Do you know anyone has or plan to do porting for Intel Atom based uboot > > or coreboot or other boot loader? > > Not specifically however I understand that the Atom is an x86/x64 based > processor which makes life a little easier. Depending on the use-case, I > think U-Boot could be useful. I notice the Atom is targeted at the Netbook > market which Linux is particularly strong which implies good hardware > driver support (i.e. re-usable driver code) which could (potentially) make > the porting process easier. > > I notice there is a work in progress for coreboot on only one Intel Atom > motherboard (the D945GCLF) - Not a lot of support there. > > Are you designing the hardware yourself or is this for an existing design? > > I suppose the next big questions relate to your experience with the x86 > architecture. It is a real pain to work with with a steep learning curve > > Regards, > > Graeme > ^ permalink raw reply [flat|nested] 8+ messages in thread
* [U-Boot] u-boot-x86 2010-08-25 14:49 ` Graeme Russ 2010-08-25 15:40 ` sk ong @ 2010-08-27 10:07 ` Detlev Zundel 1 sibling, 0 replies; 8+ messages in thread From: Detlev Zundel @ 2010-08-27 10:07 UTC (permalink / raw) To: u-boot Hi Graeme, > I notice there is a work in progress for coreboot on only one Intel Atom > motherboard (the D945GCLF) - Not a lot of support there. I would imagine that it is _very_ difficult to get at the information one needs to write such low-level software. During private talks I got the impression that the manuals/data sheets which actually put you into a position to use a hardware component from software is not what the Intel Atom is known for - if you know what I mean. Cheers Detlev -- .. most of the finest products of an applied mathematician's fancy must be rejected, as soon as they have been created, for the brutal but sufficient reason that they do not fit the facts. -- G. H. Hardy -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-40 Fax: (+49)-8142-66989-80 Email: dzu at denx.de ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2010-08-27 10:07 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-08-20 7:42 [U-Boot] u-boot-x86 sk ong
2010-08-21 1:42 ` Graeme Russ
2010-08-22 19:21 ` Robert Schwebel
[not found] ` <AANLkTinh8372fgDaiukRz-+CBS5jYfSNTjvdbK9gR8Xj@mail.gmail.com>
2010-08-25 11:16 ` Graeme Russ
2010-08-25 13:49 ` sk ong
2010-08-25 14:49 ` Graeme Russ
2010-08-25 15:40 ` sk ong
2010-08-27 10:07 ` Detlev Zundel
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox