SeaBIOS December 2011

seabios@seabios.org

25 participants
64 discussions

Re: [SeaBIOS] [PATCH RFC 1/7] Only use -fomit-frame-pointer with 32bit flat code
by Julian Pidancet 18 Dec '11

18 Dec '11

On Sat, Dec 17, 2011 at 3:50 PM, Julian Pidancet <julian.pidancet(a)gmail.com> wrote: > On Sat, Dec 17, 2011 at 3:38 PM, Kevin O'Connor <kevin(a)koconnor.net> wrote: >> On Fri, Dec 16, 2011 at 02:15:07PM +0000, Julian Pidancet wrote: >>> The real-mode emulator that Windows 7 uses to execute the vga bios rom >>> becomes very confused when the code tries to dereference the stack >>> pointer instead of using the frame pointer, leading to corruption of >>> parameters passed on the stack from one function to another. >>> >>> Signed-off-by: Julian Pidancet <julian.pidancet(a)gmail.com> >> >> Wow - I can only imagine how hard it was to find that. >> >> I'm okay with turning off omit-frame-pointer for the vga bios. >> However, I'd prefer not to turn it off for the main seabios build. >> Does the emulator also try to emulate parts of the BIOS, which would >> make that a problem? >> > > Besides the functions which are common to both the BIOS and the VGA > ROM ( src/output.c and src/util.c), it doesn't seem to make any > particular problem with the rest of the BIOS. > The VGA ROM and the BIOS currently share the same CFLAGS. Due to the way the Makefile is written at the moment, I can hardly see how to modify it to compile both binaries with different CFLAGS without making a mess in the Makefile. Do you have any recommendations ? -- Julian

1 0

Re: [SeaBIOS] [PATCH RFC 6/7] Partially implement VBE interface, create VBE driver skeleton
by Kevin O'Connor 17 Dec '11

17 Dec '11

On Fri, Dec 16, 2011 at 02:15:12PM +0000, Julian Pidancet wrote: [...] > +struct VbeInfoBlock > +{ > + u8 signature[4]; > + u16 version; > + u16 oem_string_off; > + u16 oem_string_seg; Looks like this is the same as "struct vesa_info" defined in src/bootsplash.c. It doesn't really matter which struct is used, but both areas of code should share the same definition. Also, I think it would be better to use "struct segoff_s" and to use "u32 signature". -Kevin

2 1

[PATCH RFC 0/7] VGA ROM improvements
by Julian Pidancet 17 Dec '11

17 Dec '11

This patch set contain ground work for a working VBE implementation. It adds several things: * A configuration menu in seabios Kconfig * The ability to generate a PCI ROM * PMM facilities that the ROM can use to allocate memory * A skeleton driver to implement the VBE interface As a bonus, it fixes an issue when real-mode VGA ROM code is emulated by Windows 7, and is unable to cope with code generated with the -fomit-frame-pointer option. It should make pretty easy to implement the BOCHS DISPI interface. Also, I use these patchs to implement VBE for an experimental gfx device emulation. So far I managed to get Windows 7 to boot and show up a 1280x1024 desktop. Julian Pidancet (7): Only use -fomit-frame-pointer with 32bit flat code Rename CONFIG_VBE in CONFIG_VGA_BOCHS and CONFIG_CIRRUS in CONFIG_VGA_CIRRUS Add configuration menu for the VGA ROM buildrom.py: Handle PCI header checksum Add PMM memory allocation facilities for VGA option ROMs Partially implement VBE interface, create VBE driver skeleton Add vbe_flag field in BDA Makefile | 7 +- src/Kconfig | 35 +++++++ src/biosvar.h | 3 +- tools/buildrom.py | 5 + vgasrc/pmm.c | 50 +++++++++ vgasrc/pmm.h | 16 +++ vgasrc/vbe.c | 135 ++++++++++++++++++++++++ vgasrc/vbe.h | 194 +++++++++++++++++++++++++++++++++++ vgasrc/vga.c | 290 ++++++++++++++++++++++++++++++++++++++++++++-------- vgasrc/vgaentry.S | 47 ++++++++- vgasrc/vgatables.h | 6 - 11 files changed, 731 insertions(+), 57 deletions(-) create mode 100644 vgasrc/pmm.c create mode 100644 vgasrc/pmm.h create mode 100644 vgasrc/vbe.c create mode 100644 vgasrc/vbe.h -- Julian Pidancet

4 11

Re: [SeaBIOS] how to boot seabios with a raid controller?
by Rudolf Marek 17 Dec '11

17 Dec '11

Hi, I tried with silicon Image option rom and it was working as expected. Not sure what is your problem here. Maybe the option needs big real mode... Thanks Rudolf

2 1

Re: [SeaBIOS] how to boot seabios with a raid controller?
by Kevin O'Connor 17 Dec '11

17 Dec '11

On Fri, Dec 16, 2011 at 03:39:56PM -0700, Marc Jones wrote: > I'm trying coreboot + seabios on a system with a raid controller. What > is the expected behavior in seabios for booting a sata raid > controller? The controller supports raid and shows up as a raid class > device. The raid option rom is loaded, but doesn't find any raid > arrays. Later, seabios doesn't find the sata drive. I would expect > that seabios or the optionrom (int13?) would still be able to support > a normal sata drive. Thoughts, comments? It should work. Can you post the seabios log, along with any additional info the raid option rom reports on the screen? Are the drives you are looking to boot from in RAID mode (and thus should boot from the raid option rom)? Is this a built-in device or a card? -Kevin

1 0

Re: [SeaBIOS] [PATCH RFC 3/7] Add configuration menu for the VGA ROM
by Kevin O'Connor 17 Dec '11

17 Dec '11

On Sat, Dec 17, 2011 at 04:22:01PM +0000, Julian Pidancet wrote: > On Sat, Dec 17, 2011 at 4:07 PM, Kevin O'Connor <kevin(a)koconnor.net> wrote: > > struct pci_data rom_pci_data VAR16VISIBLE = { > > .vendor = CONFIG_VGA_VID, > > .device = CONFIG_VGA_DID, > > .dlen = 0x18, > > .class_hi = 0x300, > > ... > > }; > > > > The PCI firmware specification only requires the PCI header to be > located in the first 64KB of the option ROM image, I don't know if > there's a way we can enforce that if it is declared in C. If the option rom is over 64K, then things will break anyway as the code is all 16bit and there are no provisions for far jumping. -Kevin

1 0

Re: [SeaBIOS] [patch] Skip the bootmenu if only one device is found
by Kevin O'Connor 17 Dec '11

17 Dec '11

On Fri, Dec 16, 2011 at 03:29:17PM -0700, Marc Jones wrote: > The bootmenu is only useful if there are multiple devices. We can > speed up theboot by a few seconds if there is only a single device to > try to boot from. > Signed-off-by: Marc Jones <marcj303(a)gmail.com> As Sebastian points out, this came up once before. I fear altering the timeout dynamically will cause confusion. It's possible to tune the timeout via cbfs, and CONFIG_THREAD_OPTIONROMS can also "speed up" the timeout by allowing hardware init to run in parallel with the delay. -Kevin

1 0

Re: [SeaBIOS] [PATCH RFC 3/7] Add configuration menu for the VGA ROM
by Kevin O'Connor 17 Dec '11

17 Dec '11

On Fri, Dec 16, 2011 at 02:15:09PM +0000, Julian Pidancet wrote: > This patch adds a configuration menu for the VGA ROM, it also allow > the creation of a PCI header so the ROM can be extracted from a PCI > device. [...] > +/**************************************************************** > + * PCI Data > + ****************************************************************/ > +#ifdef CONFIG_PCIBIOS > +_rom_pci_data: > + .ascii "PCIR" > +#if defined(CONFIG_VGA_VID) && defined(CONFIG_VGA_DID) > + .word CONFIG_VGA_VID /* VendorID */ > + .word CONFIG_VGA_DID /* DeviceID */ > +#elif defined(CONFIG_VGA_CIRRUS) > + .word 0x1013 > + .word 0x00b8 > +#elif defined(CONFIG_VGA_BOCHS) > + .word 0x1234 > + .word 0x1111 One other note on this - can't _rom_pci_data be declared in the C code as: struct pci_data rom_pci_data VAR16VISIBLE = { .vendor = CONFIG_VGA_VID, .device = CONFIG_VGA_DID, .dlen = 0x18, .class_hi = 0x300, ... }; -Kevin

2 1

Re: [SeaBIOS] [PATCH RFC 6/7] Partially implement VBE interface, create VBE driver skeleton
by Kevin O'Connor 17 Dec '11

17 Dec '11

On Fri, Dec 16, 2011 at 02:15:12PM +0000, Julian Pidancet wrote: [...] > +struct CRTCInfoBlock > +{ > + u16 horiz_total; > + u16 horiz_sync_start; > + u16 horiz_sync_end; > + u16 vert_total; > + u16 vert_sync_start; > + u16 vert_sync_end; > + u8 flags; > + u32 pixclock; Won't all these structs need to be declared as PACKED? -Kevin

2 1

Re: [SeaBIOS] [PATCH RFC 1/7] Only use -fomit-frame-pointer with 32bit flat code
by Kevin O'Connor 17 Dec '11

17 Dec '11

On Fri, Dec 16, 2011 at 02:15:07PM +0000, Julian Pidancet wrote: > The real-mode emulator that Windows 7 uses to execute the vga bios rom > becomes very confused when the code tries to dereference the stack > pointer instead of using the frame pointer, leading to corruption of > parameters passed on the stack from one function to another. > > Signed-off-by: Julian Pidancet <julian.pidancet(a)gmail.com> Wow - I can only imagine how hard it was to find that. I'm okay with turning off omit-frame-pointer for the vga bios. However, I'd prefer not to turn it off for the main seabios build. Does the emulator also try to emulate parts of the BIOS, which would make that a problem? -Kevin

2 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

SeaBIOS December 2011