coreboot November 2004

coreboot@coreboot.org

31 participants
183 discussions

status information
by Stefan Reinauer 02 Nov '04

02 Nov '04

Hi, the "reference build" arima hdama seems to be too big when being compiled with gcc 3.3.1 or 3.3.3.. See attached log file. Please, if anyone has time and test capabilities, have an eye on the non working ports and try to submit fixes. Stefan

7 37

RE: Weird K8 output
by YhLu 02 Nov '04

02 Nov '04

Can you let us know the incoherent device in CPU 0....CPU3..? Some device unit id can not be changed? For examples, you set it from 0 to 1, it will still be 0. So it hang there.... Regards YH -----Original Message----- From: YhLu Sent: Tuesday, November 02, 2004 4:07 PM To: 'Stefan Reinauer' Cc: linuxbios(a)clustermatic.org Subject: RE: Weird K8 output Yes, You are right, You can change to // upos = ht_c[i].upos; upos = ((reg & 0xf00)>>8) * 0x20 + 0x80; last time, I remember because it will use up regs. But this time seems OK. regards Yinghai Lu -----Original Message----- From: Stefan Reinauer [mailto:stepan@openbios.org] Sent: Tuesday, November 02, 2004 3:31 PM To: YhLu Subject: Re: Weird K8 output * YhLu <YhLu(a)tyan.com> [041102 21:27]: > MB Model? > HT routing table? It was too loud (LOGLEVEL 9) BUT: > coherent_ht_finalize > done > PCI_DEV=(0,0x18,0) 0xb4=00030000 > bus=00 id =74501022 > bus=00 id =74601022 > PCI_DEV=(0,0x19,0) 0xd4=000a0400 <HANGS HERE> static int ht_setup_chains(const struct ht_chain *ht_c, int ht_c_num) takes: [...] { /* Link 1 of CPU0 */ .udev = PCI_DEV(0, 0x18, 0), .upos = 0xa0, .devreg = 0xe0, }, { /* Link 2 of CPU1 */ .udev = PCI_DEV(0, 0x19, 0), .upos = 0xc0, .devreg = 0xe4, }, [...] and resourcemap.c: PCI_ADDR(0, 0x18, 1, 0xE0), 0x0000FC88, 0x03000103, PCI_ADDR(0, 0x18, 1, 0xE4), 0x0000FC88, 0x0a040213, PCI_ADDR(0, 0x18, 1, 0xE8), 0x0000FC88, 0x0f0b0223, PCI_ADDR(0, 0x18, 1, 0xEC), 0x0000FC88, 0x14100133, Question1: I took the bus numbers above from the running machine. is this wrong? How would I calculate them correctly? Question2: isn't .upos duplicate? ht_setup_chains reads the link to choose from .devreg. It could determine .upos from there as well, no? Stefan

1 0

RE: Weird K8 output
by YhLu 02 Nov '04

02 Nov '04

Only two times Once in crt0 One in LinuxBIOS_ram. Device. Besides updating unitid, device scan need to update bus number too. Regards Yinghai Lu -----Original Message----- From: Stefan Reinauer [mailto:stepan@openbios.org] Sent: Tuesday, November 02, 2004 4:07 PM To: YhLu Cc: linuxbios(a)clustermatic.org Subject: Re: Weird K8 output * Stefan Reinauer <stepan(a)openbios.org> [041103 00:49]: > * YhLu <YhLu(a)tyan.com> [041102 21:27]: > > done > > PCI_DEV=(0,0x18,0) 0xb4=00030000 > > bus=00 id =74501022 > > bus=00 id =74601022 > > PCI_DEV=(0,0x19,0) 0xd4=000a0400 > > Hm.. it dies here while collapsing the previous enumeration. While collapsing 74501022 again... Can anyone shed some light on why the repeated collapse is needed? Stefan

2 1

RE: Weird K8 output
by YhLu 02 Nov '04

02 Nov '04

Yes, You are right, You can change to // upos = ht_c[i].upos; upos = ((reg & 0xf00)>>8) * 0x20 + 0x80; last time, I remember because it will use up regs. But this time seems OK. regards Yinghai Lu -----Original Message----- From: Stefan Reinauer [mailto:stepan@openbios.org] Sent: Tuesday, November 02, 2004 3:31 PM To: YhLu Subject: Re: Weird K8 output * YhLu <YhLu(a)tyan.com> [041102 21:27]: > MB Model? > HT routing table? It was too loud (LOGLEVEL 9) BUT: > coherent_ht_finalize > done > PCI_DEV=(0,0x18,0) 0xb4=00030000 > bus=00 id =74501022 > bus=00 id =74601022 > PCI_DEV=(0,0x19,0) 0xd4=000a0400 <HANGS HERE> static int ht_setup_chains(const struct ht_chain *ht_c, int ht_c_num) takes: [...] { /* Link 1 of CPU0 */ .udev = PCI_DEV(0, 0x18, 0), .upos = 0xa0, .devreg = 0xe0, }, { /* Link 2 of CPU1 */ .udev = PCI_DEV(0, 0x19, 0), .upos = 0xc0, .devreg = 0xe4, }, [...] and resourcemap.c: PCI_ADDR(0, 0x18, 1, 0xE0), 0x0000FC88, 0x03000103, PCI_ADDR(0, 0x18, 1, 0xE4), 0x0000FC88, 0x0a040213, PCI_ADDR(0, 0x18, 1, 0xE8), 0x0000FC88, 0x0f0b0223, PCI_ADDR(0, 0x18, 1, 0xEC), 0x0000FC88, 0x14100133, Question1: I took the bus numbers above from the running machine. is this wrong? How would I calculate them correctly? Question2: isn't .upos duplicate? ht_setup_chains reads the link to choose from .devreg. It could determine .upos from there as well, no? Stefan

1 0

RE: Weird K8 output
by YhLu 02 Nov '04

02 Nov '04

MB Model? HT routing table? YH -----Original Message----- From: Stefan Reinauer [mailto:stepan@openbios.org] Sent: Tuesday, November 02, 2004 12:06 PM To: linuxbios(a)clustermatic.org Subject: Weird K8 output Hi, on a 4cpu K8 system I get the following output before the system hangs.. Is the machine drunk or are CPU1-3 just running away? LinuxBIOS-1.1.7.0-fallback Tue Nov 2 20:34:14 CET 2004 starting... setting up resource map....done. Enabling routing table for node 00 done. Enabling SMP settings setup_remote_node: done Renaming current temporary node to 01 done. anabling routing table for node 01 dConlee. tRiennamgin g mcutrrrenrt emporary node to 02 done. Enabling routing table for node 02 dConle.e mRernaiminngg c urmretntr tre porary node to 03 done. Enabling routing table for node 03 dColnee.a iz04n ngod esm itnirtiral ed. coherent_ht_finalize done PCI_DEV=(0,0x18,0) 0xb4=00030000 bus=00 id =74501022 bus=00 id =74601022 PCI_DEV=(0,0x19,0) 0xd4=000a0400 Stefan _______________________________________________ Linuxbios mailing list Linuxbios(a)clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios

2 2

Re: Segments in pre-C and C part
by ebiederman＠lnxi.com 02 Nov '04

02 Nov '04

Li-Ta Lo <ollie(a)lanl.gov> writes: > Eric, > > Why we use two differenct segment lay out for pre-C and > C part ? For the code that runs from ram and the code that runs directly from the rom chip. > In entry32.inc it is 0x08 for code and 0x10 for data but > in c_start we use 0x10 for code and 0x18 for data. Good question but it is probably a decent thing that will keep up from making too many assumptions that they have the same gdt. I don't believe this is something that has changed recently. Eric

2 2

Re: Interrupt Handling
by ebiederman＠lnxi.com 02 Nov '04

02 Nov '04

Li-Ta Lo <ollie(a)lanl.gov> writes: > Eric, > > Why there are two ways for interrupt handling in c_start.S There are two ways for hardware exception handling. Interrupts are not handled. And that is because on x86 exception handlers don't always give you an error code. > vec9: > pushl $0 /* error code */ > pushl $9 /* vector */ > jmp int_hand > > > vec10: > /* error code */ > pushl $10 /* vector */ > jmp int_hand > .word 0x9090 > > Vector 9 push the error code and vector but vector 10 only push the > vector number. > > BTW, are those handler actually used ? The idt is uninited. This hunk below initializes it. It seems to work out better to initialize the idt with code instead of data. /* Initialize the Interrupt Descriptor table */ leal _idt, %edi leal vec0, %ebx movl $(0x10 << 16), %eax /* cs selector */ 1: movw %bx, %ax movl %ebx, %edx movw $0x8E00, %dx /* Interrupt gate - dpl=0, present */ movl %eax, 0(%edi) movl %edx, 4(%edi) addl $6, %ebx addl $8, %edi cmpl $_idt_end, %edi jne 1b /* Load the Interrupt descriptor table */ lidt idtarg Eric

2 1

Weird K8 output
by Stefan Reinauer 02 Nov '04

02 Nov '04

Hi, on a 4cpu K8 system I get the following output before the system hangs.. Is the machine drunk or are CPU1-3 just running away? LinuxBIOS-1.1.7.0-fallback Tue Nov 2 20:34:14 CET 2004 starting... setting up resource map....done. Enabling routing table for node 00 done. Enabling SMP settings setup_remote_node: done Renaming current temporary node to 01 done. anabling routing table for node 01 dConlee. tRiennamgin g mcutrrrenrt emporary node to 02 done. Enabling routing table for node 02 dConle.e mRernaiminngg c urmretntr tre porary node to 03 done. Enabling routing table for node 03 dColnee.a iz04n ngod esm itnirtiral ed. coherent_ht_finalize done PCI_DEV=(0,0x18,0) 0xb4=00030000 bus=00 id =74501022 bus=00 id =74601022 PCI_DEV=(0,0x19,0) 0xd4=000a0400 Stefan

2 3

Re: Removing .name field
by ebiederman＠lnxi.com 02 Nov '04

02 Nov '04

Li-Ta Lo <ollie(a)lanl.gov> writes: > Eric, > > Since you relaxed the 64kB limit. Why you still removed the .name field? > We really want the field back in for debugging purpose. No one has shown me a use or an example of a use. I have simply heard that the name field is a good thing. Show me how it would be used and I don't have a problem with not removing it. As it simply looks totally like dead code to me. In addition while I have relaxed the 64KiB limit I still want to watch the code size. Eric

1 0

linuxbios external api
by Jeremy Jackson 02 Nov '04

02 Nov '04

Maybe this has been asked before, but i've been thinking about exporting functions like serial console, debugger entry points, etc. the PC BIOS uses the interrupt vectors as, well, vectors. I remember the C=64 had a jump table at a fixed address. I was wondering how hard it would be to use an ELF symbol table as a jump table, or at least as the source of information to build one. -- Jeremy Jackson Coplanar Networks (519)897-1516 http://www.coplanar.net

6 5

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

coreboot November 2004