Now I have a FireWire disk to play with. Adding an interrupt description vastly improves its performance under LinuxBIOS ;-) Potential ACKers please verify 01:0a.0 is not assigned any APIC pin so far. Comment and sign-off in the attachment. This file is quite messy.
Torsten
On Monday 17 December 2007, I wrote:
Potential ACKers please verify 01:0a.0 is not assigned any APIC pin so far.
Could someone please confirm that in src/mainboard/gigabyte/m57sli/mptable.c 01:0a.0 has fallen between the cracks? Anyone can do that.
Hardware owners could even verify it works. V1.x boards should for sure, v2 boards might.
Thanks in advance.
Torsten
Am Donnerstag, 20. Dezember 2007 11:36:36 schrieb Torsten Duwe:
On Monday 17 December 2007, I wrote:
Potential ACKers please verify 01:0a.0 is not assigned any APIC pin so far.
Could someone please confirm that in src/mainboard/gigabyte/m57sli/mptable.c 01:0a.0 has fallen between the cracks? Anyone can do that.
Hardware owners could even verify it works. V1.x boards should for sure, v2 boards might.
if i had a firefire device, i would verify if it works on v2 boards. maybe i can led some firewire hardware from friends. (but at the moment i've none available.)
Thanks in advance.
tanks for investigating that problem, harald
Torsten
Hi Torsten,
sorry, but I do not have a firewire device either :-(
Thanks
Andi
On Thu, Dec 20, 2007 at 02:19:23PM +0100, Harald Gutmann wrote:
Am Donnerstag, 20. Dezember 2007 11:36:36 schrieb Torsten Duwe:
On Monday 17 December 2007, I wrote:
Potential ACKers please verify 01:0a.0 is not assigned any APIC pin so far.
Could someone please confirm that in src/mainboard/gigabyte/m57sli/mptable.c 01:0a.0 has fallen between the cracks? Anyone can do that.
Hardware owners could even verify it works. V1.x boards should for sure, v2 boards might.
if i had a firefire device, i would verify if it works on v2 boards. maybe i can led some firewire hardware from friends. (but at the moment i've none available.)
Thanks in advance.
tanks for investigating that problem, harald
Torsten
Hi Torsten, why not add the comments and readable indentation right now? once it is committed, it won't happen :-)
But, anyway, it's ok.
Acked-by: Ronald G. Minnich rminnich@gmail.com
On Thursday 20 December 2007, Ron Minnich wrote:
Hi Torsten, why not add the comments and readable indentation right now? once it is committed, it won't happen :-)
I wanted to keep semantical change and readability improvements separate. Maybe I should have improved readablity first :) I'll follow up with a semi-trivial patch.
Torsten
On Thursday 20 December 2007, ron minnich wrote:
Hi Torsten, why not add the comments and readable indentation right now? once it is committed, it won't happen :-)
But, anyway, it's ok.
Acked-by: Ronald G. Minnich rminnich@gmail.com
Thanks. Revision 3023.
Torsten
I still owe you a beautification patch.
On Thursday 20 December 2007, ron minnich wrote:
Hi Torsten, why not add the comments and readable indentation right now? once it is committed, it won't happen :-)
It's not indentation. I found the extra long lines and useless comments extremely ugly and obfuscating.
This patch would again qualify as trivial by Russ' definition, but coding style matters so I want to bring this to discussion. IMO this is what preprocessor macros were invented for. We now have bus,dev,fn tuples together, and can easily see that 1:0a.0 maps to "pin" 18. Removing all that redundant blurb also makes room for meaningful comments 8-)
Signed-off-by: Torsten Duwe duwe@lst.de
On 06.01.2008 15:01, Torsten Duwe wrote:
I still owe you a beautification patch.
On Thursday 20 December 2007, ron minnich wrote:
Hi Torsten, why not add the comments and readable indentation right now? once it is committed, it won't happen :-)
It's not indentation. I found the extra long lines and useless comments extremely ugly and obfuscating.
This patch would again qualify as trivial by Russ' definition, but coding style matters so I want to bring this to discussion. IMO this is what preprocessor macros were invented for. We now have bus,dev,fn tuples together, and can easily see that 1:0a.0 maps to "pin" 18. Removing all that redundant blurb also makes room for meaningful comments 8-)
Signed-off-by: Torsten Duwe duwe@lst.de
This is a much needed simplification and readability improvement. Thanks! Generated code seems to be unchanged.
If you drop the #undef lines (or explain why we absolutely need them), you can take the ack from below.
Acked-by: Carl-Daniel Hailfinger c-d.hailfinger.devel.2006@gmx.net
Bonus points if you manage to convert all mptable.c files for MCP55 (or even, if that task is not too daring, all mptable.c files). MCP55 mptable.c would be: src/mainboard/msi/ms7260/mptable.c src/mainboard/msi/ms9282/mptable.c src/mainboard/tyan/s2912/mptable.c src/mainboard/supermicro/h8dmr/mptable.c src/mainboard/nvidia/l1_2pvv/mptable.c src/mainboard/gigabyte/m57sli/mptable.c
Regards, Carl-Daniel
On Sunday 06 January 2008, Carl-Daniel Hailfinger wrote:
This is a much needed simplification and readability improvement. Thanks! Generated code seems to be unchanged.
If you drop the #undef lines (or explain why we absolutely need them), you can take the ack from below.
I consider it good practise to undef such readability macros as soon as they are not needed any more, like freeing unused memory. But I really have no strong opinion about them, no problem.
Acked-by: Carl-Daniel Hailfinger c-d.hailfinger.devel.2006@gmx.net
Thanks, revision 3035.
Bonus points if you manage to convert all mptable.c files for MCP55 (or even, if that task is not too daring, all mptable.c files).
You mean something like perl -pe s:smp_write_intsrc(mc,\s*mp_INT,\s*MP_IRQ[^,]*,\s*bus_isa\s*,(([^,]*),[^,]*, ([^)]*)):ISA_INT($2, $3):g; and so on? Turn it into a full-blown script and run it per board? Maybe.
Torsten
On 07.01.2008 12:21, Torsten Duwe wrote:
On Sunday 06 January 2008, Carl-Daniel Hailfinger wrote:
This is a much needed simplification and readability improvement. Thanks! Generated code seems to be unchanged.
If you drop the #undef lines (or explain why we absolutely need them), you can take the ack from below.
I consider it good practise to undef such readability macros as soon as they are not needed any more, like freeing unused memory. But I really have no strong opinion about them, no problem.
OK, that is a good reason.
Acked-by: Carl-Daniel Hailfinger c-d.hailfinger.devel.2006@gmx.net
Thanks, revision 3035.
Bonus points if you manage to convert all mptable.c files for MCP55 (or even, if that task is not too daring, all mptable.c files).
You mean something like perl -pe s:smp_write_intsrc(mc,\s*mp_INT,\s*MP_IRQ[^,]*,\s*bus_isa\s*,(([^,]*),[^,]*, ([^)]*)):ISA_INT($2, $3):g; and so on? Turn it into a full-blown script and run it per board? Maybe.
Yes, that would be cool. However, the needed macros even differ for different mcp55 boards, so scripted replacement may need a human to look over the results.
Regards, Carl-Daniel
On Fri, Dec 21, 2007 at 06:23:27PM +0100, Torsten Duwe wrote:
On Thursday 20 December 2007, ron minnich wrote:
Hi Torsten, why not add the comments and readable indentation right now? once it is committed, it won't happen :-)
But, anyway, it's ok.
Acked-by: Ronald G. Minnich rminnich@gmail.com
Thanks. Revision 3023.
Firewire works fine on my v1 board. I've just tested a disk.
Thanks! Ward.
-
Andreas B. Mundt -
Carl-Daniel Hailfinger -
Harald Gutmann -
ron minnich -
Torsten Duwe -
Ward Vandewege