v3 still has many places where config variables are tested with #if in
v2 style. That fails silently in v3. Fix up the most obvious ones.
Signed-off-by: Carl-Daniel Hailfinger [EMAIL PROTECTED]
Index: LinuxBIOSv3-if0crud-2/include/spinlock.h
On Wed, Feb 13, 2008 at 4:44 PM, Peter Stuge [EMAIL PROTECTED] wrote:
On Wed, Feb 13, 2008 at 12:33:48PM -0800, ron minnich wrote:
The ELF parser in v3 ought to be removed.
Do you mean elfboot?
yes indeed.
One big reason is that you don't want to find out, after you have
burned an elf
I'll comment on a few.
Carl-Daniel Hailfinger wrote:
The following config variables are tested in the code, but don't exist
in Kconfig:
CONFIG_ADDR
This is actually defined but never used in v2 or v3. Remove it.
CONFIG_AGP_PLUGIN_SUPPORT
CONFIG_BIG_ENDIAN
CONFIG_CARDBUS_PLUGIN_SUPPORT
Acked-by: Ronald G. Minnich [EMAIL PROTECTED]
On Wed, Feb 13, 2008 at 3:03 PM, Carl-Daniel Hailfinger
[EMAIL PROTECTED] wrote:
Kill dead code inside #if 0.
The code was blindly copied from v2 in v3:r69. Was never enabled in v2
(committed as dead code by Eric Biederman in v2:r1664 with log
Author: hailfinger
Date: 2008-02-14 17:59:37 +0100 (Thu, 14 Feb 2008)
New Revision: 597
Modified:
coreboot-v3/device/pci_device.c
Log:
Kill dead code inside #if 0.
The code was blindly copied from v2 in v3:r69. Was never enabled in v2
(committed as dead code by Eric Biederman in v2:r1664
Author: hailfinger
Date: 2008-02-14 17:59:37 +0100 (Thu, 14 Feb 2008)
New Revision: 597
Modified:
coreboot-v3/device/pci_device.c
Log:
Kill dead code inside #if 0.
The code was blindly copied from v2 in v3:r69. Was never enabled in v2
(committed as dead code by Eric Biederman in v2:r1664
On 14.02.2008 17:55, ron minnich wrote:
Acked-by: Ronald G. Minnich [EMAIL PROTECTED]
Thanks, r597.
Regards,
Carl-Daniel
--
coreboot mailing list
coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot
This patch adds zero compression for bss segments. One of the reasons
for this is that currently, if you select no compression, the bss
segment of filo takes up 153K with just zeroes. With this patch, it
always takes up a lar header + 1 byte. I left the one byte so that
the checksum wouldn't be
Ping.
I've since realized that since lar doesn't understand pre-compressed
payloads, this has to be done for working compressed payloads.
Myles
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Myles
Watson
Sent: Tuesday, February 12, 2008 2:27 PM
The whole #ifdef vs. #if is a very old argument in linuxbios/coreboot.
initially, it was #ifdef in old v1. Then, later, there was a feeling
it should be linux style, #if.
The change from #ifdef to #if caused lots of silent failures. Now
we're changing back.
The main thing is that we have to be
I like it.
Acked-by: Ronald G. Minnich [EMAIL PROTECTED]
--
coreboot mailing list
coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot
Carl-Daniel Hailfinger wrote:
On 14.02.2008 19:15, Myles Watson wrote:
This patch removes elfboot and references to ELF from v3. Since lar
parses ELF, there's no need to have v3 parse ELF. The ELF files that
v3 could parse should always have been a subset of the ones that lar
could parse
per stefan's comment, let's see how much space we save. This is
bootblock space, which could be significant, but can we see a number?
thanks
ron
--
coreboot mailing list
coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot
Carl-Daniel Hailfinger wrote:
v3 still has many places where config variables are tested with #if in
v2 style. That fails silently in v3. Fix up the most obvious ones.
Signed-off-by: Carl-Daniel Hailfinger [EMAIL PROTECTED]
While the patch itself is discussable, it is cosmetics and not a
@ in this case seperates the type of resources (pci) from the instance
of it (device, function).
So @ is a seperator. Using _ will add ambiguity as it is NOT a
seperator. same problem for -.
Stepan, will the OFW guys kill us if we allow ':' as well as @.
The general syntax of a pathname
Carl-Daniel Hailfinger wrote:
Remove superfluous #if 1. The code has been enabled since it was
committed, the #if 1 never served any purpose.
Signed-off-by: Carl-Daniel Hailfinger [EMAIL PROTECTED]
Acked-by: Stefan Reinauer [EMAIL PROTECTED]
Index: LinuxBIOSv3-if0crud/device/pci_device.c
Stefan, one thing I don't understand, why does removing elfboot make
it so we can't unpack LAR?
ron
--
coreboot mailing list
coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot
ron minnich wrote:
per stefan's comment, let's see how much space we save. This is
bootblock space, which could be significant, but can we see a number?
Elfboot still uses bootblock space if it is disabled in Kconfig? That's
definitely something we need to fix. That space is tight.
--
On Thu, Feb 14, 2008 at 11:00 AM, Stefan Reinauer [EMAIL PROTECTED] wrote:
You unpack it, repack it, and the segments become ordinary files with
their lar headers containing bogus information.
excellent point.
So either we recreate an ELF file on unpack, or we dump the metadata to
a
On 14.02.2008 19:45, Stefan Reinauer wrote:
Carl-Daniel Hailfinger wrote:
Remove superfluous #if 1. The code has been enabled since it was
committed, the #if 1 never served any purpose.
Signed-off-by: Carl-Daniel Hailfinger
[EMAIL PROTECTED]
Acked-by: Stefan Reinauer [EMAIL PROTECTED]
Carl-Daniel Hailfinger wrote:
On 14.02.2008 19:31, Stefan Reinauer wrote:
Carl-Daniel Hailfinger wrote:
v3 still has many places where config variables are tested with #if in
v2 style. That fails silently in v3. Fix up the most obvious ones.
Signed-off-by: Carl-Daniel Hailfinger
Segher Boessenkool wrote:
@ in this case seperates the type of resources (pci) from the instance
of it (device, function).
So @ is a seperator. Using _ will add ambiguity as it is NOT a
seperator. same problem for -.
Stepan, will the OFW guys kill us if we allow ':' as well as @.
We are
Peter Stuge wrote:
I have issues with parts of this.
On Wed, Feb 13, 2008 at 10:00:21PM +0100, [EMAIL PROTECTED] wrote:
Minclude/device/path.h
Add LPC path type, replacing SUPERIO path type, since SUPERIO is
only one type of LPC. Clean up tabbing in parts of the file
(cosmetic).
We are of course free to go away from the ePAPR flat device tree (we
did already to some extent).
I am not talking about the ePAPR flat tree. I am talking about the
flat tree as used in PowerPC Linux for many years now.
I agree with Segher we should stay close to the original. That's why
Carl-Daniel Hailfinger wrote:
Index: southbridge/amd/cs5536/dts
===
--- southbridge/amd/cs5536/dts (Revision 590)
+++ southbridge/amd/cs5536/dts (Arbeitskopie)
@@ -57,4 +57,7 @@
com2_enable = 0;
com2_address =
Stefan Reinauer wrote:
Carl-Daniel Hailfinger wrote:
Index: southbridge/amd/cs5536/dts
===
--- southbridge/amd/cs5536/dts(Revision 590)
+++ southbridge/amd/cs5536/dts(Arbeitskopie)
@@ -57,4 +57,7 @@
com2_enable =
On Thu, Feb 14, 2008 at 2:16 PM, Marc Jones [EMAIL PROTECTED] wrote:
Stefan Reinauer wrote:
Should a virtual device not be disabled in the same way as a physical
one?
[EMAIL PROTECTED],0 {
disabled;
}
?
Stefan
This is an interesting idea but the devices are
* Marc Jones [EMAIL PROTECTED] [080214 23:16]:
+++ southbridge/amd/cs5536/dts(Arbeitskopie)
@@ -57,4 +57,7 @@
com2_enable = 0;
com2_address = 0x2f8;
com2_irq = 3;
+
+/* Disable unwanted virtualized PCI devices */
+unwanted_vpci[0] = 0;
};
I am blissfully
* ron minnich [EMAIL PROTECTED] [080214 23:18]:
[EMAIL PROTECTED],0 {
disabled;
}
This is an interesting idea but the devices are CS5536 specific and I
don't think we want to make dts CS5536 aware. I think it better use the
device ID rather than the device location to do
On 14.02.2008 23:25, Stefan Reinauer wrote:
* ron minnich [EMAIL PROTECTED] [080214 23:18]:
[EMAIL PROTECTED],0 {
disabled;
}
This is an interesting idea but the devices are CS5536 specific and I
don't think we want to make dts CS5536 aware. I think it better use the
On Thu, Feb 14, 2008 at 2:28 PM, Carl-Daniel Hailfinger
Hm. It's not that hard to create a MANIFEST file. And the patch I sent
twice (which was NACKed twice) to specify LAR header contents on the LAR
command line would work perfectly together with the MANIFEST. No
information lost. I can
On Thu, Feb 14, 2008 at 2:32 PM, Carl-Daniel Hailfinger
Indeed. That would solve the problem very nicely without putting hacks
into the dts.
OK, so we put nodes under the 5536, and specify 'disabled', and the
5536 code knows to actually call VSA when it sees that they are
disabled. This is
Author: hailfinger
Date: 2008-02-14 23:34:40 +0100 (Thu, 14 Feb 2008)
New Revision: 599
Modified:
coreboot-v3/arch/x86/mc146818rtc.c
coreboot-v3/arch/x86/speaker.c
coreboot-v3/device/device.c
coreboot-v3/device/pci_device.c
coreboot-v3/device/pci_rom.c
Author: hailfinger
Date: 2008-02-14 23:34:40 +0100 (Thu, 14 Feb 2008)
New Revision: 599
Modified:
coreboot-v3/arch/x86/mc146818rtc.c
coreboot-v3/arch/x86/speaker.c
coreboot-v3/device/device.c
coreboot-v3/device/pci_device.c
coreboot-v3/device/pci_rom.c
On 14.02.2008 22:23, Stefan Reinauer wrote:
Carl-Daniel Hailfinger wrote:
Remove superfluous checks for boolean CONFIG_* variables where we tested
CONFIG_* == 1. If those variables are set, they will always be 1.
Acked-by: Stefan Reinauer [EMAIL PROTECTED]
Thanks, r599.
Regards,
Thanks for the notes. We did in the end resolve in favor of '@'.
ron
On Thu, Feb 14, 2008 at 2:01 PM, Segher Boessenkool
[EMAIL PROTECTED] wrote:
We are of course free to go away from the ePAPR flat device tree (we
did already to some extent).
I am not talking about the ePAPR flat tree.
On Thu, Feb 14, 2008 at 1:10 PM, Stefan Reinauer [EMAIL PROTECTED] wrote:
Peter Stuge wrote:
You are right. I think it should stay SuperIO.
I'll try to put it back tomorrow.
{
ramsize = 128;
constructor = i440bx_constructors;
+domainid = 0x8086, 0x7190;
};
On 14.02.2008 23:32, ron minnich wrote:
On Thu, Feb 14, 2008 at 2:32 PM, Carl-Daniel Hailfinger
Indeed. That would solve the problem very nicely without putting hacks
into the dts.
OK, so we put nodes under the 5536, and specify 'disabled', and the
5536 code knows to actually
Acked-by: Ronald G. Minnich [EMAIL PROTECTED]
On Thu, Feb 14, 2008 at 2:55 PM, Carl-Daniel Hailfinger
[EMAIL PROTECTED] wrote:
On 14.02.2008 23:32, ron minnich wrote:
On Thu, Feb 14, 2008 at 2:32 PM, Carl-Daniel Hailfinger
Indeed. That would solve the problem very nicely without
Author: uwe
Date: 2008-02-15 00:22:20 +0100 (Fri, 15 Feb 2008)
New Revision: 3102
Modified:
trunk/util/flashrom/Makefile
Log:
With this small change it is possible to build flashrom again when
specifying custom CFLAGS/LDFLAGS from the make command line like:
make CFLAGS=... LDFLAGS=...
I
Author: uwe
Date: 2008-02-15 00:22:20 +0100 (Fri, 15 Feb 2008)
New Revision: 3102
Modified:
trunk/util/flashrom/Makefile
Log:
With this small change it is possible to build flashrom again when
specifying custom CFLAGS/LDFLAGS from the make command line like:
make CFLAGS=... LDFLAGS=...
I
Hi,
On Wed, Feb 13, 2008 at 03:15:09PM -0500, Clark Rawlins wrote:
Is this attempt any better?
IMHO yes, committed in r3102. It allows to override both CFLAGS and
LDFLAGS now (tested).
Signed-of-by: Clark Rawlins [EMAIL PROTECTED]
That's '-off-' not '-of-' btw.
Cheers, Uwe.
--
On Mon, Feb 11, 2008 at 05:48:38PM -0700, Marc Jones wrote:
Acked-by: Marc Jones [EMAIL PROTECTED]
Committed in r3.
Uwe.
--
http://www.hermann-uwe.de | http://www.holsham-traders.de
http://www.crazy-hacks.org | http://www.unmaintained-free-software.org
--
coreboot mailing list
ron minnich wrote:
On Thu, Feb 14, 2008 at 2:32 PM, Carl-Daniel Hailfinger
Indeed. That would solve the problem very nicely without putting hacks
into the dts.
OK, so we put nodes under the 5536, and specify 'disabled', and the
5536 code knows to actually call VSA when it sees that they
On Tue, Feb 12, 2008 at 08:15:46PM +0100, Peter Stuge wrote:
Signed-off by: Chris Kilgour [EMAIL PROTECTED]
'Signed-off-by' please. The commit hook is not in place yet, which is why
this didn't abort the commit with a warning, but we should still try to
be consistent.
Acked-by: Peter Stuge
On Thu, Feb 14, 2008 at 6:35 PM, Tom Sylla [EMAIL PROTECTED] wrote:
ron minnich wrote:
On Thu, Feb 14, 2008 at 2:32 PM, Carl-Daniel Hailfinger
Indeed. That would solve the problem very nicely without putting hacks
into the dts.
OK, so we put nodes under the 5536, and specify
Try this, simplifies the dts, removes redundant ids.
might even allow us to totally yank the pciid, domainid, etc. keywords
ron
This is a proposal in response to Peter and Stefan's comments.
I think they are correct.
Basically, this patch gets rid of the constructor arrays in the
various .c
On Thu, Feb 14, 2008 at 02:30:48PM -0800, ron minnich wrote:
It's not that hard to create a MANIFEST file.
I still think it's far simpler to have an option to extract LAR
files as LAR files.
I would prefer ELF because it's more standard - if it can do all we
need.
//Peter
--
coreboot
On Thu, Feb 14, 2008 at 4:23 PM, Peter Stuge [EMAIL PROTECTED] wrote:
I would prefer ELF because it's more standard - if it can do all we
need.
All right, so we'll have LAR extract to elf. I don't think it's that hard.
ron
--
coreboot mailing list
coreboot@coreboot.org
On Thu, Feb 14, 2008 at 03:16:24PM -0700, Marc Jones wrote:
This is an interesting idea but the devices are CS5536 specific and
I don't think we want to make dts CS5536 aware.
I would like nothing more than to make coreboot understand the Geode
architecture natively. Maybe a goal for v4, once
On Thu, Feb 14, 2008 at 4:39 PM, Peter Stuge [EMAIL PROTECTED] wrote:
Why were there arrays in the first place? Was that introduces for the
LX which has multiple devices? This new approach is definately the
right thing then!
The arrays were introduced by me over a year ago and were
Ron: I'm stealing your ack although the patch has slightly changed
function naming.
Corey/Tom: Does this name match what you had in mind?
Factor out Geode LX VPCI device disabling into a separate function which
consumes one device at a time. This helps avoid array handling in the
dts and allows
On Thu, Feb 14, 2008 at 8:19 PM, Carl-Daniel Hailfinger
[EMAIL PROTECTED] wrote:
Ron: I'm stealing your ack although the patch has slightly changed
function naming.
Corey/Tom: Does this name match what you had in mind?
Factor out Geode LX VPCI device disabling into a separate function which
On 15.02.2008 01:25, ron minnich wrote:
Try this, simplifies the dts, removes redundant ids.
might even allow us to totally yank the pciid, domainid, etc. keywords
I like it.
Index: northbridge/amd/geodelx/geodelx.c
===
On 15.02.2008 01:26, Peter Stuge wrote:
On Thu, Feb 14, 2008 at 10:23:56PM +0100, Stefan Reinauer wrote:
I stand corrected. However, the test for the #define being 1 is
superfluous if the #define is a bool.
Acked-by: Stefan Reinauer [EMAIL PROTECTED]
I'd like to nak. But
On Thu, Feb 14, 2008 at 5:31 PM, Carl-Daniel Hailfinger
Two comments about the style while it is not yet carved in stone.
- Can you prefix every constructor with struct constructor?
sure.
- Can we possibly use anonymous unions to get rid of that obnoxious .u?
(separate patch)
seperately.
On 14.02.2008 17:38, Marc Jones wrote:
I'll comment on a few.
Thanks.
Carl-Daniel Hailfinger wrote:
The following config variables are tested in the code, but don't exist
in Kconfig:
CONFIG_ADDR
This is actually defined but never used in v2 or v3. Remove it.
OK, here is try 2, with changes per your comments.
qemu boots fine. I appreciate all the comments!
I think we're close.
Thanks
ron
This is not signed off yet, but is close.
It also boots qemu just fine, which is a good sign.
Note that from now on, to pull a constructor into the coreboot
On Thu, Feb 14, 2008 at 05:34:04PM -0800, ron minnich wrote:
Note one thing the arrays in the individual files got you. You
could drag along ALL types of a (e.g.) the intel southbridge even
when you only specified one of them in a dts.
*nods* Hopefully we can be smart about this also in the
On 12.02.2008 22:27, Myles Watson wrote:
This patch uses lar to compress a payload, instead of pre-compressing
it. Since we are already using lar to add the payload, this
simplifies the build process. It also allows us to compress filo,
which didn't work with v2.
Myles
Signed-off-by:
On Thu, Feb 14, 2008 at 05:48:01PM -0800, ron minnich wrote:
This is not signed off yet, but is close.
It also boots qemu just fine, which is a good sign.
Great!
Note that from now on, to pull a constructor into the coreboot
image and make it available, some dts somewhere has to name
On 15.02.2008 02:48, ron minnich wrote:
OK, here is try 2, with changes per your comments.
qemu boots fine. I appreciate all the comments!
I think we're close.
southbridge/intel/i82371eb/ide and
northbridge/intel/i440bxemulation/domain do not exist in my local
checkout. No idea how they
On 09.02.2008 08:22, ron minnich wrote:
changes to support PIRQ tables. This now builds,runs, loads FILO, and
loads linux.
But interrupts are still not set right. That's for tomorrow.
What's the status of this patch? Superseded?
Regards,
Carl-Daniel
--
http://www.hailfinger.org/
--
I am not sure, I will have to look.
possibly superseded.
ron
--
coreboot mailing list
coreboot@coreboot.org
http://www.coreboot.org/mailman/listinfo/coreboot
On Thu, Feb 14, 2008 at 6:06 PM, Carl-Daniel Hailfinger
[EMAIL PROTECTED] wrote:
southbridge/intel/i82371eb/ide and
northbridge/intel/i440bxemulation/domain do not exist in my local
checkout. No idea how they appear to be checked in in your copy of the tree.
Can you please check whether
Author: kilgour
Date: 2008-02-15 05:44:00 +0100 (Fri, 15 Feb 2008)
New Revision: 6
Modified:
trunk/openvsa/sysmgr/vsainit.lnk
Log:
[openvsa]
ensure all data section symbols are linked into vsainit.bin
Signed-off-by: Chris Kilgour [EMAIL PROTECTED]
Acked-by: Marc Jones [EMAIL PROTECTED]
Marc Jones wrote:
Chris Kilgour wrote:
[openvsa]
ensure all data section symbols are linked into vsainit.bin
Signed-off-by: Chris Kilgour [EMAIL PROTECTED]
Acked-by: Marc Jones [EMAIL PROTECTED]
committed to r6.
--
coreboot mailing list
coreboot@coreboot.org
On Thu, Feb 14, 2008 at 07:12:54PM +0100, Segher Boessenkool wrote:
The general syntax of a pathname component in OF is:
[EMAIL PROTECTED]:args
Thanks for the insight!
unit-addr is (the text representation of) the address for the
node, in the parent's address space.
This statement
On 15.02.2008 02:48, ron minnich wrote:
OK, here is try 2, with changes per your comments.
qemu boots fine. I appreciate all the comments!
I think we're close.
Indeed. A few more comments to make this the perfect patch ;-)
This is not signed off yet, but is close.
It also boots qemu
69 matches
Mail list logo