From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from planomail.kodiaknetworks.com (planomail.kodiaknetworks.com [76.203.108.6]) by ozlabs.org (Postfix) with ESMTP id 18962DDDFF for ; Wed, 3 Dec 2008 10:40:36 +1100 (EST) MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C954D5.ADF74BEE" Subject: Marvell mv64462 + mpc744x PCI Interrupt Cause Error Handling Date: Tue, 2 Dec 2008 17:28:27 -0600 Message-ID: <295C5089A56CE143B316E5F67CA99CB001F4389F@cowboy.inovate.inovate.com> From: "Stephen Horton" To: List-Id: Linux on Embedded PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , This is a multi-part message in MIME format. ------_=_NextPart_001_01C954D5.ADF74BEE Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi folks, =20 I'm really hoping Mark Greer or Dale Farnsworth sees this email. In my current project, we are in the process of updating a board that has a Marvell mv64462 system controller and a Freescale mpc7447A processor on it. The board was originally developed with Gentoo Linux 2.6.9, but we have ported Linux 2.6.24 kernel onto it. The new kernel is working well thanks to help from people on this list. However, in examining the differences between the 2 kernels, I can see that the 2.6.24 kernel registers no interrupt error handling routines (see PCI Error Report Register Map in the Marvell documentation). The error handlers for cpu, sram, and specifically for the PCI bus that were previously in: /arch/ppc/syslib/mv64460_pic.c are absent from: /arch/powerpc/sysdev/mv64x60_pic.c =20 Can someone who is familiar with the history of the ppc->powerpc migration for this architecture help me figure out why the error handling wasn't ported over? Is there some reason for this, such as it didn't work properly or had some disadvantages? Is this error handling done in 2.6.24 in some more generic manner that I have not yet found? Is this an exercise left for me to implement? =20 Thanks, Stephen ------_=_NextPart_001_01C954D5.ADF74BEE Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable

Hi folks,

I’m really hoping Mark Greer or Dale Farnsworth = sees this email. In my current project, we are in the process of updating a = board that has a Marvell mv64462 system controller and a Freescale mpc7447A = processor on it. The board was originally developed with Gentoo Linux 2.6.9, but = we have ported Linux 2.6.24 kernel onto it. The new kernel is working well = thanks to help from people on this list. However, in examining the differences = between the 2 kernels, I can see that the 2.6.24 kernel registers no interrupt = error handling routines (see PCI Error Report Register Map in the Marvell documentation). The error handlers for cpu, sram, and specifically for = the PCI bus that were previously in:

= /arch/ppc/syslib/mv64460_pic.c

are absent from:

= = /arch/powerpc/sysdev/mv64x60_pic.c

Can someone who is familiar with the history of the ppc->powerpc migration for this architecture help me figure out why = the error handling wasn’t ported over? Is there some reason for this, = such as it didn’t work properly or had some disadvantages? Is this error = handling done in 2.6.24 in some more generic manner that I have not yet found? Is = this an exercise left for me to implement?

Thanks,

Stephen

------_=_NextPart_001_01C954D5.ADF74BEE-- From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.wrs.com (mail.windriver.com [147.11.1.11]) by ozlabs.org (Postfix) with ESMTP id B2840DDDDF for ; Wed, 3 Dec 2008 12:51:04 +1100 (EST) Message-ID: <4935E65B.30305@windriver.com> Date: Wed, 03 Dec 2008 09:52:27 +0800 From: "tiejun.chen" MIME-Version: 1.0 To: Stephen Horton Subject: Re: Marvell mv64462 + mpc744x PCI Interrupt Cause Error Handling References: <295C5089A56CE143B316E5F67CA99CB001F4389F@cowboy.inovate.inovate.com> In-Reply-To: <295C5089A56CE143B316E5F67CA99CB001F4389F@cowboy.inovate.inovate.com> Content-Type: text/plain; charset=UTF-8 Cc: linuxppc-embedded@ozlabs.org List-Id: Linux on Embedded PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Stephen Horton wrote: > Hi folks, > > > > I'm really hoping Mark Greer or Dale Farnsworth sees this email. In my > current project, we are in the process of updating a board that has a > Marvell mv64462 system controller and a Freescale mpc7447A processor on > it. The board was originally developed with Gentoo Linux 2.6.9, but we > have ported Linux 2.6.24 kernel onto it. The new kernel is working well > thanks to help from people on this list. However, in examining the > differences between the 2 kernels, I can see that the 2.6.24 kernel > registers no interrupt error handling routines (see PCI Error Report > Register Map in the Marvell documentation). The error handlers for cpu, > sram, and specifically for the PCI bus that were previously in: > > /arch/ppc/syslib/mv64460_pic.c > > are absent from: > > /arch/powerpc/sysdev/mv64x60_pic.c > > > > Can someone who is familiar with the history of the ppc->powerpc > migration for this architecture help me figure out why the error > handling wasn't ported over? Is there some reason for this, such as it > didn't work properly or had some disadvantages? Is this error handling > done in 2.6.24 in some more generic manner that I have not yet found? Is > this an exercise left for me to implement? > > Untill v2.6.27 you can find corresponding dts content such as cpu-error, sram-ctrl, pci-error and mem-ctrl, in file, prpmc2800.dts. All events are handled on EDAC, error detection and reporting, as an unified mode. So you can track the directory drivers/edac/ for more information. I think these are not upgraded completely for v2.6.24. Best Regards Tiejun > > Thanks, > > Stephen > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > Linuxppc-embedded mailing list > Linuxppc-embedded@ozlabs.org > https://ozlabs.org/mailman/listinfo/linuxppc-embedded From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from planomail.kodiaknetworks.com (planomail.kodiaknetworks.com [76.203.108.6]) by ozlabs.org (Postfix) with ESMTP id 30022474C3 for ; Thu, 4 Dec 2008 08:06:25 +1100 (EST) MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Subject: RE: Marvell mv64462 + mpc744x PCI Interrupt Cause Error Handling Date: Wed, 3 Dec 2008 15:06:23 -0600 Message-ID: <295C5089A56CE143B316E5F67CA99CB001F4398C@cowboy.inovate.inovate.com> In-Reply-To: <4935E65B.30305@windriver.com> References: <295C5089A56CE143B316E5F67CA99CB001F4389F@cowboy.inovate.inovate.com> <4935E65B.30305@windriver.com> From: "Stephen Horton" To: "tiejun.chen" Cc: linuxppc-embedded@ozlabs.org List-Id: Linux on Embedded PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Hi Tiejun, Thanks. That is very helpful; I had not thought to look inside the drivers directory. Regards, stephen -----Original Message----- From: tiejun.chen [mailto:tiejun.chen@windriver.com]=20 Sent: Tuesday, December 02, 2008 7:52 PM To: Stephen Horton Cc: linuxppc-embedded@ozlabs.org Subject: Re: Marvell mv64462 + mpc744x PCI Interrupt Cause Error Handling Stephen Horton wrote: > Hi folks, >=20 > =20 >=20 > I'm really hoping Mark Greer or Dale Farnsworth sees this email. In my > current project, we are in the process of updating a board that has a > Marvell mv64462 system controller and a Freescale mpc7447A processor on > it. The board was originally developed with Gentoo Linux 2.6.9, but we > have ported Linux 2.6.24 kernel onto it. The new kernel is working well > thanks to help from people on this list. However, in examining the > differences between the 2 kernels, I can see that the 2.6.24 kernel > registers no interrupt error handling routines (see PCI Error Report > Register Map in the Marvell documentation). The error handlers for cpu, > sram, and specifically for the PCI bus that were previously in: >=20 > /arch/ppc/syslib/mv64460_pic.c >=20 > are absent from: >=20 > /arch/powerpc/sysdev/mv64x60_pic.c >=20 > =20 >=20 > Can someone who is familiar with the history of the ppc->powerpc > migration for this architecture help me figure out why the error > handling wasn't ported over? Is there some reason for this, such as it > didn't work properly or had some disadvantages? Is this error handling > done in 2.6.24 in some more generic manner that I have not yet found? Is > this an exercise left for me to implement? >=20 > =20 Untill v2.6.27 you can find corresponding dts content such as cpu-error, sram-ctrl, pci-error and mem-ctrl, in file, prpmc2800.dts. All events are handled on EDAC, error detection and reporting, as an unified mode. So you can track the directory drivers/edac/ for more information. I think these are not upgraded completely for v2.6.24. Best Regards Tiejun >=20 > Thanks, >=20 > Stephen >=20 >=20 >=20 >=20 > ------------------------------------------------------------------------ >=20 > _______________________________________________ > Linuxppc-embedded mailing list > Linuxppc-embedded@ozlabs.org > https://ozlabs.org/mailman/listinfo/linuxppc-embedded