Message ID | 20090111153013.GA31791@oksana.dev.rtsoft.ru (mailing list archive) |
---|---|
State | Accepted, archived |
Commit | 6c9789de2a02755101f757789ffcb17312f3fa9b |
Delegated to: | Kumar Gala |
Headers | show |
On Jan 12, 2009, at 1:55 AM, Li Yang wrote: >> -----Original Message----- >> From: linuxppc-dev-bounces+leoli=freescale.com@ozlabs.org >> [mailto:linuxppc-dev-bounces+leoli=freescale.com@ozlabs.org] >> On Behalf Of Anton Vorontsov >> Sent: Sunday, January 11, 2009 11:30 PM >> To: Kumar Gala >> Cc: linuxppc-dev@ozlabs.org >> Subject: [PATCH] powerpc/83xx: Make serial ports work on >> MPC8315E-RDB w/FSL U-Boots >> >> FSL U-Boots use /soc8315@e0000000 node to search and fixup >> serial nodes' clock-frequency properties. Though in upstream >> kernels we use new naming convention -- for IMMR address >> space dts files specify /immr@e0000000 nodes. >> >> This makes FSL U-Boots fail to fixup the clock frequencies, >> and that leads to serial ports misbehaviour. We can >> workaround the issue by filling the clock frequency values manually. > > Freescale BSP is for customer who needs the out-of-box experience. > It's > better tested, but doesn't update very frequently. I would suggest > the > customer to use the upstream u-boot, if they decide to use latest > upstream kernel. > > - Leo Agreed. However I don't think there is any harm w/the patch as upstream u-boot will update the property. - k
On Jan 11, 2009, at 9:30 AM, Anton Vorontsov wrote: > FSL U-Boots use /soc8315@e0000000 node to search and fixup serial > nodes' clock-frequency properties. Though in upstream kernels we use > new naming convention -- for IMMR address space dts files specify > /immr@e0000000 nodes. > > This makes FSL U-Boots fail to fixup the clock frequencies, and that > leads to serial ports misbehaviour. We can workaround the issue by > filling the clock frequency values manually. > > p.s. For the same reason FSL U-Boots fail to fixup MAC addresses for > ethernet nodes, so users should either change the .dts file locally > or set MAC address via `ifconfig hw ether' command. > > Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> > --- > > Leon, > > With this patch 2.6.28 kernel boots and works fine with the U-Boot > I took from this image (downloadable from freescale.com): > > $ md5sum MPC8315ERDB_20080321-ltib.iso > 009730826366d593347b88fa4fdb9be0 MPC8315ERDB_20080321-ltib.iso > > That is, U-Boot 1.3.0-rc2 (Mar 21 2008 - 13:36:09) MPC83XX > > arch/powerpc/boot/dts/mpc8315erdb.dts | 4 ++-- > 1 files changed, 2 insertions(+), 2 deletions(-) applied - k
Hello, On Mon, Jan 12, 2009 at 4:07 PM, Kumar Gala <galak@kernel.crashing.org> wrote: > On Jan 12, 2009, at 1:55 AM, Li Yang wrote: >>> -----Original Message----- >> Freescale BSP is for customer who needs the out-of-box experience. It's >> better tested, but doesn't update very frequently. I would suggest the >> customer to use the upstream u-boot, if they decide to use latest >> upstream kernel. >> > The upstream u-boot did however not support PCI Express, which the latest Freescale kernel did. So it was a chicken-egg problem experience for people wanting to test both the latest Freescale kernel as well as Antov's work to bring the PCI Express support upstream. Thanks Antov et al,
On Jan 12, 2009, at 11:20 AM, Leon Woestenberg wrote: > Hello, > > On Mon, Jan 12, 2009 at 4:07 PM, Kumar Gala > <galak@kernel.crashing.org> wrote: >> On Jan 12, 2009, at 1:55 AM, Li Yang wrote: >>>> -----Original Message----- >>> Freescale BSP is for customer who needs the out-of-box >>> experience. It's >>> better tested, but doesn't update very frequently. I would >>> suggest the >>> customer to use the upstream u-boot, if they decide to use latest >>> upstream kernel. >>> >> > The upstream u-boot did however not support PCI Express, which the > latest Freescale kernel did. > > So it was a chicken-egg problem experience for people wanting to test > both the latest Freescale kernel as well as Antov's work to bring the > PCI Express support upstream. > > Thanks Antov et al, Please gripe at your Freescale contacts about getting code into mainline/upstream kernel. (The more they hear it from customers the more importance it will get. The team I work on at Freescale can focus on only so much and the other groups need to hear that customers want code in mainline.) And thanks Anton for cleaning all this up and getting it in a state that is acceptable :) - k
diff --git a/arch/powerpc/boot/dts/mpc8315erdb.dts b/arch/powerpc/boot/dts/mpc8315erdb.dts index 9a4fa2a..88d691c 100644 --- a/arch/powerpc/boot/dts/mpc8315erdb.dts +++ b/arch/powerpc/boot/dts/mpc8315erdb.dts @@ -257,7 +257,7 @@ device_type = "serial"; compatible = "ns16550"; reg = <0x4500 0x100>; - clock-frequency = <0>; + clock-frequency = <133333333>; interrupts = <9 0x8>; interrupt-parent = <&ipic>; }; @@ -267,7 +267,7 @@ device_type = "serial"; compatible = "ns16550"; reg = <0x4600 0x100>; - clock-frequency = <0>; + clock-frequency = <133333333>; interrupts = <10 0x8>; interrupt-parent = <&ipic>; };
FSL U-Boots use /soc8315@e0000000 node to search and fixup serial nodes' clock-frequency properties. Though in upstream kernels we use new naming convention -- for IMMR address space dts files specify /immr@e0000000 nodes. This makes FSL U-Boots fail to fixup the clock frequencies, and that leads to serial ports misbehaviour. We can workaround the issue by filling the clock frequency values manually. p.s. For the same reason FSL U-Boots fail to fixup MAC addresses for ethernet nodes, so users should either change the .dts file locally or set MAC address via `ifconfig hw ether' command. Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> --- Leon, With this patch 2.6.28 kernel boots and works fine with the U-Boot I took from this image (downloadable from freescale.com): $ md5sum MPC8315ERDB_20080321-ltib.iso 009730826366d593347b88fa4fdb9be0 MPC8315ERDB_20080321-ltib.iso That is, U-Boot 1.3.0-rc2 (Mar 21 2008 - 13:36:09) MPC83XX arch/powerpc/boot/dts/mpc8315erdb.dts | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)