General question : Debugging Kernel on PPC

2012-02-02 Thread Arshad, Farrukh
Greetings All, I have general question regarding kernel debugging on PowerPC platform. Except printk what are other ways for kernel / kernel modules debugging on PowerPC architecture (i.e P1022DS/RDK boards). I am specifically referring to the option of connecting hardware probe and debug the k

Re: [PATCH v2] powerpc/usb: fix issue of CPU halt when missing USB PHY clock

2012-02-02 Thread Sergei Shtylyov
Hello. On 02-02-2012 7:23, Shengzhou Liu wrote: when missing USB PHY clock, kernel booting up will halt during USB initialization. We should check USBGP[PHY_CLK_VALID] bit to avoid CPU hang in this case. Signed-off-by: Shengzhou Liu [...] diff --git a/drivers/usb/host/ehci-fsl.h b/drivers

Re: General question : Debugging Kernel on PPC

2012-02-02 Thread McClintock Matthew-B29882
On Thu, Feb 2, 2012 at 2:59 AM, Arshad, Farrukh wrote: > Greetings All, > > I have general question regarding kernel debugging on PowerPC platform. > Except printk what are other ways for kernel / kernel modules debugging on > PowerPC architecture (i.e P1022DS/RDK boards). I am specifically referr

[PATCH] Implement GET_IP/SET_IP for powerpc architecture.

2012-02-02 Thread Srikar Dronamraju
With this change, helpers such as instruction_pointer() et al, get defined in the generic header in terms of GET_IP Signed-off-by: Srikar Dronamraju Signed-off-by: Ananth N Mavinakayanahalli --- arch/powerpc/include/asm/ptrace.h | 22 ++ 1 files changed, 14 insertions(+),

Re: [PATCH] Implement GET_IP/SET_IP for powerpc architecture.

2012-02-02 Thread Mike Frysinger
On Thursday 02 February 2012 09:41:25 Srikar Dronamraju wrote: > +#define GET_FP(regs) (0) > +#define SET_FP(regs, val) ppc doesn't have a standard FP location ? > +#define profile_pc(regs) GET_IP(regs) pretty sure you don't need this as asm-generic/ptrace.h already has a definition for

Re: [PATCH] Implement GET_IP/SET_IP for powerpc architecture.

2012-02-02 Thread Srikar Dronamraju
* Mike Frysinger [2012-02-02 11:28:46]: > On Thursday 02 February 2012 09:41:25 Srikar Dronamraju wrote: > > +#define GET_FP(regs) (0) > > +#define SET_FP(regs, val) > > ppc doesn't have a standard FP location ? > > > +#define profile_pc(regs) GET_IP(regs) > > pretty sure you don

Re: Efika (mpc5200b): sound doesn't build/work from linux-2.6.38.x

2012-02-02 Thread acrux
On Thu, 2 Feb 2012 03:48:03 +0100 acrux wrote: > board: Genesi Efika (MPC5200B) > problem: unable to use sound subsystem instead it works fine with > linux-2.6.36.4 (+ device tree supplement, attached) . > > With linux-2.6.38.8 and linux-2.6.39.4 it doesn't build. > That's from my build log: >

Re: [PATCH] Implement GET_IP/SET_IP for powerpc architecture.

2012-02-02 Thread Mike Frysinger
On Thursday 02 February 2012 11:36:59 Srikar Dronamraju wrote: > * Mike Frysinger [2012-02-02 11:28:46]: > > On Thursday 02 February 2012 09:41:25 Srikar Dronamraju wrote: > > > +#define profile_pc(regs) GET_IP(regs) > > > > pretty sure you don't need this as asm-generic/ptrace.h already has a >

Re: [PATCH] powerpc: dont include 8250 pre-setup if 8250 driver isn't enabled.

2012-02-02 Thread Paul Gortmaker
On Fri, Jan 20, 2012 at 10:50 PM, Paul Gortmaker wrote: > The legacy_serial code was setting things up based on the assumption > that the main 8250 driver would be loaded shortly after.  But some > randconfigs exposed an issue where early debug (UDB) UART support was > enabled, yet the core UART 8

Re: MPC5200B FEC TX packets getting stuck

2012-02-02 Thread Joey Nelson
I ran my test overnight with the TX irq time-stamping and finally got one delayed packet. Based on the time-stamp data the packets are getting stuck in the FEC TX FIFO. The calls to the xmit function are spaced our on about 150 us interval for a larger TCP socket write. The BestComm tx irq is han

Re: [PATCH v2] powerpc/usb: fix issue of CPU halt when missing USB PHY clock

2012-02-02 Thread Greg KH
On Thu, Feb 02, 2012 at 03:39:48PM +0400, Sergei Shtylyov wrote: > Hello. > > On 02-02-2012 7:23, Shengzhou Liu wrote: > > >when missing USB PHY clock, kernel booting up will halt during USB > >initialization. We should check USBGP[PHY_CLK_VALID] bit to avoid > >CPU hang in this case. > > >Signe

Re: [PATCH] Implement GET_IP/SET_IP for powerpc architecture.

2012-02-02 Thread Benjamin Herrenschmidt
On Thu, 2012-02-02 at 11:28 -0500, Mike Frysinger wrote: > On Thursday 02 February 2012 09:41:25 Srikar Dronamraju wrote: > > +#define GET_FP(regs) (0) > > +#define SET_FP(regs, val) > > ppc doesn't have a standard FP location ? Not really no, it's the sp (r1). r31 might be consider

Re: [PATCH] powerpc: dont include 8250 pre-setup if 8250 driver isn't enabled.

2012-02-02 Thread Benjamin Herrenschmidt
On Thu, 2012-02-02 at 12:56 -0500, Paul Gortmaker wrote: > On Fri, Jan 20, 2012 at 10:50 PM, Paul Gortmaker > wrote: > > The legacy_serial code was setting things up based on the assumption > > that the main 8250 driver would be loaded shortly after. But some > > randconfigs exposed an issue wher

Re: Efika (mpc5200b): sound doesn't build/work from linux-2.6.38.x

2012-02-02 Thread Tabi Timur-B04825
On Thu, Feb 2, 2012 at 10:57 AM, acrux wrote: > well, i got the same error with also linux-2.6.37. Btw, this was already > reported about a year ago: > http://lists.ozlabs.org/pipermail/linuxppc-dev/2011-February/088415.html I think this was fixed already. You're using an obsolete kernel. --

[RFC] Multi queue support in ethernet/freescale/ucc_geth.c

2012-02-02 Thread Paul Gortmaker
Hi Li, A while back DaveM mentioned that it would be good to break out the ring allocations[1] in this driver. I was looking at it, and in the process noticed this: $ grep 'numQueues.*=' drivers/net/ethernet/freescale/ucc_geth.c .numQueuesTx = 1, .numQueuesRx = 1, $ My interpr

[PATCH] powerpc: Rework lazy-interrupt handling

2012-02-02 Thread Benjamin Herrenschmidt
The current implementation of lazy interrupts handling has some issues that this tries to address. Except on iSeries, we don't do the various workarounds we need to do on re-enable when returning from an interrupt, which can do an implicit re-enable, and thus we may still lose or get delayed decre