On Wed, Jul 17, 2019 at 10:56:02AM -0600, Mathieu Poirier wrote:
> On Fri, Jul 12, 2019 at 07:46:27PM +0530, Sai Prakash Ranjan wrote:
> > Add support for coresight CPU debug module on Qualcomm
> > Kryo CPUs. This patch adds the UCI entries for Kryo CPUs
> > found on MSM8996 which shares the same PIDs as ETMs.
> >
> > Without this, below error is observed on MSM8996:
> >
> > [ 5.429867] OF: graph: no port node found in /soc/debug@3810000
> > [ 5.429938] coresight-etm4x: probe of 3810000.debug failed with error -22
> > [ 5.435415] coresight-cpu-debug 3810000.debug: Coresight debug-CPU0 initialized
> > [ 5.446474] OF: graph: no port node found in /soc/debug@3910000
> > [ 5.448927] coresight-etm4x: probe of 3910000.debug failed with error -22
> > [ 5.454681] coresight-cpu-debug 3910000.debug: Coresight debug-CPU1 initialized
> > [ 5.487765] OF: graph: no port node found in /soc/debug@3a10000
> > [ 5.488007] coresight-etm4x: probe of 3a10000.debug failed with error -22
> > [ 5.493024] coresight-cpu-debug 3a10000.debug: Coresight debug-CPU2 initialized
> > [ 5.501802] OF: graph: no port node found in /soc/debug@3b10000
> > [ 5.512901] coresight-etm4x: probe of 3b10000.debug failed with error -22
> > [ 5.513192] coresight-cpu-debug 3b10000.debug: Coresight debug-CPU3 initialized
> >
> > Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@xxxxxxxxxxxxxx>
> > ---
> > .../hwtracing/coresight/coresight-cpu-debug.c | 33 +++++++++----------
> > drivers/hwtracing/coresight/coresight-priv.h | 10 +++---
> > 2 files changed, 21 insertions(+), 22 deletions(-)
> >
> > diff --git a/drivers/hwtracing/coresight/coresight-cpu-debug.c b/drivers/hwtracing/coresight/coresight-cpu-debug.c
> > index 2463aa7ab4f6..96544b348c27 100644
> > --- a/drivers/hwtracing/coresight/coresight-cpu-debug.c
> > +++ b/drivers/hwtracing/coresight/coresight-cpu-debug.c
> > @@ -646,24 +646,23 @@ static int debug_remove(struct amba_device *adev)
> > return 0;
> > }
> >
> > +static const struct amba_cs_uci_id uci_id_debug[] = {
> > + {
> > + /* CPU Debug UCI data */
> > + .devarch = 0x47706a15,
> > + .devarch_mask = 0xfff0ffff,
> > + .devtype = 0x00000015,
> > + }
> > +};
> > +
> > static const struct amba_id debug_ids[] = {
> > - { /* Debug for Cortex-A53 */
> > - .id = 0x000bbd03,
> > - .mask = 0x000fffff,
> > - },
> > - { /* Debug for Cortex-A57 */
> > - .id = 0x000bbd07,
> > - .mask = 0x000fffff,
> > - },
> > - { /* Debug for Cortex-A72 */
> > - .id = 0x000bbd08,
> > - .mask = 0x000fffff,
> > - },
> > - { /* Debug for Cortex-A73 */
> > - .id = 0x000bbd09,
> > - .mask = 0x000fffff,
> > - },
> > - { 0, 0 },
> > + CS_AMBA_ID(0x000bbd03), /* Cortex-A53 */
> > + CS_AMBA_ID(0x000bbd07), /* Cortex-A57 */
> > + CS_AMBA_ID(0x000bbd08), /* Cortex-A72 */
> > + CS_AMBA_ID(0x000bbd09), /* Cortex-A73 */
> > + CS_AMBA_UCI_ID(0x000f0205, uci_id_debug), /* Qualcomm Kryo */
> > + CS_AMBA_UCI_ID(0x000f0211, uci_id_debug), /* Qualcomm Kryo */
> > + {},
> > };
> >
> > static struct amba_driver debug_driver = {
> > diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h
> > index 7d401790dd7e..41ae5863104d 100644
> > --- a/drivers/hwtracing/coresight/coresight-priv.h
> > +++ b/drivers/hwtracing/coresight/coresight-priv.h
> > @@ -185,11 +185,11 @@ static inline int etm_writel_cp14(u32 off, u32 val) { return 0; }
> > }
> >
> > /* coresight AMBA ID, full UCI structure: id table entry. */
> > -#define CS_AMBA_UCI_ID(pid, uci_ptr) \
> > - { \
> > - .id = pid, \
> > - .mask = 0x000fffff, \
> > - .data = uci_ptr \
> > +#define CS_AMBA_UCI_ID(pid, uci_ptr) \
> > + { \
> > + .id = pid, \
> > + .mask = 0x000fffff, \
> > + .data = (void *)uci_ptr \
> > }
>
> I will pickup this patch - it will show up in my next tree when rc1 comes out.
I tested this patch on the mainline kernel with latest commit
f1a3b43cc1f5 ("Merge branch 'for-linus' of
git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input"). FWIW:
Tested-by: Leo Yan <leo.yan@xxxxxxxxxx>
P.s. Acutally I tested this patch for 5.2-rcx a few days ago and found
a regression for CPU debug module: I observed the CPU debug module
panic dump will stuck. After I pulled to latest kernel code base the
CPU debug module can work well; also works well with this patch. F.Y.I.
Thanks,
Leo Yan
