Message ID | 1431985887-23379-1-git-send-email-ankgupta@codeaurora.org (mailing list archive) |
---|---|
State | Not Applicable, archived |
Delegated to: | Andy Gross |
Headers | show |
+Steven Rostedt On 05/18/15 14:51, Ankit Gupta wrote: > Add tracepoints to retrieve information about read, write > and non-data commands. For performance measurement support > tracepoints are added at the beginning and at the end of > transfers. Following is a list showing the new tracepoint > events. The "cmd" parameter here represents the opcode, SID, > and full 16-bit address. > > spmi_write_begin: cmd and data buffer. > spmi_write_end : cmd and return value. > spmi_read_begin : cmd. > spmi_read_end : cmd, return value and data buffer. > spmi_cmd : cmd. > > The reason that cmd appears at both the beginning and at > the end event is that SPMI drivers can request commands > concurrently. cmd helps in matching the corresponding > events. > > SPMI tracepoints can be enabled like: > > echo 1 >/sys/kernel/debug/tracing/events/spmi/enable > > and will dump messages that can be viewed in > /sys/kernel/debug/tracing/trace that look like: > > ... spmi_read_begin: opc=56 sid=00 addr=0x0000 > ... spmi_read_end: opc=56 sid=00 addr=0x0000 ret=0 len=02 buf=0x[01-40] > ... spmi_write_begin: opc=48 sid=00 addr=0x0000 len=3 buf=0x[ff-ff-ff] > > Signed-off-by: Ankit Gupta <ankgupta@codeaurora.org> > Signed-off-by: Gilad Avidov <gavidov@codeaurora.org> > Suggested-by: Sagar Dharia <sdharia@codeaurora.org> > --- Your Signed-off-by should be last. > drivers/spmi/spmi.c | 22 ++++++- > include/trace/events/spmi.h | 136 ++++++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 155 insertions(+), 3 deletions(-) > create mode 100644 include/trace/events/spmi.h > > diff --git a/drivers/spmi/spmi.c b/drivers/spmi/spmi.c > index 1d92f51..36fe010 100644 > --- a/drivers/spmi/spmi.c > +++ b/drivers/spmi/spmi.c > @@ -21,6 +21,8 @@ > #include <linux/pm_runtime.h> > > #include <dt-bindings/spmi/spmi.h> > +#define CREATE_TRACE_POINTS > +#include <trace/events/spmi.h> > > static DEFINE_IDA(ctrl_ida); > > @@ -95,28 +97,42 @@ EXPORT_SYMBOL_GPL(spmi_device_remove); > static inline int > spmi_cmd(struct spmi_controller *ctrl, u8 opcode, u8 sid) > { > + int ret; > + > if (!ctrl || !ctrl->cmd || ctrl->dev.type != &spmi_ctrl_type) > return -EINVAL; > > - return ctrl->cmd(ctrl, opcode, sid); > + ret = ctrl->cmd(ctrl, opcode, sid); > + trace_spmi_cmd(opcode, sid, ret); > + return ret; > } > > static inline int spmi_read_cmd(struct spmi_controller *ctrl, u8 opcode, > u8 sid, u16 addr, u8 *buf, size_t len) > { > + int ret; > + > if (!ctrl || !ctrl->read_cmd || ctrl->dev.type != &spmi_ctrl_type) > return -EINVAL; > > - return ctrl->read_cmd(ctrl, opcode, sid, addr, buf, len); > + trace_spmi_read_begin(opcode, sid, addr); > + ret = ctrl->read_cmd(ctrl, opcode, sid, addr, buf, len); > + trace_spmi_read_end(opcode, sid, addr, ret, len, buf); > + return ret; > } > > static inline int spmi_write_cmd(struct spmi_controller *ctrl, u8 opcode, > u8 sid, u16 addr, const u8 *buf, size_t len) > { > + int ret; > + > if (!ctrl || !ctrl->write_cmd || ctrl->dev.type != &spmi_ctrl_type) > return -EINVAL; > > - return ctrl->write_cmd(ctrl, opcode, sid, addr, buf, len); > + trace_spmi_write_begin(opcode, sid, addr, len, buf); > + ret = ctrl->write_cmd(ctrl, opcode, sid, addr, buf, len); > + trace_spmi_write_end(opcode, sid, addr, ret); > + return ret; > } > > /** > diff --git a/include/trace/events/spmi.h b/include/trace/events/spmi.h > new file mode 100644 > index 0000000..241fd1e > --- /dev/null > +++ b/include/trace/events/spmi.h > @@ -0,0 +1,136 @@ > +#undef TRACE_SYSTEM > +#define TRACE_SYSTEM spmi > + > +#if !defined(_TRACE_SPMI_H) || defined(TRACE_HEADER_MULTI_READ) > +#define _TRACE_SPMI_H > + > +#include <linux/spmi.h> > +#include <linux/tracepoint.h> > + > +/* > + * drivers/spmi/spmi.c > + */ > + > +TRACE_EVENT(spmi_write_begin, > + TP_PROTO(u8 opcode, u8 sid, u16 addr, u8 len, const u8 *buf), > + TP_ARGS(opcode, sid, addr, len, buf), > + > + TP_STRUCT__entry( > + __field ( u8, opcode ) > + __field ( u8, sid ) > + __field ( u16, addr ) > + __field ( u8, len ) > + __dynamic_array ( u8, buf, len + 1 ) > + ), > + > + TP_fast_assign( > + __entry->opcode = opcode; > + __entry->sid = sid; > + __entry->addr = addr; > + __entry->len = len + 1; > + memcpy(__get_dynamic_array(buf), buf, len + 1); > + ), > + > + TP_printk("opc=%d sid=%02d addr=0x%04x len=%d buf=0x[%*phD]", > + (int)__entry->opcode, (int)__entry->sid, > + (int)__entry->addr, (int)__entry->len, > + (int)__entry->len, __get_dynamic_array(buf)) > +); > + > +TRACE_EVENT(spmi_write_end, > + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret), > + TP_ARGS(opcode, sid, addr, ret), > + > + TP_STRUCT__entry( > + __field ( u8, opcode ) > + __field ( u8, sid ) > + __field ( u16, addr ) > + __field ( int, ret ) > + ), > + > + TP_fast_assign( > + __entry->opcode = opcode; > + __entry->sid = sid; > + __entry->addr = addr; > + __entry->ret = ret; > + ), > + > + TP_printk("opc=%d sid=%02d addr=0x%04x ret=%d", > + (int)__entry->opcode, (int)__entry->sid, > + (int)__entry->addr, __entry->ret) > +); > + > +TRACE_EVENT(spmi_read_begin, > + TP_PROTO(u8 opcode, u8 sid, u16 addr), > + TP_ARGS(opcode, sid, addr), > + > + TP_STRUCT__entry( > + __field ( u8, opcode ) > + __field ( u8, sid ) > + __field ( u16, addr ) > + ), > + > + TP_fast_assign( > + __entry->opcode = opcode; > + __entry->sid = sid; > + __entry->addr = addr; > + ), > + > + TP_printk("opc=%d sid=%02d addr=0x%04x", > + (int)__entry->opcode, (int)__entry->sid, > + (int)__entry->addr) > +); > + > +TRACE_EVENT(spmi_read_end, > + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret, u8 len, Should "len" be size_t instead of u8? It would at least match the implementation of spmi_controller::read_cmd(). Same comment for the write side. -Stephen > + const u8 *buf), > + TP_ARGS(opcode, sid, addr, ret, len, buf), > + > + TP_STRUCT__entry( > + __field ( u8, opcode ) > + __field ( u8, sid ) > + __field ( u16, addr ) > + __field ( int, ret ) > + __field ( u8, len ) > + __dynamic_array ( u8, buf, len + 1 ) > + ), > + > + TP_fast_assign( > + __entry->opcode = opcode; > + __entry->sid = sid; > + __entry->addr = addr; > + __entry->ret = ret; > + __entry->len = len + 1; > + memcpy(__get_dynamic_array(buf), buf, len + 1); > + ), > + > + TP_printk("opc=%d sid=%02d addr=0x%04x ret=%d len=%02d buf=0x[%*phD]", > + (int)__entry->opcode, (int)__entry->sid, > + (int)__entry->addr, __entry->ret, (int)__entry->len, > + (int)__entry->len, __get_dynamic_array(buf)) > +); > + > +TRACE_EVENT(spmi_cmd, > + TP_PROTO(u8 opcode, u8 sid, int ret), > + TP_ARGS(opcode, sid, ret), > + > + TP_STRUCT__entry( > + __field ( u8, opcode ) > + __field ( u8, sid ) > + __field ( int, ret ) > + ), > + > + TP_fast_assign( > + __entry->opcode = opcode; > + __entry->sid = sid; > + __entry->ret = ret; > + ), > + > + TP_printk("opc=%d sid=%02d ret=%d", (int)__entry->opcode, > + (int)__entry->sid, ret) > +); > + > +#endif /* _TRACE_SPMI_H */ > + > +/* This part must be outside protection */ > +#include <trace/define_trace.h>
On Tue, 19 May 2015 17:36:41 -0700 Stephen Boyd <sboyd@codeaurora.org> wrote: > +Steven Rostedt > > On 05/18/15 14:51, Ankit Gupta wrote: > > Add tracepoints to retrieve information about read, write > > and non-data commands. For performance measurement support > > tracepoints are added at the beginning and at the end of > > transfers. Following is a list showing the new tracepoint > > events. The "cmd" parameter here represents the opcode, SID, > > and full 16-bit address. > > > > spmi_write_begin: cmd and data buffer. > > spmi_write_end : cmd and return value. > > spmi_read_begin : cmd. > > spmi_read_end : cmd, return value and data buffer. > > spmi_cmd : cmd. > > > > The reason that cmd appears at both the beginning and at > > the end event is that SPMI drivers can request commands > > concurrently. cmd helps in matching the corresponding > > events. > > > > SPMI tracepoints can be enabled like: > > > > echo 1 >/sys/kernel/debug/tracing/events/spmi/enable > > > > and will dump messages that can be viewed in > > /sys/kernel/debug/tracing/trace that look like: > > > > ... spmi_read_begin: opc=56 sid=00 addr=0x0000 > > ... spmi_read_end: opc=56 sid=00 addr=0x0000 ret=0 len=02 buf=0x[01-40] > > ... spmi_write_begin: opc=48 sid=00 addr=0x0000 len=3 buf=0x[ff-ff-ff] > > > > Signed-off-by: Ankit Gupta <ankgupta@codeaurora.org> > > Signed-off-by: Gilad Avidov <gavidov@codeaurora.org> > > Suggested-by: Sagar Dharia <sdharia@codeaurora.org> > > --- > > Your Signed-off-by should be last. > The tracing part looks fine to me. For that: Acked-by: Steven Rostedt <rostedt@goodmis.org> -- Steve -- To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, 19 May 2015 17:36:41 -0700 Stephen Boyd <sboyd@codeaurora.org> wrote: > +Steven Rostedt > > On 05/18/15 14:51, Ankit Gupta wrote: > > Add tracepoints to retrieve information about read, write > > and non-data commands. For performance measurement support > > tracepoints are added at the beginning and at the end of > > transfers. Following is a list showing the new tracepoint > > events. The "cmd" parameter here represents the opcode, SID, > > and full 16-bit address. > > > > spmi_write_begin: cmd and data buffer. > > spmi_write_end : cmd and return value. > > spmi_read_begin : cmd. > > spmi_read_end : cmd, return value and data buffer. > > spmi_cmd : cmd. > > > > The reason that cmd appears at both the beginning and at > > the end event is that SPMI drivers can request commands > > concurrently. cmd helps in matching the corresponding > > events. > > > > SPMI tracepoints can be enabled like: > > > > echo 1 >/sys/kernel/debug/tracing/events/spmi/enable > > > > and will dump messages that can be viewed in > > /sys/kernel/debug/tracing/trace that look like: > > > > ... spmi_read_begin: opc=56 sid=00 addr=0x0000 > > ... spmi_read_end: opc=56 sid=00 addr=0x0000 ret=0 len=02 > > buf=0x[01-40] ... spmi_write_begin: opc=48 sid=00 addr=0x0000 len=3 > > buf=0x[ff-ff-ff] > > > > Signed-off-by: Ankit Gupta <ankgupta@codeaurora.org> > > Signed-off-by: Gilad Avidov <gavidov@codeaurora.org> > > Suggested-by: Sagar Dharia <sdharia@codeaurora.org> > > --- > > Your Signed-off-by should be last. Will change that. > > > drivers/spmi/spmi.c | 22 ++++++- > > include/trace/events/spmi.h | 136 > > ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 155 > > insertions(+), 3 deletions(-) create mode 100644 > > include/trace/events/spmi.h > > > > diff --git a/drivers/spmi/spmi.c b/drivers/spmi/spmi.c > > index 1d92f51..36fe010 100644 > > --- a/drivers/spmi/spmi.c > > +++ b/drivers/spmi/spmi.c > > @@ -21,6 +21,8 @@ > > #include <linux/pm_runtime.h> > > > > #include <dt-bindings/spmi/spmi.h> > > +#define CREATE_TRACE_POINTS > > +#include <trace/events/spmi.h> > > > > static DEFINE_IDA(ctrl_ida); > > > > @@ -95,28 +97,42 @@ EXPORT_SYMBOL_GPL(spmi_device_remove); > > static inline int > > spmi_cmd(struct spmi_controller *ctrl, u8 opcode, u8 sid) > > { > > + int ret; > > + > > if (!ctrl || !ctrl->cmd || ctrl->dev.type != > > &spmi_ctrl_type) return -EINVAL; > > > > - return ctrl->cmd(ctrl, opcode, sid); > > + ret = ctrl->cmd(ctrl, opcode, sid); > > + trace_spmi_cmd(opcode, sid, ret); > > + return ret; > > } > > > > static inline int spmi_read_cmd(struct spmi_controller *ctrl, u8 > > opcode, u8 sid, u16 addr, u8 *buf, size_t len) > > { > > + int ret; > > + > > if (!ctrl || !ctrl->read_cmd || ctrl->dev.type != > > &spmi_ctrl_type) return -EINVAL; > > > > - return ctrl->read_cmd(ctrl, opcode, sid, addr, buf, len); > > + trace_spmi_read_begin(opcode, sid, addr); > > + ret = ctrl->read_cmd(ctrl, opcode, sid, addr, buf, len); > > + trace_spmi_read_end(opcode, sid, addr, ret, len, buf); > > + return ret; > > } > > > > static inline int spmi_write_cmd(struct spmi_controller *ctrl, u8 > > opcode, u8 sid, u16 addr, const u8 *buf, size_t len) > > { > > + int ret; > > + > > if (!ctrl || !ctrl->write_cmd || ctrl->dev.type != > > &spmi_ctrl_type) return -EINVAL; > > > > - return ctrl->write_cmd(ctrl, opcode, sid, addr, buf, len); > > + trace_spmi_write_begin(opcode, sid, addr, len, buf); > > + ret = ctrl->write_cmd(ctrl, opcode, sid, addr, buf, len); > > + trace_spmi_write_end(opcode, sid, addr, ret); > > + return ret; > > } > > > > /** > > diff --git a/include/trace/events/spmi.h > > b/include/trace/events/spmi.h new file mode 100644 > > index 0000000..241fd1e > > --- /dev/null > > +++ b/include/trace/events/spmi.h > > @@ -0,0 +1,136 @@ > > +#undef TRACE_SYSTEM > > +#define TRACE_SYSTEM spmi > > + > > +#if !defined(_TRACE_SPMI_H) || defined(TRACE_HEADER_MULTI_READ) > > +#define _TRACE_SPMI_H > > + > > +#include <linux/spmi.h> > > +#include <linux/tracepoint.h> > > + > > +/* > > + * drivers/spmi/spmi.c > > + */ > > + > > +TRACE_EVENT(spmi_write_begin, > > + TP_PROTO(u8 opcode, u8 sid, u16 addr, u8 len, const u8 > > *buf), > > + TP_ARGS(opcode, sid, addr, len, buf), > > + > > + TP_STRUCT__entry( > > + __field ( u8, opcode ) > > + __field ( u8, sid ) > > + __field ( u16, addr ) > > + __field ( u8, len ) > > + __dynamic_array ( u8, buf, len + 1 ) > > + ), > > + > > + TP_fast_assign( > > + __entry->opcode = opcode; > > + __entry->sid = sid; > > + __entry->addr = addr; > > + __entry->len = len + 1; > > + memcpy(__get_dynamic_array(buf), buf, len + 1); > > + ), > > + > > + TP_printk("opc=%d sid=%02d addr=0x%04x len=%d > > buf=0x[%*phD]", > > + (int)__entry->opcode, (int)__entry->sid, > > + (int)__entry->addr, (int)__entry->len, > > + (int)__entry->len, __get_dynamic_array(buf)) > > +); > > + > > +TRACE_EVENT(spmi_write_end, > > + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret), > > + TP_ARGS(opcode, sid, addr, ret), > > + > > + TP_STRUCT__entry( > > + __field ( u8, opcode ) > > + __field ( u8, sid ) > > + __field ( u16, addr ) > > + __field ( int, ret ) > > + ), > > + > > + TP_fast_assign( > > + __entry->opcode = opcode; > > + __entry->sid = sid; > > + __entry->addr = addr; > > + __entry->ret = ret; > > + ), > > + > > + TP_printk("opc=%d sid=%02d addr=0x%04x ret=%d", > > + (int)__entry->opcode, (int)__entry->sid, > > + (int)__entry->addr, __entry->ret) > > +); > > + > > +TRACE_EVENT(spmi_read_begin, > > + TP_PROTO(u8 opcode, u8 sid, u16 addr), > > + TP_ARGS(opcode, sid, addr), > > + > > + TP_STRUCT__entry( > > + __field ( u8, opcode ) > > + __field ( u8, sid ) > > + __field ( u16, addr ) > > + ), > > + > > + TP_fast_assign( > > + __entry->opcode = opcode; > > + __entry->sid = sid; > > + __entry->addr = addr; > > + ), > > + > > + TP_printk("opc=%d sid=%02d addr=0x%04x", > > + (int)__entry->opcode, (int)__entry->sid, > > + (int)__entry->addr) > > +); > > + > > +TRACE_EVENT(spmi_read_end, > > + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret, u8 len, > > Should "len" be size_t instead of u8? It would at least match the > implementation of spmi_controller::read_cmd(). Same comment for the > write side. > > -Stephen I see no reason to spend to 4-8 bytes when spmi spec allows for maximum buffer size of 16. Do you suggest changing the API of read_cmd()? > > > + const u8 *buf), > > + TP_ARGS(opcode, sid, addr, ret, len, buf), > > + > > + TP_STRUCT__entry( > > + __field ( u8, opcode ) > > + __field ( u8, sid ) > > + __field ( u16, addr ) > > + __field ( int, ret ) > > + __field ( u8, len ) > > + __dynamic_array ( u8, buf, len + 1 ) > > + ), > > + > > + TP_fast_assign( > > + __entry->opcode = opcode; > > + __entry->sid = sid; > > + __entry->addr = addr; > > + __entry->ret = ret; > > + __entry->len = len + 1; > > + memcpy(__get_dynamic_array(buf), buf, len + 1); > > + ), > > + > > + TP_printk("opc=%d sid=%02d addr=0x%04x ret=%d len=%02d > > buf=0x[%*phD]", > > + (int)__entry->opcode, (int)__entry->sid, > > + (int)__entry->addr, __entry->ret, > > (int)__entry->len, > > + (int)__entry->len, __get_dynamic_array(buf)) > > +); > > + > > +TRACE_EVENT(spmi_cmd, > > + TP_PROTO(u8 opcode, u8 sid, int ret), > > + TP_ARGS(opcode, sid, ret), > > + > > + TP_STRUCT__entry( > > + __field ( u8, opcode ) > > + __field ( u8, sid ) > > + __field ( int, ret ) > > + ), > > + > > + TP_fast_assign( > > + __entry->opcode = opcode; > > + __entry->sid = sid; > > + __entry->ret = ret; > > + ), > > + > > + TP_printk("opc=%d sid=%02d ret=%d", (int)__entry->opcode, > > + (int)__entry->sid, ret) > > +); > > + > > +#endif /* _TRACE_SPMI_H */ > > + > > +/* This part must be outside protection */ > > +#include <trace/define_trace.h> > > Thanks, Ankit -- To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 05/20, Ankit Gupta wrote: > On Tue, 19 May 2015 17:36:41 -0700 > Stephen Boyd <sboyd@codeaurora.org> wrote: > > On 05/18/15 14:51, Ankit Gupta wrote: > > > + > > > +TRACE_EVENT(spmi_read_end, > > > + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret, u8 len, > > > > Should "len" be size_t instead of u8? It would at least match the > > implementation of spmi_controller::read_cmd(). Same comment for the > > write side. > > > > -Stephen > > I see no reason to spend to 4-8 bytes when spmi spec allows for maximum > buffer size of 16. Do you suggest changing the API of read_cmd()? > > Is that a maximum buffer size of 16 bytes? I'd prefer consistency with the API that's being traced, that's all. Changing it to u8 to save a few bytes probably doesn't make any difference if the architecture passes function arguments in CPU registers which are 32 or 64 bits wide or if the function can be inlined enough by the compiler to where the len parameter is kept in a register.
On Wed, 20 May 2015 11:29:55 -0700 Stephen Boyd <sboyd@codeaurora.org> wrote: > > I see no reason to spend to 4-8 bytes when spmi spec allows for maximum > > buffer size of 16. Do you suggest changing the API of read_cmd()? > > > > > Is that a maximum buffer size of 16 bytes? I'd prefer consistency > with the API that's being traced, that's all. Changing it to u8 > to save a few bytes probably doesn't make any difference if the > architecture passes function arguments in CPU registers which are > 32 or 64 bits wide or if the function can be inlined enough by > the compiler to where the len parameter is kept in a register. > I believe the worry is about wasting bytes in the ring buffer if not necessary. But we do that in other tracepoints, so it's really up to the maintainer. -- Steve -- To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 05/20/15 11:58, Steven Rostedt wrote: > On Wed, 20 May 2015 11:29:55 -0700 > Stephen Boyd <sboyd@codeaurora.org> wrote: > > >>> I see no reason to spend to 4-8 bytes when spmi spec allows for maximum >>> buffer size of 16. Do you suggest changing the API of read_cmd()? >> Is that a maximum buffer size of 16 bytes? I'd prefer consistency >> with the API that's being traced, that's all. Changing it to u8 >> to save a few bytes probably doesn't make any difference if the >> architecture passes function arguments in CPU registers which are >> 32 or 64 bits wide or if the function can be inlined enough by >> the compiler to where the len parameter is kept in a register. >> > > I believe the worry is about wasting bytes in the ring buffer if not > necessary. But we do that in other tracepoints, so it's really up to > the maintainer. Ah ok. Fair enough.
On 05/18/15 14:51, Ankit Gupta wrote: > Add tracepoints to retrieve information about read, write > and non-data commands. For performance measurement support > tracepoints are added at the beginning and at the end of > transfers. Following is a list showing the new tracepoint > events. The "cmd" parameter here represents the opcode, SID, > and full 16-bit address. > > spmi_write_begin: cmd and data buffer. > spmi_write_end : cmd and return value. > spmi_read_begin : cmd. > spmi_read_end : cmd, return value and data buffer. > spmi_cmd : cmd. > > The reason that cmd appears at both the beginning and at > the end event is that SPMI drivers can request commands > concurrently. cmd helps in matching the corresponding > events. > > SPMI tracepoints can be enabled like: > > echo 1 >/sys/kernel/debug/tracing/events/spmi/enable > > and will dump messages that can be viewed in > /sys/kernel/debug/tracing/trace that look like: > > ... spmi_read_begin: opc=56 sid=00 addr=0x0000 > ... spmi_read_end: opc=56 sid=00 addr=0x0000 ret=0 len=02 buf=0x[01-40] > ... spmi_write_begin: opc=48 sid=00 addr=0x0000 len=3 buf=0x[ff-ff-ff] > > Signed-off-by: Ankit Gupta <ankgupta@codeaurora.org> > Signed-off-by: Gilad Avidov <gavidov@codeaurora.org> > Suggested-by: Sagar Dharia <sdharia@codeaurora.org> > --- Assuming you fix sign-off-chain: Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
> On 05/20, Ankit Gupta wrote: >> On Tue, 19 May 2015 17:36:41 -0700 >> Stephen Boyd <sboyd@codeaurora.org> wrote: >> > On 05/18/15 14:51, Ankit Gupta wrote: >> > > + >> > > +TRACE_EVENT(spmi_read_end, >> > > + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret, u8 len, >> > >> > Should "len" be size_t instead of u8? It would at least match the >> > implementation of spmi_controller::read_cmd(). Same comment for the >> > write side. >> > >> > -Stephen >> >> I see no reason to spend to 4-8 bytes when spmi spec allows for maximum >> buffer size of 16. Do you suggest changing the API of read_cmd()? >> > > > Is that a maximum buffer size of 16 bytes? I'd prefer consistency > with the API that's being traced, that's all. Changing it to u8 > to save a few bytes probably doesn't make any difference if the > architecture passes function arguments in CPU registers which are > 32 or 64 bits wide or if the function can be inlined enough by > the compiler to where the len parameter is kept in a register. > Thanks Stephen for your comments. Will change it to size_t. > -- > Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, > a Linux Foundation Collaborative Project > -- To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
> On 05/18/15 14:51, Ankit Gupta wrote: >> Add tracepoints to retrieve information about read, write >> and non-data commands. For performance measurement support >> tracepoints are added at the beginning and at the end of >> transfers. Following is a list showing the new tracepoint >> events. The "cmd" parameter here represents the opcode, SID, >> and full 16-bit address. >> >> spmi_write_begin: cmd and data buffer. >> spmi_write_end : cmd and return value. >> spmi_read_begin : cmd. >> spmi_read_end : cmd, return value and data buffer. >> spmi_cmd : cmd. >> >> The reason that cmd appears at both the beginning and at >> the end event is that SPMI drivers can request commands >> concurrently. cmd helps in matching the corresponding >> events. >> >> SPMI tracepoints can be enabled like: >> >> echo 1 >/sys/kernel/debug/tracing/events/spmi/enable >> >> and will dump messages that can be viewed in >> /sys/kernel/debug/tracing/trace that look like: >> >> ... spmi_read_begin: opc=56 sid=00 addr=0x0000 >> ... spmi_read_end: opc=56 sid=00 addr=0x0000 ret=0 len=02 buf=0x[01-40] >> ... spmi_write_begin: opc=48 sid=00 addr=0x0000 len=3 buf=0x[ff-ff-ff] >> >> Signed-off-by: Ankit Gupta <ankgupta@codeaurora.org> >> Signed-off-by: Gilad Avidov <gavidov@codeaurora.org> >> Suggested-by: Sagar Dharia <sdharia@codeaurora.org> >> --- > > Assuming you fix sign-off-chain: > > Reviewed-by: Stephen Boyd <sboyd@codeaurora.org> Sorry, Missed your this email, will upload a new patch with just change of sign-off order. Thanks > > -- > Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, > a Linux Foundation Collaborative Project > > -- To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/spmi/spmi.c b/drivers/spmi/spmi.c index 1d92f51..36fe010 100644 --- a/drivers/spmi/spmi.c +++ b/drivers/spmi/spmi.c @@ -21,6 +21,8 @@ #include <linux/pm_runtime.h> #include <dt-bindings/spmi/spmi.h> +#define CREATE_TRACE_POINTS +#include <trace/events/spmi.h> static DEFINE_IDA(ctrl_ida); @@ -95,28 +97,42 @@ EXPORT_SYMBOL_GPL(spmi_device_remove); static inline int spmi_cmd(struct spmi_controller *ctrl, u8 opcode, u8 sid) { + int ret; + if (!ctrl || !ctrl->cmd || ctrl->dev.type != &spmi_ctrl_type) return -EINVAL; - return ctrl->cmd(ctrl, opcode, sid); + ret = ctrl->cmd(ctrl, opcode, sid); + trace_spmi_cmd(opcode, sid, ret); + return ret; } static inline int spmi_read_cmd(struct spmi_controller *ctrl, u8 opcode, u8 sid, u16 addr, u8 *buf, size_t len) { + int ret; + if (!ctrl || !ctrl->read_cmd || ctrl->dev.type != &spmi_ctrl_type) return -EINVAL; - return ctrl->read_cmd(ctrl, opcode, sid, addr, buf, len); + trace_spmi_read_begin(opcode, sid, addr); + ret = ctrl->read_cmd(ctrl, opcode, sid, addr, buf, len); + trace_spmi_read_end(opcode, sid, addr, ret, len, buf); + return ret; } static inline int spmi_write_cmd(struct spmi_controller *ctrl, u8 opcode, u8 sid, u16 addr, const u8 *buf, size_t len) { + int ret; + if (!ctrl || !ctrl->write_cmd || ctrl->dev.type != &spmi_ctrl_type) return -EINVAL; - return ctrl->write_cmd(ctrl, opcode, sid, addr, buf, len); + trace_spmi_write_begin(opcode, sid, addr, len, buf); + ret = ctrl->write_cmd(ctrl, opcode, sid, addr, buf, len); + trace_spmi_write_end(opcode, sid, addr, ret); + return ret; } /** diff --git a/include/trace/events/spmi.h b/include/trace/events/spmi.h new file mode 100644 index 0000000..241fd1e --- /dev/null +++ b/include/trace/events/spmi.h @@ -0,0 +1,136 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM spmi + +#if !defined(_TRACE_SPMI_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_SPMI_H + +#include <linux/spmi.h> +#include <linux/tracepoint.h> + +/* + * drivers/spmi/spmi.c + */ + +TRACE_EVENT(spmi_write_begin, + TP_PROTO(u8 opcode, u8 sid, u16 addr, u8 len, const u8 *buf), + TP_ARGS(opcode, sid, addr, len, buf), + + TP_STRUCT__entry( + __field ( u8, opcode ) + __field ( u8, sid ) + __field ( u16, addr ) + __field ( u8, len ) + __dynamic_array ( u8, buf, len + 1 ) + ), + + TP_fast_assign( + __entry->opcode = opcode; + __entry->sid = sid; + __entry->addr = addr; + __entry->len = len + 1; + memcpy(__get_dynamic_array(buf), buf, len + 1); + ), + + TP_printk("opc=%d sid=%02d addr=0x%04x len=%d buf=0x[%*phD]", + (int)__entry->opcode, (int)__entry->sid, + (int)__entry->addr, (int)__entry->len, + (int)__entry->len, __get_dynamic_array(buf)) +); + +TRACE_EVENT(spmi_write_end, + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret), + TP_ARGS(opcode, sid, addr, ret), + + TP_STRUCT__entry( + __field ( u8, opcode ) + __field ( u8, sid ) + __field ( u16, addr ) + __field ( int, ret ) + ), + + TP_fast_assign( + __entry->opcode = opcode; + __entry->sid = sid; + __entry->addr = addr; + __entry->ret = ret; + ), + + TP_printk("opc=%d sid=%02d addr=0x%04x ret=%d", + (int)__entry->opcode, (int)__entry->sid, + (int)__entry->addr, __entry->ret) +); + +TRACE_EVENT(spmi_read_begin, + TP_PROTO(u8 opcode, u8 sid, u16 addr), + TP_ARGS(opcode, sid, addr), + + TP_STRUCT__entry( + __field ( u8, opcode ) + __field ( u8, sid ) + __field ( u16, addr ) + ), + + TP_fast_assign( + __entry->opcode = opcode; + __entry->sid = sid; + __entry->addr = addr; + ), + + TP_printk("opc=%d sid=%02d addr=0x%04x", + (int)__entry->opcode, (int)__entry->sid, + (int)__entry->addr) +); + +TRACE_EVENT(spmi_read_end, + TP_PROTO(u8 opcode, u8 sid, u16 addr, int ret, u8 len, + const u8 *buf), + TP_ARGS(opcode, sid, addr, ret, len, buf), + + TP_STRUCT__entry( + __field ( u8, opcode ) + __field ( u8, sid ) + __field ( u16, addr ) + __field ( int, ret ) + __field ( u8, len ) + __dynamic_array ( u8, buf, len + 1 ) + ), + + TP_fast_assign( + __entry->opcode = opcode; + __entry->sid = sid; + __entry->addr = addr; + __entry->ret = ret; + __entry->len = len + 1; + memcpy(__get_dynamic_array(buf), buf, len + 1); + ), + + TP_printk("opc=%d sid=%02d addr=0x%04x ret=%d len=%02d buf=0x[%*phD]", + (int)__entry->opcode, (int)__entry->sid, + (int)__entry->addr, __entry->ret, (int)__entry->len, + (int)__entry->len, __get_dynamic_array(buf)) +); + +TRACE_EVENT(spmi_cmd, + TP_PROTO(u8 opcode, u8 sid, int ret), + TP_ARGS(opcode, sid, ret), + + TP_STRUCT__entry( + __field ( u8, opcode ) + __field ( u8, sid ) + __field ( int, ret ) + ), + + TP_fast_assign( + __entry->opcode = opcode; + __entry->sid = sid; + __entry->ret = ret; + ), + + TP_printk("opc=%d sid=%02d ret=%d", (int)__entry->opcode, + (int)__entry->sid, ret) +); + +#endif /* _TRACE_SPMI_H */ + +/* This part must be outside protection */ +#include <trace/define_trace.h>