| Message ID | 1411467581-2982-1-git-send-email-aik@ozlabs.ru |
|---|---|
| State | New |
| Headers | show |
On Di, 2014-09-23 at 20:19 +1000, Alexey Kardashevskiy wrote: > This adds a stub for ohci_td_pkt() function (which traces packets) > when configured without --enable-trace-backend Ah, cool. Just noticed that, while doing usb patch queue test builds for the next pull req. > It should probably be squashed to > [PATCH] ohci: Convert fprint/DPRINTF/print to traces Done. thanks, Gerd
On Tue, Sep 23, 2014 at 08:19:41PM +1000, Alexey Kardashevskiy wrote: > This adds a stub for ohci_td_pkt() function (which traces packets) > when configured without --enable-trace-backend > > Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> > --- > > It should probably be squashed to > [PATCH] ohci: Convert fprint/DPRINTF/print to traces > > Sorry about that... > > --- > hw/usb/hcd-ohci.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/hw/usb/hcd-ohci.c b/hw/usb/hcd-ohci.c > index 4a4dd02..7ea871d 100644 > --- a/hw/usb/hcd-ohci.c > +++ b/hw/usb/hcd-ohci.c > @@ -920,6 +920,7 @@ static int ohci_service_iso_td(OHCIState *ohci, struct ohci_ed *ed, > return 1; > } > > +#ifdef trace_event_get_state > static void ohci_td_pkt(const char *msg, const uint8_t *buf, size_t len) > { > bool print16 = !!trace_event_get_state(TRACE_USB_OHCI_TD_PKT_SHORT); The preferred define for this is CONFIG_TRACE_NOP. Checking for trace_event_get_state could break if trace/control.h changes in the future. Stefan
On 09/23/2014 08:52 PM, Gerd Hoffmann wrote: > On Di, 2014-09-23 at 20:19 +1000, Alexey Kardashevskiy wrote: >> This adds a stub for ohci_td_pkt() function (which traces packets) >> when configured without --enable-trace-backend > > Ah, cool. Just noticed that, while doing usb patch queue test builds > for the next pull req. > >> It should probably be squashed to >> [PATCH] ohci: Convert fprint/DPRINTF/print to traces > > Done. Posted another patch for dtrace backend... Sorry for too much noise.
On 09/24/2014 07:47 PM, Stefan Hajnoczi wrote: > On Tue, Sep 23, 2014 at 08:19:41PM +1000, Alexey Kardashevskiy wrote: >> This adds a stub for ohci_td_pkt() function (which traces packets) >> when configured without --enable-trace-backend >> >> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> >> --- >> >> It should probably be squashed to >> [PATCH] ohci: Convert fprint/DPRINTF/print to traces >> >> Sorry about that... >> >> --- >> hw/usb/hcd-ohci.c | 6 ++++++ >> 1 file changed, 6 insertions(+) >> >> diff --git a/hw/usb/hcd-ohci.c b/hw/usb/hcd-ohci.c >> index 4a4dd02..7ea871d 100644 >> --- a/hw/usb/hcd-ohci.c >> +++ b/hw/usb/hcd-ohci.c >> @@ -920,6 +920,7 @@ static int ohci_service_iso_td(OHCIState *ohci, struct ohci_ed *ed, >> return 1; >> } >> >> +#ifdef trace_event_get_state >> static void ohci_td_pkt(const char *msg, const uint8_t *buf, size_t len) >> { >> bool print16 = !!trace_event_get_state(TRACE_USB_OHCI_TD_PKT_SHORT); > > The preferred define for this is CONFIG_TRACE_NOP. > > Checking for trace_event_get_state could break if trace/control.h > changes in the future. I am not familiar with "dtrace" - are traces configurable with it? With --enable-trace-backend=dtrace, trace_event_get_state is not defined.
On Thu, Sep 25, 2014 at 10:19:32AM +1000, Alexey Kardashevskiy wrote: > On 09/24/2014 07:47 PM, Stefan Hajnoczi wrote: > > On Tue, Sep 23, 2014 at 08:19:41PM +1000, Alexey Kardashevskiy wrote: > >> This adds a stub for ohci_td_pkt() function (which traces packets) > >> when configured without --enable-trace-backend > >> > >> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> > >> --- > >> > >> It should probably be squashed to > >> [PATCH] ohci: Convert fprint/DPRINTF/print to traces > >> > >> Sorry about that... > >> > >> --- > >> hw/usb/hcd-ohci.c | 6 ++++++ > >> 1 file changed, 6 insertions(+) > >> > >> diff --git a/hw/usb/hcd-ohci.c b/hw/usb/hcd-ohci.c > >> index 4a4dd02..7ea871d 100644 > >> --- a/hw/usb/hcd-ohci.c > >> +++ b/hw/usb/hcd-ohci.c > >> @@ -920,6 +920,7 @@ static int ohci_service_iso_td(OHCIState *ohci, struct ohci_ed *ed, > >> return 1; > >> } > >> > >> +#ifdef trace_event_get_state > >> static void ohci_td_pkt(const char *msg, const uint8_t *buf, size_t len) > >> { > >> bool print16 = !!trace_event_get_state(TRACE_USB_OHCI_TD_PKT_SHORT); > > > > The preferred define for this is CONFIG_TRACE_NOP. > > > > Checking for trace_event_get_state could break if trace/control.h > > changes in the future. > > I am not familiar with "dtrace" - are traces configurable with it? With > --enable-trace-backend=dtrace, trace_event_get_state is not defined. dtrace has its own tools for enabling/disabling trace events. QEMU's command cannot be used with dtrace.
diff --git a/hw/usb/hcd-ohci.c b/hw/usb/hcd-ohci.c index 4a4dd02..7ea871d 100644 --- a/hw/usb/hcd-ohci.c +++ b/hw/usb/hcd-ohci.c @@ -920,6 +920,7 @@ static int ohci_service_iso_td(OHCIState *ohci, struct ohci_ed *ed, return 1; } +#ifdef trace_event_get_state static void ohci_td_pkt(const char *msg, const uint8_t *buf, size_t len) { bool print16 = !!trace_event_get_state(TRACE_USB_OHCI_TD_PKT_SHORT); @@ -950,6 +951,11 @@ static void ohci_td_pkt(const char *msg, const uint8_t *buf, size_t len) p += sprintf(p, " %.2x", buf[i]); } } +#else +static void ohci_td_pkt(const char *msg, const uint8_t *buf, size_t len) +{ +} +#endif /* Service a transport descriptor. Returns nonzero to terminate processing of this endpoint. */
This adds a stub for ohci_td_pkt() function (which traces packets) when configured without --enable-trace-backend Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> --- It should probably be squashed to [PATCH] ohci: Convert fprint/DPRINTF/print to traces Sorry about that... --- hw/usb/hcd-ohci.c | 6 ++++++ 1 file changed, 6 insertions(+)