| Message ID | 20230904140705.1620708-3-james.clark@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | coresight: Allow guests to be traced when FEAT_TRF and VHE are present | expand |
Hi James On 04/09/2023 15:07, James Clark wrote: > Currently the userspace and kernel filters for guests are never set, so > no trace will be generated for them. Add it by writing to the guest > filters when exclude_guest isn't set. By writing either E1TRE or E0TRE, > filtering on either guest kernel or guest userspace is also supported. > > Since TRFCR_EL1 access is trapped, this can't be modified by the guest. > > This change also brings exclude_host support which is difficult to add > as a separate commit without excess churn and resulting in no trace at > all. > > Testing > ======= > > The addresses were counted with the following: > > $ perf report -D | grep -Eo 'EL2|EL1|EL0' | sort | uniq -c > > Guest kernel only: > > $ perf record -e cs_etm//Gk -a -- true > 535 EL1 > 1 EL2 > > Guest user only (0 addresses expected because the guest OS hasn't reached > userspace yet): > > $ perf record -e cs_etm//Gu -a -- true > > Host kernel only: > > $ perf record -e cs_etm//Hk -a -- true > 3501 EL2 > > Host userspace only: > > $ perf record -e cs_etm//Hu -a -- true > 408 EL0 > 1 EL2 > > Reviewed-by: Mark Brown <broonie@kernel.org> (sysreg) > Signed-off-by: James Clark <james.clark@arm.com> > --- > arch/arm64/tools/sysreg | 4 ++ > .../coresight/coresight-etm4x-core.c | 51 ++++++++++++++++--- > drivers/hwtracing/coresight/coresight-etm4x.h | 2 +- > drivers/hwtracing/coresight/coresight-priv.h | 3 ++ > 4 files changed, 53 insertions(+), 7 deletions(-) > > diff --git a/arch/arm64/tools/sysreg b/arch/arm64/tools/sysreg > index 6ca7db69d6c9..cae9139b6c05 100644 > --- a/arch/arm64/tools/sysreg > +++ b/arch/arm64/tools/sysreg > @@ -2531,3 +2531,7 @@ EndSysreg > Sysreg TRFCR_EL2 3 4 1 2 1 > Fields TRFCR_EL2 > EndSysreg > + > +Sysreg TRFCR_EL12 3 5 1 2 1 > +Fields TRFCR_ELx > +EndSysreg > diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c > index 77b0271ce6eb..6c16a14d6fbe 100644 > --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c > +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c > @@ -274,6 +274,18 @@ static void etm4x_prohibit_trace(struct etmv4_drvdata *drvdata) > cpu_prohibit_trace(); > } > > +static u64 etm4x_get_kern_user_filter(struct etmv4_drvdata *drvdata) > +{ > + u64 trfcr = drvdata->trfcr; > + > + if (drvdata->config.mode & ETM_MODE_EXCL_KERN) > + trfcr &= ~TRFCR_ELx_ExTRE; > + if (drvdata->config.mode & ETM_MODE_EXCL_USER) > + trfcr &= ~TRFCR_ELx_E0TRE; > + > + return trfcr; > +} > + > /* > * etm4x_allow_trace - Allow CPU tracing in the respective ELs, > * as configured by the drvdata->config.mode for the current > @@ -286,18 +298,39 @@ static void etm4x_prohibit_trace(struct etmv4_drvdata *drvdata) > */ > static void etm4x_allow_trace(struct etmv4_drvdata *drvdata) > { > - u64 trfcr = drvdata->trfcr; > + u64 trfcr; > > /* If the CPU doesn't support FEAT_TRF, nothing to do */ > - if (!trfcr) > + if (!drvdata->trfcr) > return; > > - if (drvdata->config.mode & ETM_MODE_EXCL_KERN) > - trfcr &= ~TRFCR_ELx_ExTRE; > - if (drvdata->config.mode & ETM_MODE_EXCL_USER) > - trfcr &= ~TRFCR_ELx_E0TRE; > + if (drvdata->config.mode & ETM_MODE_EXCL_HOST) > + trfcr = drvdata->trfcr & ~(TRFCR_ELx_ExTRE | TRFCR_ELx_E0TRE); > + else > + trfcr = etm4x_get_kern_user_filter(drvdata); > > write_trfcr(trfcr); > + > + /* > + * Filters for EL1 and EL0 (when running a guest) are stored in > + * TRFCR_EL1 so write it there for VHE. For nVHE, the filters in > + * have to be re-applied when switching to the guest instead. > + */ > + if (!is_kernel_in_hyp_mode()) > + return; > + > + if (drvdata->config.mode & ETM_MODE_EXCL_GUEST) > + trfcr = drvdata->trfcr & ~(TRFCR_ELx_ExTRE | TRFCR_ELx_E0TRE); > + else > + trfcr = etm4x_get_kern_user_filter(drvdata); > + > + /* > + * TRFCR_EL1 doesn't have CX and TRFCR_EL1.TS has no effect when TS is > + * set in EL2 so mask them out. > + */ > + trfcr &= ~(TRFCR_ELx_TS_MASK | TRFCR_EL2_CX); > + > + write_sysreg_s(trfcr, SYS_TRFCR_EL12); While I agree with the intention of the patch, I am not sure if we should do the TRFCR_EL1 (via TRFCR_EL12) updates from CoreSight driver. It might be much better to do it from the KVM for both VHE and nVHE. Marc What are your thoughts on this ? Suzuki > } > > #ifdef CONFIG_ETM4X_IMPDEF_FEATURE > @@ -655,6 +688,12 @@ static int etm4_parse_event_config(struct coresight_device *csdev, > if (attr->exclude_user) > config->mode = ETM_MODE_EXCL_USER; > > + if (attr->exclude_host) > + config->mode |= ETM_MODE_EXCL_HOST; > + > + if (attr->exclude_guest) > + config->mode |= ETM_MODE_EXCL_GUEST; > + > /* Always start from the default config */ > etm4_set_default_config(config); > > diff --git a/drivers/hwtracing/coresight/coresight-etm4x.h b/drivers/hwtracing/coresight/coresight-etm4x.h > index 20e2e4cb7614..3f170599822f 100644 > --- a/drivers/hwtracing/coresight/coresight-etm4x.h > +++ b/drivers/hwtracing/coresight/coresight-etm4x.h > @@ -841,7 +841,7 @@ enum etm_impdef_type { > * @s_ex_level: Secure ELs where tracing is supported. > */ > struct etmv4_config { > - u32 mode; > + u64 mode; > u32 pe_sel; > u32 cfg; > u32 eventctrl0; > diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h > index 767076e07970..727dd27ba800 100644 > --- a/drivers/hwtracing/coresight/coresight-priv.h > +++ b/drivers/hwtracing/coresight/coresight-priv.h > @@ -39,6 +39,9 @@ > > #define ETM_MODE_EXCL_KERN BIT(30) > #define ETM_MODE_EXCL_USER BIT(31) > +#define ETM_MODE_EXCL_HOST BIT(32) > +#define ETM_MODE_EXCL_GUEST BIT(33) > + > struct cs_pair_attribute { > struct device_attribute attr; > u32 lo_off;
On 12/09/2023 18:37, Suzuki K Poulose wrote: > Hi James > > > On 04/09/2023 15:07, James Clark wrote: >> Currently the userspace and kernel filters for guests are never set, so >> no trace will be generated for them. Add it by writing to the guest >> filters when exclude_guest isn't set. By writing either E1TRE or E0TRE, >> filtering on either guest kernel or guest userspace is also supported. >> >> Since TRFCR_EL1 access is trapped, this can't be modified by the guest. >> >> This change also brings exclude_host support which is difficult to add >> as a separate commit without excess churn and resulting in no trace at >> all. >> >> Testing >> ======= >> >> The addresses were counted with the following: >> >> $ perf report -D | grep -Eo 'EL2|EL1|EL0' | sort | uniq -c >> >> Guest kernel only: >> >> $ perf record -e cs_etm//Gk -a -- true >> 535 EL1 >> 1 EL2 >> >> Guest user only (0 addresses expected because the guest OS hasn't reached >> userspace yet): >> >> $ perf record -e cs_etm//Gu -a -- true >> >> Host kernel only: >> >> $ perf record -e cs_etm//Hk -a -- true >> 3501 EL2 >> >> Host userspace only: >> >> $ perf record -e cs_etm//Hu -a -- true >> 408 EL0 >> 1 EL2 >> >> Reviewed-by: Mark Brown <broonie@kernel.org> (sysreg) >> Signed-off-by: James Clark <james.clark@arm.com> >> --- >> arch/arm64/tools/sysreg | 4 ++ >> .../coresight/coresight-etm4x-core.c | 51 ++++++++++++++++--- >> drivers/hwtracing/coresight/coresight-etm4x.h | 2 +- >> drivers/hwtracing/coresight/coresight-priv.h | 3 ++ >> 4 files changed, 53 insertions(+), 7 deletions(-) >> >> diff --git a/arch/arm64/tools/sysreg b/arch/arm64/tools/sysreg >> index 6ca7db69d6c9..cae9139b6c05 100644 >> --- a/arch/arm64/tools/sysreg >> +++ b/arch/arm64/tools/sysreg >> @@ -2531,3 +2531,7 @@ EndSysreg >> Sysreg TRFCR_EL2 3 4 1 2 1 >> Fields TRFCR_EL2 >> EndSysreg >> + >> +Sysreg TRFCR_EL12 3 5 1 2 1 >> +Fields TRFCR_ELx >> +EndSysreg >> diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c >> b/drivers/hwtracing/coresight/coresight-etm4x-core.c >> index 77b0271ce6eb..6c16a14d6fbe 100644 >> --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c >> +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c >> @@ -274,6 +274,18 @@ static void etm4x_prohibit_trace(struct >> etmv4_drvdata *drvdata) >> cpu_prohibit_trace(); >> } >> +static u64 etm4x_get_kern_user_filter(struct etmv4_drvdata *drvdata) >> +{ >> + u64 trfcr = drvdata->trfcr; >> + >> + if (drvdata->config.mode & ETM_MODE_EXCL_KERN) >> + trfcr &= ~TRFCR_ELx_ExTRE; >> + if (drvdata->config.mode & ETM_MODE_EXCL_USER) >> + trfcr &= ~TRFCR_ELx_E0TRE; >> + >> + return trfcr; >> +} >> + >> /* >> * etm4x_allow_trace - Allow CPU tracing in the respective ELs, >> * as configured by the drvdata->config.mode for the current >> @@ -286,18 +298,39 @@ static void etm4x_prohibit_trace(struct >> etmv4_drvdata *drvdata) >> */ >> static void etm4x_allow_trace(struct etmv4_drvdata *drvdata) >> { >> - u64 trfcr = drvdata->trfcr; >> + u64 trfcr; >> /* If the CPU doesn't support FEAT_TRF, nothing to do */ >> - if (!trfcr) >> + if (!drvdata->trfcr) >> return; >> - if (drvdata->config.mode & ETM_MODE_EXCL_KERN) >> - trfcr &= ~TRFCR_ELx_ExTRE; >> - if (drvdata->config.mode & ETM_MODE_EXCL_USER) >> - trfcr &= ~TRFCR_ELx_E0TRE; >> + if (drvdata->config.mode & ETM_MODE_EXCL_HOST) >> + trfcr = drvdata->trfcr & ~(TRFCR_ELx_ExTRE | TRFCR_ELx_E0TRE); >> + else >> + trfcr = etm4x_get_kern_user_filter(drvdata); >> write_trfcr(trfcr); >> + >> + /* >> + * Filters for EL1 and EL0 (when running a guest) are stored in >> + * TRFCR_EL1 so write it there for VHE. For nVHE, the filters in >> + * have to be re-applied when switching to the guest instead. >> + */ >> + if (!is_kernel_in_hyp_mode()) >> + return; >> + >> + if (drvdata->config.mode & ETM_MODE_EXCL_GUEST) >> + trfcr = drvdata->trfcr & ~(TRFCR_ELx_ExTRE | TRFCR_ELx_E0TRE); >> + else >> + trfcr = etm4x_get_kern_user_filter(drvdata); >> + >> + /* >> + * TRFCR_EL1 doesn't have CX and TRFCR_EL1.TS has no effect when >> TS is >> + * set in EL2 so mask them out. >> + */ >> + trfcr &= ~(TRFCR_ELx_TS_MASK | TRFCR_EL2_CX); >> + >> + write_sysreg_s(trfcr, SYS_TRFCR_EL12); > > > > While I agree with the intention of the patch, I am not sure if we > should do the TRFCR_EL1 (via TRFCR_EL12) updates from CoreSight driver. > > It might be much better to do it from the KVM for both VHE and nVHE. > For the next version of the nVHE change [1] I planned to re-write it just in terms of the guest trfcr register value rather than exclude_guest/exclude host. That means that it would be pretty easy to move the write to SYS_TRFCR_EL12 to the VHE implementation of kvm_etm_set_guest_trfcr(). So yes I can make this change I think it makes sense. [1]: https://lore.kernel.org/kvmarm/20230804101317.460697-1-james.clark@arm.com/ James > > Marc > > What are your thoughts on this ? > > Suzuki > > >> } >> #ifdef CONFIG_ETM4X_IMPDEF_FEATURE >> @@ -655,6 +688,12 @@ static int etm4_parse_event_config(struct >> coresight_device *csdev, >> if (attr->exclude_user) >> config->mode = ETM_MODE_EXCL_USER; >> + if (attr->exclude_host) >> + config->mode |= ETM_MODE_EXCL_HOST; >> + >> + if (attr->exclude_guest) >> + config->mode |= ETM_MODE_EXCL_GUEST; >> + >> /* Always start from the default config */ >> etm4_set_default_config(config); >> diff --git a/drivers/hwtracing/coresight/coresight-etm4x.h >> b/drivers/hwtracing/coresight/coresight-etm4x.h >> index 20e2e4cb7614..3f170599822f 100644 >> --- a/drivers/hwtracing/coresight/coresight-etm4x.h >> +++ b/drivers/hwtracing/coresight/coresight-etm4x.h >> @@ -841,7 +841,7 @@ enum etm_impdef_type { >> * @s_ex_level: Secure ELs where tracing is supported. >> */ >> struct etmv4_config { >> - u32 mode; >> + u64 mode; >> u32 pe_sel; >> u32 cfg; >> u32 eventctrl0; >> diff --git a/drivers/hwtracing/coresight/coresight-priv.h >> b/drivers/hwtracing/coresight/coresight-priv.h >> index 767076e07970..727dd27ba800 100644 >> --- a/drivers/hwtracing/coresight/coresight-priv.h >> +++ b/drivers/hwtracing/coresight/coresight-priv.h >> @@ -39,6 +39,9 @@ >> #define ETM_MODE_EXCL_KERN BIT(30) >> #define ETM_MODE_EXCL_USER BIT(31) >> +#define ETM_MODE_EXCL_HOST BIT(32) >> +#define ETM_MODE_EXCL_GUEST BIT(33) >> + >> struct cs_pair_attribute { >> struct device_attribute attr; >> u32 lo_off; >
diff --git a/arch/arm64/tools/sysreg b/arch/arm64/tools/sysreg index 6ca7db69d6c9..cae9139b6c05 100644 --- a/arch/arm64/tools/sysreg +++ b/arch/arm64/tools/sysreg @@ -2531,3 +2531,7 @@ EndSysreg Sysreg TRFCR_EL2 3 4 1 2 1 Fields TRFCR_EL2 EndSysreg + +Sysreg TRFCR_EL12 3 5 1 2 1 +Fields TRFCR_ELx +EndSysreg diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c index 77b0271ce6eb..6c16a14d6fbe 100644 --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c @@ -274,6 +274,18 @@ static void etm4x_prohibit_trace(struct etmv4_drvdata *drvdata) cpu_prohibit_trace(); } +static u64 etm4x_get_kern_user_filter(struct etmv4_drvdata *drvdata) +{ + u64 trfcr = drvdata->trfcr; + + if (drvdata->config.mode & ETM_MODE_EXCL_KERN) + trfcr &= ~TRFCR_ELx_ExTRE; + if (drvdata->config.mode & ETM_MODE_EXCL_USER) + trfcr &= ~TRFCR_ELx_E0TRE; + + return trfcr; +} + /* * etm4x_allow_trace - Allow CPU tracing in the respective ELs, * as configured by the drvdata->config.mode for the current @@ -286,18 +298,39 @@ static void etm4x_prohibit_trace(struct etmv4_drvdata *drvdata) */ static void etm4x_allow_trace(struct etmv4_drvdata *drvdata) { - u64 trfcr = drvdata->trfcr; + u64 trfcr; /* If the CPU doesn't support FEAT_TRF, nothing to do */ - if (!trfcr) + if (!drvdata->trfcr) return; - if (drvdata->config.mode & ETM_MODE_EXCL_KERN) - trfcr &= ~TRFCR_ELx_ExTRE; - if (drvdata->config.mode & ETM_MODE_EXCL_USER) - trfcr &= ~TRFCR_ELx_E0TRE; + if (drvdata->config.mode & ETM_MODE_EXCL_HOST) + trfcr = drvdata->trfcr & ~(TRFCR_ELx_ExTRE | TRFCR_ELx_E0TRE); + else + trfcr = etm4x_get_kern_user_filter(drvdata); write_trfcr(trfcr); + + /* + * Filters for EL1 and EL0 (when running a guest) are stored in + * TRFCR_EL1 so write it there for VHE. For nVHE, the filters in + * have to be re-applied when switching to the guest instead. + */ + if (!is_kernel_in_hyp_mode()) + return; + + if (drvdata->config.mode & ETM_MODE_EXCL_GUEST) + trfcr = drvdata->trfcr & ~(TRFCR_ELx_ExTRE | TRFCR_ELx_E0TRE); + else + trfcr = etm4x_get_kern_user_filter(drvdata); + + /* + * TRFCR_EL1 doesn't have CX and TRFCR_EL1.TS has no effect when TS is + * set in EL2 so mask them out. + */ + trfcr &= ~(TRFCR_ELx_TS_MASK | TRFCR_EL2_CX); + + write_sysreg_s(trfcr, SYS_TRFCR_EL12); } #ifdef CONFIG_ETM4X_IMPDEF_FEATURE @@ -655,6 +688,12 @@ static int etm4_parse_event_config(struct coresight_device *csdev, if (attr->exclude_user) config->mode = ETM_MODE_EXCL_USER; + if (attr->exclude_host) + config->mode |= ETM_MODE_EXCL_HOST; + + if (attr->exclude_guest) + config->mode |= ETM_MODE_EXCL_GUEST; + /* Always start from the default config */ etm4_set_default_config(config); diff --git a/drivers/hwtracing/coresight/coresight-etm4x.h b/drivers/hwtracing/coresight/coresight-etm4x.h index 20e2e4cb7614..3f170599822f 100644 --- a/drivers/hwtracing/coresight/coresight-etm4x.h +++ b/drivers/hwtracing/coresight/coresight-etm4x.h @@ -841,7 +841,7 @@ enum etm_impdef_type { * @s_ex_level: Secure ELs where tracing is supported. */ struct etmv4_config { - u32 mode; + u64 mode; u32 pe_sel; u32 cfg; u32 eventctrl0; diff --git a/drivers/hwtracing/coresight/coresight-priv.h b/drivers/hwtracing/coresight/coresight-priv.h index 767076e07970..727dd27ba800 100644 --- a/drivers/hwtracing/coresight/coresight-priv.h +++ b/drivers/hwtracing/coresight/coresight-priv.h @@ -39,6 +39,9 @@ #define ETM_MODE_EXCL_KERN BIT(30) #define ETM_MODE_EXCL_USER BIT(31) +#define ETM_MODE_EXCL_HOST BIT(32) +#define ETM_MODE_EXCL_GUEST BIT(33) + struct cs_pair_attribute { struct device_attribute attr; u32 lo_off;
Currently the userspace and kernel filters for guests are never set, so no trace will be generated for them. Add it by writing to the guest filters when exclude_guest isn't set. By writing either E1TRE or E0TRE, filtering on either guest kernel or guest userspace is also supported. Since TRFCR_EL1 access is trapped, this can't be modified by the guest. This change also brings exclude_host support which is difficult to add as a separate commit without excess churn and resulting in no trace at all. Testing ======= The addresses were counted with the following: $ perf report -D | grep -Eo 'EL2|EL1|EL0' | sort | uniq -c Guest kernel only: $ perf record -e cs_etm//Gk -a -- true 535 EL1 1 EL2 Guest user only (0 addresses expected because the guest OS hasn't reached userspace yet): $ perf record -e cs_etm//Gu -a -- true Host kernel only: $ perf record -e cs_etm//Hk -a -- true 3501 EL2 Host userspace only: $ perf record -e cs_etm//Hu -a -- true 408 EL0 1 EL2 Reviewed-by: Mark Brown <broonie@kernel.org> (sysreg) Signed-off-by: James Clark <james.clark@arm.com> --- arch/arm64/tools/sysreg | 4 ++ .../coresight/coresight-etm4x-core.c | 51 ++++++++++++++++--- drivers/hwtracing/coresight/coresight-etm4x.h | 2 +- drivers/hwtracing/coresight/coresight-priv.h | 3 ++ 4 files changed, 53 insertions(+), 7 deletions(-)