On Thu, Jul 16, 2026 at 01:48:15PM +0100, Suzuki K Poulose wrote:
On 29/06/2026 09:59, Yeoreum Yun wrote:
The current ETMx configuration via sysfs can lead to the following inconsistencies:
If a configuration is modified via sysfs while a perf session is active, the running configuration may differ between before a sched-out and after a subsequent sched-in.
If a perf session and sysfs session tries to enable concurrently, configuration from configfs could be corrupted (etm4).
There is chance to corrupt drvdata->config if perf session tries to enabled among handling cscfg_csdev_disable_active_config() in etm4_disable_sysfs() (etm4).
To resolve these inconsistencies, the configuration should be separated into:
- active_config, which is applied configuration for the current session
- config, which stores the settings configured via sysfs.
and apply configuration from configfs after taking a mode.
Also, This patch set includes some small fixes:
- missing trace id release in etm4x.
- underflow issue for nrseqstate.
- wrong check in etm4x_sspcicrn_present().
- missing call of cscfg_csdev_disable_active_config()
This patch based on coresight tree's next
Please could you move all the "Fixes" to the top of the series, that way it is cleaner to backport for the stable kernels
Okay. but the fix which based on the cleanup wouldn't be top of it. I'll deploy below patches first:
coresight: etm3x: change drvdata->spinlock type to raw_spin_lock_t coresight: etm4x: fix leaked trace id coresight: etm4x: fix underflow for usage of (nrseqstate - 1) coresight: etm4x: fix wrong check of etm4x_sspcicrn_present()
Sorry. Ignore this please, I've moved all fix in the top properly.
Thanks!
[...]