答复: [PATCH 1/3] perf/x86/amd/uncore: Add common PMU helper functions

[Qi Liu]

-----邮件原件-----
发件人: Peter Zijlstra <peterz@xxxxxxxxxxxxx> 
发送时间: 2026年7月1日 15:00
收件人: Qi Liu <liuqi@xxxxxxxx>
抄送: mingo@xxxxxxxxxx; namhyung@xxxxxxxxxx; alexander.shishkin@xxxxxxxxxxxxxxx; jolsa@xxxxxxxxxx; x86@xxxxxxxxxx; linux-perf-users@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Jing Li <lijing@xxxxxxxx>; Zhenglang Hu <huzhenglang@xxxxxxxx>
主题: Re: [PATCH 1/3] perf/x86/amd/uncore: Add common PMU helper functions

On Wed, Jul 01, 2026 at 03:44:11AM +0000, Qi Liu wrote:
> Add common helper functions for AMD-family uncore PMU handling.
> 
> The helpers cover event initialization, counter allocation, counter 
> read/update, event start/stop and per-CPU context management.  These 
> paths are not tied to a specific uncore unit and can be reused by 
> drivers with a similar uncore PMU programming model.
> 
> Signed-off-by: Qi Liu <liuqi@xxxxxxxx>
> Tested-by: Zhenglang Hu <huzhenglang@xxxxxxxx>

> +EXPORT_SYMBOL_GPL(uncore_common_set_update_interval);
> +EXPORT_SYMBOL_GPL(event_to_uncore_common_pmu);
> +EXPORT_SYMBOL_GPL(uncore_common_attr_group);
> +EXPORT_SYMBOL_GPL(uncore_common_start_hrtimer);
> +EXPORT_SYMBOL_GPL(uncore_common_read);
> +EXPORT_SYMBOL_GPL(uncore_common_start);
> +EXPORT_SYMBOL_GPL(uncore_common_stop);
> +EXPORT_SYMBOL_GPL(uncore_common_event_init);
> +EXPORT_SYMBOL_GPL(uncore_common_add);
> +EXPORT_SYMBOL_GPL(uncore_common_del);
> +EXPORT_SYMBOL_GPL(uncore_common_ctx_init);
> +EXPORT_SYMBOL_GPL(uncore_common_ctx_free);
> +EXPORT_SYMBOL_GPL(uncore_common_ctx_move);

I like the idea, but I don't particularly like the namespace.
'uncore_common' is too generic, there are a ton of uncore drivers, even on x86 and these functions are specific to the AMD family. That is, they aren't really common at all.

Hi Peter,

Thanks for your reply!

Yes, that namespace is a fair point ― "common" is too broad. Here are two options:
A) amd_uncore_* ― The prefix describes the programming model origin, not the vendor running it.
B) amd_family_uncore_* ― makes it explicit that this is AMD-family uncore infrastructure.

Do you have a preference?  Or something else entirely?

Thanks,
Qi Liu