Message-ID: <4b41e3e2-ebac-48c3-b1a3-ce24a5bed475@linux.alibaba.com>
Date: Fri, 25 Jul 2025 10:31:45 +0800
From: Shuai Xue <xueshuai@...ux.alibaba.com>
To: Steven Rostedt <rostedt@...dmis.org>, kernel test robot <lkp@...el.com>,
 Bjorn Helgaas <bhelgaas@...gle.com>
Cc: lukas@...ner.de, linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org,
 linux-edac@...r.kernel.org, linux-trace-kernel@...r.kernel.org,
 helgaas@...nel.org, ilpo.jarvinen@...ux.intel.com, mattc@...estorage.com,
 Jonathan.Cameron@...wei.com, oe-kbuild-all@...ts.linux.dev,
 tony.luck@...el.com, bp@...en8.de, mhiramat@...nel.org,
 mathieu.desnoyers@...icios.com, oleg@...hat.com, naveen@...nel.org,
 davem@...emloft.net, anil.s.keshavamurthy@...el.com, mark.rutland@....com,
 peterz@...radead.org, tianruidong@...ux.alibaba.com
Subject: Re: [PATCH v9 2/2] PCI: trace: Add a RAS tracepoint to monitor link
 speed changes

Hi, Steven,

在 2025/7/24 22:06, Steven Rostedt 写道:
> On Thu, 24 Jul 2025 03:30:17 +0800
> kernel test robot <lkp@...el.com> wrote:
> 
>> url:    https://github.com/intel-lab-lkp/linux/commits/Shuai-Xue/PCI-trace-Add-a-generic-RAS-tracepoint-for-hotplug-event/20250723-113454
>> base:   https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git for-linus
>> patch link:    https://lore.kernel.org/r/20250723033108.61587-3-xueshuai%40linux.alibaba.com
>> patch subject: [PATCH v9 2/2] PCI: trace: Add a RAS tracepoint to monitor link speed changes
>> config: sparc-sparc64_defconfig (https://download.01.org/0day-ci/archive/20250724/202507240322.nJGdyXsy-lkp@intel.com/config)
>> compiler: sparc64-linux-gcc (GCC) 15.1.0
>> reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20250724/202507240322.nJGdyXsy-lkp@intel.com/reproduce)
>>
>> If you fix the issue in a separate patch/commit (i.e. not just a new version of
>> the same patch/commit), kindly add following tags
>> | Reported-by: kernel test robot <lkp@...el.com>
>> | Closes: https://lore.kernel.org/oe-kbuild-all/202507240322.nJGdyXsy-lkp@intel.com/
>>
>> All errors (new ones prefixed by >>):
>>
>>     sparc64-linux-ld: drivers/pci/probe.o: in function `pcie_update_link_speed':
>>>> probe.c:(.text+0x370): undefined reference to `__tracepoint_pcie_link_event'
>>>> sparc64-linux-ld: probe.c:(.text+0x37c): undefined reference to `__tracepoint_pcie_link_event'
>>>> sparc64-linux-ld: probe.c:(.text+0x3dc): undefined reference to `__traceiter_pcie_link_event'
> 
> The config has:
> 
> # CONFIG_CPU_HOTPLUG_STATE_CONTROL is not set
> 
> Which looks to me from the first patch, would enable the code that
> defines the trace events via the:
> 
> +#define CREATE_TRACE_POINTS
> +#include <trace/events/pci.h>
> 
> Thus, without compiling that file, the tracepoints would not be created
> and you would get the above errors.

Thanks for the detailed analysis! You're absolutely right about the root
cause.

The issue is that I have the tracepoint usage in drivers/pci/probe.c
(which is always compiled), but the tracepoint definition with:

#define CREATE_TRACE_POINTS
#include<trace/events/pci.h>

is in drivers/pci/hotplug/pciehp_ctrl.c that's conditionally compiled
based on CONFIG_HOTPLUG_PCI_PCIE=y, creating the undefined reference
when that config is not set.

I can see a few ways to fix this:

- Move tracepoint definition to always-compiled code - Create a dedicated drivers/pci/trace.c that always defines the PCI tracepoints
- Make the tracepoint usage conditional - Wrap the trace calls with the same config dependency
- Create a PCI-specific trace config - Add a CONFIG_PCI_TRACE option

I'm leaning toward option 1 (dedicated trace.c file) as it's the
cleanest approach and follows the pattern used in other subsystems. Does
that sound reasonable to you?

> 
> -- Steve

Thanks for valuable feedback.

Best Regards,
Shuai

Powered by blists - more mailing lists