Message-ID: <07f52870-3a49-404c-af64-00fef9e5bb94@amd.com>
Date: Thu, 13 Nov 2025 16:39:28 -0600
From: "Bowman, Terry" <terry.bowman@....com>
To: Alison Schofield <alison.schofield@...el.com>
Cc: dave@...olabs.net, jonathan.cameron@...wei.com, dave.jiang@...el.com,
 dan.j.williams@...el.com, bhelgaas@...gle.com, shiju.jose@...wei.com,
 ming.li@...omail.com, Smita.KoralahalliChannabasappa@....com,
 rrichter@....com, dan.carpenter@...aro.org,
 PradeepVineshReddy.Kodamati@....com, lukas@...ner.de,
 Benjamin.Cheatham@....com, sathyanarayanan.kuppuswamy@...ux.intel.com,
 linux-cxl@...r.kernel.org, alucerop@....com, ira.weiny@...el.com,
 linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org
Subject: Re: [RESEND v13 17/25] cxl: Introduce cxl_pci_drv_bound() to check
 for bound driver



On 11/13/2025 3:42 PM, Alison Schofield wrote:
> On Tue, Nov 11, 2025 at 12:33:53AM -0800, Alison Schofield wrote:
>> On Tue, Nov 04, 2025 at 11:02:57AM -0600, Terry Bowman wrote:
>>> CXL devices handle protocol errors via driver-specific callbacks rather
>>> than the generic pci_driver::err_handlers by default. The callbacks are
>>> implemented in the cxl_pci driver and are not part of struct pci_driver, so
>>> cxl_core must verify that a device is actually bound to the cxl_pci
>>> module's driver before invoking the callbacks (the device could be bound
>>> to another driver, e.g. VFIO).
>>>
>>> However, cxl_core can not reference symbols in the cxl_pci module because
>>> it creates a circular dependency. This prevents cxl_core from checking the
>>> EP's bound driver and calling the callbacks.
>>>
>>> To fix this, move drivers/cxl/pci.c into drivers/cxl/core/pci_drv.c and
>>> build it as part of the cxl_core module. Compile into cxl_core using
>>> CXL_PCI and CXL_CORE Kconfig dependencies. This removes the standalone
>>> cxl_pci module, consolidates the cxl_pci driver code into cxl_core, and
>>> eliminates the circular dependency so cxl_core can safely perform
>>> bound-driver checks and invoke the CXL PCI callbacks.
>>>
>>> Introduce cxl_pci_drv_bound() to return boolean depending on if the PCI EP
>>> parameter is bound to a CXL driver instance. This will be used in future
>>> patch when dequeuing work from the kfifo.
>> This one was troublesome in cxl-test, more circular dependencies.
>> I noticed you and GregP chatting about it, so I simply remove it from
>> the set for now (made all callsites true).
>>
>> With it gone, the set builds cxl-test and passes the test suite.
>> I'll watch what happens with this one, and can take another look at
>> the cxl-test issues if they persist.
> Hi Terry -
>
> I took another look, suspecting the circle issue started with the
> move of pci.c into the core, and not necessarily your new additions.
> There are two functions that are wrapped in cxl-test and now with
> this move are being called from the core and creating the 'circle':
>
> cxl_await_media_ready()
> cxl_rcd_component_reg_phys()
>
> Both those need the 'restrict' method, like for Patch 14.
>
> Once that is resolved, the new function cxl_pci_drv_bound()
> seems like it needs mocking and will require the same treatment.
>
> Suggest doing it in separate patches. First patch does the move
> and the cxl-test work.  Then a second patch adds the new function
> and it's cxl-test support.
>
> --Alison
>

Hi Alison,

Thanks for finding the issue. I'll start on the fix using the changes 
you described.

-Terry


>> A bit below...
>>
>> snip
>>
>>> diff --git a/drivers/cxl/pci.c b/drivers/cxl/core/pci_drv.c
>>> similarity index 99%
>>> rename from drivers/cxl/pci.c
>>> rename to drivers/cxl/core/pci_drv.c
>>> index bd95be1f3d5c..06f2fd993cb0 100644
>>> --- a/drivers/cxl/pci.c
>>> +++ b/drivers/cxl/core/pci_drv.c
>> Needs:
>> +#include "core.h"
>>
>> Compiler is warning: no previous prototypes.
>>
>> snip
>>

Powered by blists - more mailing lists