Linux-PCI Archive on lore.kernel.org
 help / color / Atom feed
From: "Sean V Kelley" <sean.v.kelley@intel.com>
To: "Jonathan Cameron" <Jonathan.Cameron@Huawei.com>
Cc: bhelgaas@google.com, rjw@rjwysocki.net, tony.luck@intel.com,
	"Raj, Ashok" <ashok.raj@intel.com>,
	sathyanarayanan.kuppuswamy@linux.intel.com,
	linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [RFC PATCH 0/9] Add RCEC handling to PCI/AER
Date: Mon, 27 Jul 2020 07:56:35 -0700
Message-ID: <E04FFB14-95BA-4D04-BC53-037B5EB2FE80@intel.com> (raw)
In-Reply-To: <20200727133736.00001066@Huawei.com>

Hi Jonathan,

On 27 Jul 2020, at 5:37, Jonathan Cameron wrote:

> On Fri, 24 Jul 2020 10:22:14 -0700
> Sean V Kelley <sean.v.kelley@intel.com> wrote:
>
>> Root Complex Event Collectors (RCEC) provide support for terminating 
>> error
>> and PME messages from Root Complex Integrated Endpoints (RCiEPs).  An 
>> RCEC
>> resides on a Bus in the Root Complex. Multiple RCECs can in fact 
>> reside on
>> a single bus. An RCEC will explicitly declare supported RCiEPs 
>> through the
>> Root Complex Endpoint Association Extended Capability.
>>
>> (See PCIe 5.0-1, sections 1.3.2.3 (RCiEP), and 7.9.10 (RCEC Ext. 
>> Cap.))
>>
>> The kernel lacks handling for these RCECs and the error messages 
>> received
>> from their respective associated RCiEPs. More recently, a new CPU
>> interconnect, Compute eXpress Link (CXL) depends on RCEC capabilities 
>> for
>> purposes of error messaging from CXL 1.1 supported RCiEP devices.
>>
>> DocLink: https://www.computeexpresslink.org/
>>
>> This use case is not limited to CXL. Existing hardware today includes
>> support for RCECs, such as the Denverton microserver product
>> family. Future hardware will be forthcoming.
>>
>> (See Intel Document, Order number: 33061-003US)
>>
>> So services such as AER or PME could be associated with an RCEC 
>> driver.
>> In the case of CXL, if an RCiEP (i.e., CXL 1.1 device) is associated 
>> with a
>> platform's RCEC it shall signal PME and AER error conditions through 
>> that
>> RCEC.
>>
>> Towards the above use cases, add the missing RCEC class and extend 
>> the
>> PCIe Root Port and service drivers to allow association of RCiEPs to 
>> their
>> respective parent RCEC and facilitate handling of terminating error 
>> and PME
>> messages.
>
> Silly question number 1.  Why an RFC? I always find it helps to 
> highlight which
> bits you are unsure on / want particular input on.

I suppose it was because we were continuing the conversation from 
discussion on the mailing list.
And I was not fully sure about the impact to your use case.  No worries, 
I will remove it.

>
> Otherwise, I've left the PME and error injection patches as I don't 
> really know
> anything about those two paths.
>
> I'll fire up my APEI etc test VMs in the nex day or so and report back 
> if there
> any problems with that case (fairly sure there is one in patch 6, 
> highlighted in
> review but it is possible I've missed others.
>
> It all seems to have come together rather simpler than I was expecting 
> which is
> great!

Sounds good, looking forward to your feedback.

Sean

>
> Thanks,
>
> Jonathan
>
>
>>
>> TESTING:
>>
>>    Results:
>>     1) Show RCiEPs which are associated with RCECs:
>> 	Run dmesg | grep "RCiEP"
>> 	Log:
>> 	[    8.981698] pcieport 0000:e8:00.4: RCiEP(under an RCEC) 
>> 0000:e8:01.0
>> 	[    8.988830] pcieport 0000:e8:00.4: RCiEP(under an RCEC) 
>> 0000:e8:02.0
>> 	[    8.995956] pcieport 0000:e8:00.4: RCiEP(under an RCEC) 
>> 0000:e9:00.0
>> 	[    9.023034] pcieport 0000:ed:00.4: RCiEP(under an RCEC) 
>> 0000:ed:01.0
>> 	[    9.030159] pcieport 0000:ed:00.4: RCiEP(under an RCEC) 
>> 0000:ed:02.0
>> 	[    9.037282] pcieport 0000:ed:00.4: RCiEP(under an RCEC) 
>> 0000:ee:00.0
>> 	[    9.064294] pcieport 0000:f2:00.4: RCiEP(under an RCEC) 
>> 0000:f2:01.0
>> 	[    9.071409] pcieport 0000:f2:00.4: RCiEP(under an RCEC) 
>> 0000:f2:02.0
>> 	[    9.078526] pcieport 0000:f2:00.4: RCiEP(under an RCEC) 
>> 0000:f3:00.0
>> 	[    9.105535] pcieport 0000:f7:00.4: RCiEP(under an RCEC) 
>> 0000:f7:01.0
>> 	[    9.112652] pcieport 0000:f7:00.4: RCiEP(under an RCEC) 
>> 0000:f7:02.0
>> 	[    9.119774] pcieport 0000:f7:00.4: RCiEP(under an RCEC) 
>> 0000:f8:00.0
>>
>>     2) Inject a correctable error to the RCiEP 0000:e9:00.0
>> 	Run ./aer_inject <a parameter file as below>:
>> 	AER
>> 	PCI_ID 0000:e9:00.0
>> 	COR_STATUS BAD_TLP
>> 	HEADER_LOG 0 1 2 3
>>
>> 	Log:
>> 	[  253.248362] pcieport 0000:e8:00.4: aer_inject: Injecting errors 
>> 00000040/00000000 into device 0000:e9:00.0
>> 	[  253.260656] pcieport 0000:e8:00.4: AER: Corrected error received: 
>> 0000:e9:00.0
>> 	[  253.269919] pci 0000:e9:00.0: AER: PCIe Bus Error: 
>> severity=Corrected, type=Data Link Layer, (Receiver ID)
>> 	[  253.282549] pci 0000:e9:00.0: AER:   device [8086:4940] error 
>> status/mask=00000040/00002000
>> 	[  253.293937] pci 0000:e9:00.0: AER:    [ 6] BadTLP
>>
>>     3) Inject a non-fatal error to the RCiEP 0000:e8:01.0
>> 	Run ./aer_inject <a parameter file as below>:
>> 	AER
>> 	PCI_ID 0000:e8:01.0
>> 	UNCOR_STATUS COMP_ABORT
>> 	HEADER_LOG 0 1 2 3
>>
>> 	Log:
>> 	[  288.405326] pcieport 0000:e8:00.4: aer_inject: Injecting errors 
>> 00000000/00008000 into device 0000:e8:01.0
>> 	[  288.416881] pcieport 0000:e8:00.4: AER: Uncorrected (Non-Fatal) 
>> error received: 0000:e8:01.0
>> 	[  288.427487] igen6_edac 0000:e8:01.0: AER: PCIe Bus Error: 
>> severity=Uncorrected (Non-Fatal), type=Transaction Layer, (Completer 
>> ID)
>> 	[  288.442098] igen6_edac 0000:e8:01.0: AER:   device [8086:0b25] 
>> error status/mask=00008000/00100000
>> 	[  288.452869] igen6_edac 0000:e8:01.0: AER:    [15] CmpltAbrt
>> 	[  288.461118] igen6_edac 0000:e8:01.0: AER:   TLP Header: 00000000 
>> 00000001 00000002 00000003
>> 	[  288.471192] igen6_edac 0000:e8:01.0: AER: device recovery 
>> successful
>>
>>     4) Inject a fatal error to the RCiEP 0000:ed:01.0
>> 	Run ./aer_inject <a parameter file as below>:
>> 	AER
>> 	PCI_ID 0000:ed:01.0
>> 	UNCOR_STATUS MALF_TLP
>> 	HEADER_LOG 0 1 2 3
>>
>> 	Log:
>> 	[  535.537281] pcieport 0000:ed:00.4: aer_inject: Injecting errors 
>> 00000000/00040000 into device 0000:ed:01.0
>> 	[  535.551911] pcieport 0000:ed:00.4: AER: Uncorrected (Fatal) error 
>> received: 0000:ed:01.0
>> 	[  535.561556] igen6_edac 0000:ed:01.0: AER: PCIe Bus Error: 
>> severity=Uncorrected (Fatal), type=Inaccessible, (Unregistered Agent 
>> ID)
>> 	[  535.684964] igen6_edac 0000:ed:01.0: AER: device recovery 
>> successful
>>
>>
>> Jonathan Cameron (1):
>>   PCI/AER: Extend AER error handling to RCECs
>>
>> Qiuxu Zhuo (6):
>>   pci_ids: Add class code and extended capability for RCEC
>>   PCI: Extend Root Port Driver to support RCEC
>>   PCI/portdrv: Add pcie_walk_rcec() to walk RCiEPs associated with 
>> RCEC
>>   PCI/AER: Apply function level reset to RCiEP on fatal error
>>   PCI: Add 'rcec' field to pci_dev for associated RCiEPs
>>   PCI/AER: Add RCEC AER error injection support
>>
>> Sean V Kelley (2):
>>   PCI/AER: Add RCEC AER handling
>>   PCI/PME: Add RCEC PME handling
>>
>>  drivers/pci/pcie/aer.c          | 43 ++++++++++++-----
>>  drivers/pci/pcie/aer_inject.c   |  5 +-
>>  drivers/pci/pcie/err.c          | 85 
>> +++++++++++++++++++++++++++------
>>  drivers/pci/pcie/pme.c          | 15 ++++--
>>  drivers/pci/pcie/portdrv.h      |  2 +
>>  drivers/pci/pcie/portdrv_core.c | 82 +++++++++++++++++++++++++++++++
>>  drivers/pci/pcie/portdrv_pci.c  | 20 +++++++-
>>  include/linux/pci.h             |  3 ++
>>  include/linux/pci_ids.h         |  1 +
>>  include/uapi/linux/pci_regs.h   |  7 +++
>>  10 files changed, 232 insertions(+), 31 deletions(-)
>>
>> --
>> 2.27.0
>>

      reply index

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-24 17:22 Sean V Kelley
2020-07-24 17:22 ` [RFC PATCH 1/9] pci_ids: Add class code and extended capability for RCEC Sean V Kelley
2020-07-27 10:00   ` Jonathan Cameron
2020-07-27 10:21     ` Jonathan Cameron
2020-07-27 15:22       ` Sean V Kelley
2020-07-24 17:22 ` [RFC PATCH 2/9] PCI: Extend Root Port Driver to support RCEC Sean V Kelley
2020-07-27 12:30   ` Jonathan Cameron
2020-07-27 15:05     ` Sean V Kelley
2020-07-24 17:22 ` [RFC PATCH 3/9] PCI/portdrv: Add pcie_walk_rcec() to walk RCiEPs associated with RCEC Sean V Kelley
2020-07-27 10:49   ` Jonathan Cameron
2020-07-27 15:21     ` Sean V Kelley
2020-07-24 17:22 ` [RFC PATCH 4/9] PCI/AER: Extend AER error handling to RCECs Sean V Kelley
2020-07-27 11:00   ` Jonathan Cameron
2020-07-27 14:58     ` Sean V Kelley
2020-07-27 14:04   ` Jonathan Cameron
2020-07-27 15:00     ` Sean V Kelley
2020-07-24 17:22 ` [RFC PATCH 5/9] PCI/AER: Apply function level reset to RCiEP on fatal error Sean V Kelley
2020-07-27 11:17   ` Jonathan Cameron
2020-07-28 13:27     ` Zhuo, Qiuxu
2020-07-28 16:14       ` Sean V Kelley
2020-07-28 17:02         ` Jonathan Cameron
2020-07-28 17:42           ` Sean V Kelley
2020-07-24 17:22 ` [RFC PATCH 6/9] PCI: Add 'rcec' field to pci_dev for associated RCiEPs Sean V Kelley
2020-07-27 11:23   ` Jonathan Cameron
2020-07-27 15:39     ` Sean V Kelley
2020-07-27 16:11     ` Jonathan Cameron
2020-07-27 16:28       ` Sean V Kelley
2020-07-24 17:22 ` [RFC PATCH 7/9] PCI/AER: Add RCEC AER handling Sean V Kelley
2020-07-27 12:22   ` Jonathan Cameron
2020-07-27 15:19     ` Sean V Kelley
2020-07-27 17:14       ` Jonathan Cameron
2020-07-24 17:22 ` [RFC PATCH 8/9] PCI/PME: Add RCEC PME handling Sean V Kelley
2020-08-04  8:35   ` Jay Fang
2020-08-04  9:47     ` Jonathan Cameron
2020-07-24 17:22 ` [RFC PATCH 9/9] PCI/AER: Add RCEC AER error injection support Sean V Kelley
2020-07-27 12:37 ` [RFC PATCH 0/9] Add RCEC handling to PCI/AER Jonathan Cameron
2020-07-27 14:56   ` Sean V Kelley [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=E04FFB14-95BA-4D04-BC53-037B5EB2FE80@intel.com \
    --to=sean.v.kelley@intel.com \
    --cc=Jonathan.Cameron@Huawei.com \
    --cc=ashok.raj@intel.com \
    --cc=bhelgaas@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=rjw@rjwysocki.net \
    --cc=sathyanarayanan.kuppuswamy@linux.intel.com \
    --cc=tony.luck@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-PCI Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-pci/0 linux-pci/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-pci linux-pci/ https://lore.kernel.org/linux-pci \
		linux-pci@vger.kernel.org
	public-inbox-index linux-pci

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-pci


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git