Received-SPF: pass (google.com: domain of linux-kernel+bounces-168388-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Fri, 3 May 2024 17:53:05 -0500
From: Bjorn Helgaas <helgaas@kernel.org>
To: Ilpo =?utf-8?B?SsOkcnZpbmVu?= <ilpo.jarvinen@linux.intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>, linux-pci@vger.kernel.org,
	Mahesh J Salgaonkar <mahesh@linux.ibm.com>,
	Oliver O'Halloran <oohall@gmail.com>,
	Lukas Wunner <lukas@wunner.de>, linux-kernel@vger.kernel.org,
	linuxppc-dev@lists.ozlabs.org
Subject: Re: [PATCH v3 1/2] PCI: Add TLP Prefix reading into
 pcie_read_tlp_log()
Message-ID: <20240503225305.GA1609388@bhelgaas>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20240412133635.3831-2-ilpo.jarvinen@linux.intel.com>

On Fri, Apr 12, 2024 at 04:36:34PM +0300, Ilpo Järvinen wrote:
> pcie_read_tlp_log() handles only 4 TLP Header Log DWORDs but TLP Prefix
> Log (PCIe r6.1 secs 7.8.4.12 & 7.9.14.13) may also be present.
> 
> Generalize pcie_read_tlp_log() and struct pcie_tlp_log to handle also
> TLP Prefix Log. The layout of relevant registers in AER and DPC
> Capability is not identical because the offsets of TLP Header Log and
> TLP Prefix Log vary so the callers must pass the offsets to
> pcie_read_tlp_log().

I think the layouts of the Header Log and the TLP Prefix Log *are*
identical, but they are at different offsets in the AER Capability vs
the DPC Capability.  Lukas and I have both stumbled over this.

Similar and more comments at:
https://lore.kernel.org/r/20240322193011.GA701027@bhelgaas

> Convert eetlp_prefix_path into integer called eetlp_prefix_max and
> make is available also when CONFIG_PCI_PASID is not configured to
> be able to determine the number of E-E Prefixes.

s/make is/make it/

I think this could be a separate patch.

> --- a/include/linux/aer.h
> +++ b/include/linux/aer.h
> @@ -20,6 +20,7 @@ struct pci_dev;
>  
>  struct pcie_tlp_log {
>  	u32 dw[4];
> +	u32 prefix[4];
>  };
>  
>  struct aer_capability_regs {
> @@ -37,7 +38,9 @@ struct aer_capability_regs {
>  	u16 uncor_err_source;
>  };
>  
> -int pcie_read_tlp_log(struct pci_dev *dev, int where, struct pcie_tlp_log *log);
> +int pcie_read_tlp_log(struct pci_dev *dev, int where, int where2,
> +		      unsigned int tlp_len, struct pcie_tlp_log *log);
> +unsigned int aer_tlp_log_len(struct pci_dev *dev);

I think it was a mistake to expose pcie_read_tlp_log() outside
drivers/pci, and I don't think we should expose aer_tlp_log_len()
either.

We might be stuck with exposing struct pcie_tlp_log since it looks
like ras_event.h uses it.

Bjorn