netdev - Re: [PATCH net-next 3/7] qed: Add support for RoCE hw init

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 7 Oct 2022 10:48:30 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Yuval Mintz <Yuval.Mintz@...iumnetworks.com>
Cc:     netdev@...r.kernel.org, davem@...emloft.net,
        linux-rdma@...r.kernel.org, Ram.Amrani@...iumnetworks.com,
        Michal.Kalderon@...iumnetworks.com, Ariel.Elior@...iumnetworks.com,
        dledford@...hat.com, Manish Chopra <manishc@...vell.com>,
        linux-pci@...r.kernel.org
Subject: Re: [PATCH net-next 3/7] qed: Add support for RoCE hw init

[+cc Manish, linux-pci]

On Sat, Oct 01, 2016 at 09:59:57PM +0300, Yuval Mintz wrote:
> From: Ram Amrani <Ram.Amrani@...iumnetworks.com>
> 
> This adds the backbone required for the various HW initalizations
> which are necessary for the qedr driver - FW notification, resource
> initializations, etc.
> ...

> diff --git a/drivers/net/ethernet/qlogic/qed/qed_roce.c b/drivers/net/ethernet/qlogic/qed/qed_roce.c
> ...
> +	/* Check atomic operations support in PCI configuration space. */
> +	pci_read_config_dword(cdev->pdev,
> +			      cdev->pdev->pcie_cap + PCI_EXP_DEVCTL2,
> +			      &pci_status_control);
> +
> +	if (pci_status_control & PCI_EXP_DEVCTL2_LTR_EN)
> +		SET_FIELD(dev->dev_caps, QED_RDMA_DEV_CAP_ATOMIC_OP, 1);

I don't understand this.

  1) PCI_EXP_DEVCTL2 is a 16-bit register ("word"), not a 32-bit one
  ("dword").

  2) QED_RDMA_DEV_CAP_ATOMIC_OP is set here but is not read anywhere
  in this patch.  Is it used by the qed device itself?

  3) PCI_EXP_DEVCTL2_LTR_EN is for Latency Tolerance Reporting and is
  not related to atomic ops.  I don't know what
  QED_RDMA_DEV_CAP_ATOMIC_OP means, but possibly one of these was
  intended instead?

    - PCI_EXP_DEVCAP2_ATOMIC_COMP32 means the device supports 32-bit
      AtomicOps as a completer.
    - PCI_EXP_DEVCAP2_ATOMIC_COMP64 means the device supports 64-bit
      AtomicOps as a completer.
    - PCI_EXP_DEVCAP2_ATOMIC_COMP128 means the device supports 128-bit
      AtomicOps as a completer.
    - PCI_EXP_DEVCTL2_ATOMIC_REQ means the device is allowed to
      initiate AtomicOps.

(This code is now in qed_rdma.c)