[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <68952ab060b6d_cff9910033@dwillia2-xfh.jf.intel.com.notmuch>
Date: Thu, 7 Aug 2025 15:37:36 -0700
From: <dan.j.williams@...el.com>
To: Bjorn Helgaas <helgaas@...nel.org>, Dan Williams
<dan.j.williams@...el.com>
CC: <linux-coco@...ts.linux.dev>, <linux-pci@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, <bhelgaas@...gle.com>, <aik@....com>,
<lukas@...ner.de>, Yilun Xu <yilun.xu@...el.com>, Jonathan Cameron
<Jonathan.Cameron@...wei.com>, Aneesh Kumar K.V <aneesh.kumar@...nel.org>
Subject: Re: [PATCH v4 02/10] PCI/IDE: Enumerate Selective Stream IDE
capabilities
Bjorn Helgaas wrote:
> On Thu, Jul 17, 2025 at 11:33:50AM -0700, Dan Williams wrote:
> > Link encryption is a new PCIe feature enumerated by "PCIe 6.2 section
> > 7.9.26 IDE Extended Capability".
>
> > +++ b/drivers/pci/ide.c
> > @@ -0,0 +1,93 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/* Copyright(c) 2024 Intel Corporation. All rights reserved. */
> > +
> > +/* PCIe 6.2 section 6.33 Integrity & Data Encryption (IDE) */
> > +
> > +#define dev_fmt(fmt) "PCI/IDE: " fmt
> > +#include <linux/pci.h>
> > +#include <linux/bitfield.h>
>
> Trend is to alphabetize these. And I think there should be more
> #includes here instead of using other things pulled in indirectly:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submit-checklist.rst?id=v6.16#n17
In this case I think it was only missing a:
#include <linux/pci_regs.h>
...but more includes are needed in follow-on patches. Added those and
alphabetized.
>
> > +++ b/include/uapi/linux/pci_regs.h
>
> > +#define PCI_IDE_CAP_ALG_MASK __GENMASK(12, 8) /* Supported Algorithms */
> > +#define PCI_IDE_CAP_ALG_AES_GCM_256 0 /* AES-GCM 256 key size, 96b MAC */
> > +#define PCI_IDE_CAP_LINK_TC_NUM_MASK __GENMASK(15, 13) /* Link IDE TCs */
> > +#define PCI_IDE_CAP_SEL_NUM_MASK __GENMASK(23, 16)/* Supported Selective IDE Streams */
>
> I'm totally OK with dropping the "_MASK" suffix since I think uses are
> completely readable without it, especially with __GENMASK()/FIELD_GET()/
> FIELD_PREP().
Sounds good, and helps with the column width pressure. There might be
isolated cases of "mask vs value" confusion, but I think proximity to
FIELD_PREP()/FIELD_GET(), like you say, makes this clear.
Powered by blists - more mailing lists