Alpha arch development list
 help / color / mirror / Atom feed
From: "Krzysztof Wilczyński" <kwilczynski@kernel.org>
To: Bjorn Helgaas <helgaas@kernel.org>
Cc: "Bjorn Helgaas" <bhelgaas@google.com>,
	"Manivannan Sadhasivam" <mani@kernel.org>,
	"Lorenzo Pieralisi" <lpieralisi@kernel.org>,
	"Magnus Lindholm" <linmag7@gmail.com>,
	"Matt Turner" <mattst88@gmail.com>,
	"Richard Henderson" <richard.henderson@linaro.org>,
	"Christophe Leroy" <chleroy@kernel.org>,
	"Madhavan Srinivasan" <maddy@linux.ibm.com>,
	"Michael Ellerman" <mpe@ellerman.id.au>,
	"Nicholas Piggin" <npiggin@gmail.com>,
	"Dexuan Cui" <decui@microsoft.com>,
	"Krzysztof Hałasa" <khalasa@piap.pl>,
	"Lukas Wunner" <lukas@wunner.de>,
	"Oliver O'Halloran" <oohall@gmail.com>,
	"Saurabh Singh Sengar" <ssengar@microsoft.com>,
	"Shuan He" <heshuan@bytedance.com>,
	"Srivatsa Bhat" <srivatsabhat@microsoft.com>,
	"Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>,
	linux-pci@vger.kernel.org, linux-alpha@vger.kernel.org,
	linuxppc-dev@lists.ozlabs.org
Subject: Re: [PATCH v6 04/24] PCI/sysfs: Use BAR length in pci_llseek_resource() when attr->size is zero
Date: Thu, 30 Apr 2026 05:53:49 +0900	[thread overview]
Message-ID: <20260429203625.GA3724801@rocinante> (raw)
In-Reply-To: <20260429195055.GA312811@bhelgaas>

Hello,

> > @@ -909,11 +909,21 @@ static const struct attribute_group pci_dev_config_attr_group = {
> >   */
> >  static __maybe_unused loff_t
> >  pci_llseek_resource(struct file *filep,
> > -		    struct kobject *kobj __always_unused,
> > +		    struct kobject *kobj,
> >  		    const struct bin_attribute *attr,
> >  		    loff_t offset, int whence)
> >  {
> > -	return fixed_size_llseek(filep, offset, whence, attr->size);
> > +	struct pci_dev *pdev;
> > +	int bar;
> > +
> > +	if (attr->size)
> > +		return fixed_size_llseek(filep, offset, whence, attr->size);
> > +
> > +	pdev = to_pci_dev(kobj_to_dev(kobj));
> > +	bar = (unsigned long)attr->private;
> > +
> > +	return fixed_size_llseek(filep, offset, whence,
> > +				 pci_resource_len(pdev, bar));
>
> Is there a case where using "attr->size" is better than using
> "pci_resource_len(pdev, bar)"?
>
> In other words, would the following be equivalent?
>
>   pci_llseek_resource(...)
>   {
>     ...
>     pdev = to_pci_dev(kobj_to_dev(kobj));
>     bar = (unsigned long)attr->private;
>
>     return fixed_size_llseek(filep, offset, whence,
>                              pci_resource_len(pdev, bar));
>   }

Sadly, the simplified version would break legacy attributes.

pci_llseek_resource() is shared between device-level resource
attributes and bus-level legacy attributes, both have different
semantics:

  - Resource attributes (resource0, resource0_wc, ...) are per-device,
    carry a BAR index in attr->private, and will have .size == 0 with
    the static conversion.
  
  - Legacy attributes (legacy_io, legacy_mem, ...) are per-bus, have
    no BAR index in attr->private, and carry a fixed .size
    (PCI_LEGACY_IO_SIZE, PCI_LEGACY_MEM_SIZE, etc.).

The if (attr->size) check distinguishes the two cases, where legacy
attributes have size set at compile time (no BAR index), and the
resource attributes derive it from the BAR at runtime.

For legacy attributes, the kobj belongs to a struct pci_bus, not a
struct pci_dev, so to_pci_dev(kobj_to_dev(kobj)) would be a wrong
type for container_of().  Also, the pci_resource_len() helper would
not work there either.

Thus, dropping the attr->size check and always using pci_resource_len()
would break the legacy attributes case.

The alternative would be separate llseek callbacks for both the legacy
and resource attributes, which we can add if this would be the preference
here.

I hope this clears this up a little bit.

Thank you!

	Krzysztof

  reply	other threads:[~2026-04-29 20:53 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-22 16:13 [PATCH v6 00/24] PCI: Convert all dynamic sysfs attributes to static Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 01/24] PCI/sysfs: Use PCI resource accessor macros Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 02/24] PCI: Add pci_resource_is_io() and pci_resource_is_mem() helpers Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 03/24] PCI/sysfs: Only allow supported resource types in I/O and MMIO helpers Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 04/24] PCI/sysfs: Use BAR length in pci_llseek_resource() when attr->size is zero Krzysztof Wilczyński
2026-04-29 19:50   ` Bjorn Helgaas
2026-04-29 20:53     ` Krzysztof Wilczyński [this message]
2026-04-30 23:32       ` Krzysztof Wilczyński
2026-05-05 21:46         ` Bjorn Helgaas
2026-05-05 23:19           ` Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 05/24] PCI/sysfs: Add CAP_SYS_ADMIN check to __resource_resize_store() Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 06/24] PCI/sysfs: Add static PCI resource attribute macros Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 07/24] PCI/sysfs: Convert PCI resource files to static attributes Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 08/24] PCI/sysfs: Warn about BAR resize failure in __resource_resize_store() Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 09/24] PCI/sysfs: Add stubs for pci_{create,remove}_sysfs_dev_files() Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 10/24] PCI/sysfs: Limit pci_sysfs_init() late_initcall compile scope Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 11/24] alpha/PCI: Add security_locked_down() check to pci_mmap_resource() Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 12/24] alpha/PCI: Use BAR index in sysfs attr->private instead of resource pointer Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 13/24] alpha/PCI: Use PCI resource accessor macros Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 14/24] alpha/PCI: Fix __pci_mmap_fits() overflow for zero-length BARs Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 15/24] alpha/PCI: Remove WARN from __pci_mmap_fits() Krzysztof Wilczyński
2026-04-22 16:13 ` [PATCH v6 16/24] alpha/PCI: Add static PCI resource attribute macros Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 17/24] alpha/PCI: Convert resource files to static attributes Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 18/24] PCI/sysfs: Remove pci_{create,remove}_sysfs_dev_files() Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 19/24] PCI: Add macros for legacy I/O and memory address space sizes Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 20/24] alpha/PCI: Compute legacy size in pci_mmap_legacy_page_range() Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 21/24] PCI/sysfs: Add __weak pci_legacy_has_sparse() helper Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 22/24] PCI/sysfs: Convert legacy I/O and memory attributes to static definitions Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 23/24] PCI/sysfs: Remove pci_create_legacy_files() and pci_sysfs_init() Krzysztof Wilczyński
2026-04-22 16:14 ` [PATCH v6 24/24] PCI/sysfs: Limit BAR resize attribute scope to platforms with PCI mmap Krzysztof Wilczyński
2026-04-29 20:49   ` Bjorn Helgaas
2026-04-29 21:53     ` Krzysztof Wilczyński
2026-04-30  2:51       ` Krzysztof Wilczyński
2026-05-05 22:01         ` Bjorn Helgaas
2026-05-05 23:42           ` Krzysztof Wilczyński
2026-05-06 17:48             ` Bjorn Helgaas
2026-04-23 17:22 ` [PATCH v6 00/24] PCI: Convert all dynamic sysfs attributes to static Krzysztof Wilczyński
2026-04-23 18:23   ` Bjorn Helgaas
2026-04-23 19:00     ` Krzysztof Wilczyński
2026-04-27 14:09   ` Shivaprasad G Bhat
2026-04-27 23:34     ` Krzysztof Wilczyński
2026-04-29 20:04 ` Bjorn Helgaas
2026-04-30  1:31   ` Krzysztof Wilczyński

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=20260429203625.GA3724801@rocinante \
    --to=kwilczynski@kernel.org \
    --cc=bhelgaas@google.com \
    --cc=chleroy@kernel.org \
    --cc=decui@microsoft.com \
    --cc=helgaas@kernel.org \
    --cc=heshuan@bytedance.com \
    --cc=ilpo.jarvinen@linux.intel.com \
    --cc=khalasa@piap.pl \
    --cc=linmag7@gmail.com \
    --cc=linux-alpha@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=lpieralisi@kernel.org \
    --cc=lukas@wunner.de \
    --cc=maddy@linux.ibm.com \
    --cc=mani@kernel.org \
    --cc=mattst88@gmail.com \
    --cc=mpe@ellerman.id.au \
    --cc=npiggin@gmail.com \
    --cc=oohall@gmail.com \
    --cc=richard.henderson@linaro.org \
    --cc=srivatsabhat@microsoft.com \
    --cc=ssengar@microsoft.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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox