linux-pci.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: LKML <linux-kernel@vger.kernel.org>
Cc: Bjorn Helgaas <helgaas@kernel.org>, Marc Zygnier <maz@kernel.org>,
	Alex Williamson <alex.williamson@redhat.com>,
	Kevin Tian <kevin.tian@intel.com>,
	Jason Gunthorpe <jgg@nvidia.com>, Megha Dey <megha.dey@intel.com>,
	Ashok Raj <ashok.raj@intel.com>,
	Michael Ellerman <mpe@ellerman.id.au>,
	Andrew Cooper <amc96@cam.ac.uk>, Juergen Gross <jgross@suse.com>,
	linux-pci@vger.kernel.org, xen-devel@lists.xenproject.org
Subject: [patch 00/10] genirq/msi, PCI/MSI: Support for dynamic MSI-X vector expansion - Part 4
Date: Sat, 27 Nov 2021 02:24:30 +0100 (CET)	[thread overview]
Message-ID: <20211126233124.618283684@linutronix.de> (raw)

This is finally the point where dynamically expanding MSI-X vectors after
enabling MSI-X is implemented.

The first three parts of this work can be found here:

    https://lore.kernel.org/r/20211126222700.862407977@linutronix.de
    https://lore.kernel.org/r/20211126224100.303046749@linutronix.de
    https://lore.kernel.org/r/20211126230957.239391799@linutronix.de

This last and smallest part of the overall series contains the following
changes:

   1) Prepare the core MSI irq domain code to handle range based allocation
      and free

   2) Prepare the PCI/MSI code to handle range based allocation and free
  
   3) Implement a new interface which allows to expand the MSI-X vector
      space after initialization

   4) Enable support for the X86 PCI/MSI irq domains

      This is unfortunate, but some PCI/MSI irq domain implementations,
      e.g. powerpc and the x86/XEN irqdomain wrappers are not really ready
      to support this out of the box.

      I looked at the 30 places which implement PCI/MSI irq domains and
      many of them look like they could support it out of the box, but as
      we have two which definitely don't, making this opt-in is the only
      safe option.

I've tested this by hacking up the XHCI driver and it works like a charm.

There is certainly some more room for consolidating the PCI/MSI-X usage in
drivers, i.e. getting rid of pci_enable_msix*(), but this would have made
this overall series even larger and is an orthogonal issue.

This fourth series is based on:

     git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git msi-v1-part-3

and also available from git:

     git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git msi-v1-part-4

Thanks,

	tglx
---
 arch/powerpc/platforms/pseries/msi.c |    6 +-
 arch/x86/kernel/apic/msi.c           |    4 -
 arch/x86/pci/xen.c                   |   10 +--
 drivers/base/platform-msi.c          |    3 -
 drivers/pci/msi/irqdomain.c          |   39 ++++++++++----
 drivers/pci/msi/msi.c                |   97 +++++++++++++++++++++++++++--------
 drivers/pci/msi/msi.h                |    4 -
 include/linux/msi.h                  |   46 +++++++++++-----
 include/linux/pci.h                  |   13 ++++
 kernel/irq/msi.c                     |   75 +++++++++++++++------------
 10 files changed, 208 insertions(+), 89 deletions(-)

WARNING: multiple messages have this Message-ID (diff)
From: Thomas Gleixner <tglx@linutronix.de>
To: LKML <linux-kernel@vger.kernel.org>
Cc: Bjorn Helgaas <helgaas@kernel.org>, Marc Zygnier <maz@kernel.org>,
	Alex Williamson <alex.williamson@redhat.com>,
	Kevin Tian <kevin.tian@intel.com>,
	Jason Gunthorpe <jgg@nvidia.com>, Megha Dey <megha.dey@intel.com>,
	Ashok Raj <ashok.raj@intel.com>,
	Michael Ellerman <mpe@ellerman.id.au>,
	Andrew Cooper <amc96@cam.ac.uk>, Juergen Gross <jgross@suse.com>,
	linux-pci@vger.kernel.org, xen-devel@lists.xenproject.org
Subject: [patch 00/10] genirq/msi, PCI/MSI: Support for dynamic MSI-X vector expansion - Part 4
Date: Sat, 27 Nov 2021 02:24:57 +0100 (CET)	[thread overview]
Message-ID: <20211126233124.618283684@linutronix.de> (raw)
Message-ID: <20211127012457.r6xYh-TuqB2phIkYdJAhJeWPz0HzlPOQNBl_UKa9LqY@z> (raw)

This is finally the point where dynamically expanding MSI-X vectors after
enabling MSI-X is implemented.

The first three parts of this work can be found here:

    https://lore.kernel.org/r/20211126222700.862407977@linutronix.de
    https://lore.kernel.org/r/20211126224100.303046749@linutronix.de
    https://lore.kernel.org/r/20211126230957.239391799@linutronix.de

This last and smallest part of the overall series contains the following
changes:

   1) Prepare the core MSI irq domain code to handle range based allocation
      and free

   2) Prepare the PCI/MSI code to handle range based allocation and free
  
   3) Implement a new interface which allows to expand the MSI-X vector
      space after initialization

   4) Enable support for the X86 PCI/MSI irq domains

      This is unfortunate, but some PCI/MSI irq domain implementations,
      e.g. powerpc and the x86/XEN irqdomain wrappers are not really ready
      to support this out of the box.

      I looked at the 30 places which implement PCI/MSI irq domains and
      many of them look like they could support it out of the box, but as
      we have two which definitely don't, making this opt-in is the only
      safe option.

I've tested this by hacking up the XHCI driver and it works like a charm.

There is certainly some more room for consolidating the PCI/MSI-X usage in
drivers, i.e. getting rid of pci_enable_msix*(), but this would have made
this series even larger and is an orthogonal issue.

The series is based on:

     git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git msi-v1-part-3

and also available from git:

     git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git msi-v1-part-4

Thanks,

	tglx
---
 arch/powerpc/platforms/pseries/msi.c |    6 +-
 arch/x86/kernel/apic/msi.c           |    4 -
 arch/x86/pci/xen.c                   |   10 +--
 drivers/base/platform-msi.c          |    3 -
 drivers/pci/msi/irqdomain.c          |   39 ++++++++++----
 drivers/pci/msi/msi.c                |   97 +++++++++++++++++++++++++++--------
 drivers/pci/msi/msi.h                |    4 -
 include/linux/msi.h                  |   46 +++++++++++-----
 include/linux/pci.h                  |   13 ++++
 kernel/irq/msi.c                     |   75 +++++++++++++++------------
 10 files changed, 208 insertions(+), 89 deletions(-)

             reply	other threads:[~2021-11-27  1:33 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-27  1:24 Thomas Gleixner [this message]
2021-11-27  1:24 ` [patch 01/10] genirq/msi: Add range argument to alloc/free MSI domain ops Thomas Gleixner
2021-11-27  1:24   ` Thomas Gleixner
2021-11-27  1:24 ` [patch 02/10] genirq/msi: Add range argument to msi_domain_alloc/free_descs_locked() Thomas Gleixner
2021-11-27  1:25   ` Thomas Gleixner
2021-11-27  1:24 ` [patch 03/10] genirq/msi: Make MSI descriptor alloc/free ready for range allocations Thomas Gleixner
2021-11-27  1:25   ` Thomas Gleixner
2021-11-28 15:57   ` Marc Zyngier
2021-11-28 19:17     ` Thomas Gleixner
2021-11-29 17:28       ` Thomas Gleixner
2021-11-27  1:24 ` [patch 05/10] genirq/msi: Add domain info flag MSI_FLAG_CAN_EXPAND Thomas Gleixner
2021-11-27  1:25   ` Thomas Gleixner
2021-11-27  1:24 ` [patch 06/10] PCI/MSI: Use range in allocation path Thomas Gleixner
2021-11-27  1:25   ` Thomas Gleixner
2021-11-27  1:24 ` [patch 08/10] PCI/MSI: Provide pci_msi_domain_supports_expand() Thomas Gleixner
2021-11-27  1:25   ` Thomas Gleixner
2021-11-27  1:24 ` [patch 10/10] x86/apic/msi: Support MSI-X vector expansion Thomas Gleixner
2021-11-27  1:25   ` Thomas Gleixner
2021-11-27  1:24 ` [patch 00/10] genirq/msi, PCI/MSI: Support for dynamic MSI-X vector expansion - Part 4 Thomas Gleixner
2021-11-27  1:25 ` [patch 04/10] genirq/msi: Prepare MSI domain alloc/free for range irq allocation Thomas Gleixner
2021-11-27  1:24   ` Thomas Gleixner
2021-11-27  1:25 ` [patch 07/10] PCI/MSI: Make free related functions range based Thomas Gleixner
2021-11-27  1:24   ` Thomas Gleixner
2021-11-27  1:25 ` [patch 09/10] PCI/MSI: Provide pci_msix_expand_vectors[_at]() Thomas Gleixner
2021-11-27  1:24   ` Thomas Gleixner
2021-12-02  1:08   ` Dey, Megha
2021-12-02 10:16     ` Thomas Gleixner
2021-12-02 19:21       ` Raj, Ashok
2021-12-02 20:40         ` Thomas Gleixner
2021-12-03  0:45           ` Raj, Ashok
2021-12-03 12:29             ` Thomas Gleixner

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=20211126233124.618283684@linutronix.de \
    --to=tglx@linutronix.de \
    --cc=alex.williamson@redhat.com \
    --cc=amc96@cam.ac.uk \
    --cc=ashok.raj@intel.com \
    --cc=helgaas@kernel.org \
    --cc=jgg@nvidia.com \
    --cc=jgross@suse.com \
    --cc=kevin.tian@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=maz@kernel.org \
    --cc=megha.dey@intel.com \
    --cc=mpe@ellerman.id.au \
    --cc=xen-devel@lists.xenproject.org \
    /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;
as well as URLs for NNTP newsgroup(s).