From: Jonathan Cameron <Jonathan.Cameron@Huawei.com>
To: Lukas Wunner <lukas@wunner.de>
Cc: Bjorn Helgaas <helgaas@kernel.org>, <linux-pci@vger.kernel.org>,
Gregory Price <gregory.price@memverge.com>,
Ira Weiny <ira.weiny@intel.com>,
Dan Williams <dan.j.williams@intel.com>,
"Alison Schofield" <alison.schofield@intel.com>,
Vishal Verma <vishal.l.verma@intel.com>,
Dave Jiang <dave.jiang@intel.com>,
"Li, Ming" <ming4.li@intel.com>, Hillf Danton <hdanton@sina.com>,
Ben Widawsky <bwidawsk@kernel.org>, <linux-cxl@vger.kernel.org>
Subject: Re: [PATCH v2 06/10] PCI/DOE: Allow mailbox creation without devres management
Date: Tue, 24 Jan 2023 12:18:12 +0000 [thread overview]
Message-ID: <20230124121759.00007793@huawei.com> (raw)
In-Reply-To: <20230124121543.00002600@Huawei.com>
On Tue, 24 Jan 2023 12:15:43 +0000
Jonathan Cameron <Jonathan.Cameron@Huawei.com> wrote:
> On Mon, 23 Jan 2023 11:16:00 +0100
> Lukas Wunner <lukas@wunner.de> wrote:
>
> > DOE mailbox creation is currently only possible through a devres-managed
> > API. The lifetime of mailboxes thus ends with driver unbinding.
> >
> > An upcoming commit will create DOE mailboxes upon device enumeration by
> > the PCI core. Their lifetime shall not be limited by a driver.
> >
> > Therefore rework pcim_doe_create_mb() into the non-devres-managed
> > pci_doe_create_mb(). Add pci_doe_destroy_mb() for mailbox destruction
> > on device removal.
> >
> > Provide a devres-managed wrapper under the existing pcim_doe_create_mb()
> > name.
> >
> > Tested-by: Ira Weiny <ira.weiny@intel.com>
> > Signed-off-by: Lukas Wunner <lukas@wunner.de>
> Hi Lukas,
>
> A few comments inline.
>
> In particular I'd like to understand why flushing in the tear down
> can't always be done as that makes the code more complex.
>
> Might become clear in later patches though as I've not read ahead yet!
Ah.. It's in the patch description of the next patch. So ignore this question.
Thanks,
Jonathan
>
> Jonathan
>
> > ---
> > drivers/pci/doe.c | 103 +++++++++++++++++++++++++++++++---------------
> > 1 file changed, 70 insertions(+), 33 deletions(-)
> >
> > diff --git a/drivers/pci/doe.c b/drivers/pci/doe.c
> > index 066400531d09..cc1fdd75ad2a 100644
> > --- a/drivers/pci/doe.c
> > +++ b/drivers/pci/doe.c
> > @@ -37,7 +37,7 @@
> > *
> > * This state is used to manage a single DOE mailbox capability. All fields
> > * should be considered opaque to the consumers and the structure passed into
> > - * the helpers below after being created by devm_pci_doe_create()
> > + * the helpers below after being created by pci_doe_create_mb().
> > *
> > * @pdev: PCI device this mailbox belongs to
> > * @cap_offset: Capability offset
> > @@ -412,20 +412,6 @@ static int pci_doe_cache_protocols(struct pci_doe_mb *doe_mb)
> > return 0;
> > }
> >
> > -static void pci_doe_xa_destroy(void *mb)
> > -{
> > - struct pci_doe_mb *doe_mb = mb;
> > -
> > - xa_destroy(&doe_mb->prots);
> > -}
> > -
> > -static void pci_doe_destroy_workqueue(void *mb)
> > -{
> > - struct pci_doe_mb *doe_mb = mb;
> > -
> > - destroy_workqueue(doe_mb->work_queue);
> > -}
> > -
> > static void pci_doe_flush_mb(void *mb)
> > {
> > struct pci_doe_mb *doe_mb = mb;
> > @@ -442,7 +428,7 @@ static void pci_doe_flush_mb(void *mb)
> > }
> >
> > /**
> > - * pcim_doe_create_mb() - Create a DOE mailbox object
> > + * pci_doe_create_mb() - Create a DOE mailbox object
> > *
> > * @pdev: PCI device to create the DOE mailbox for
> > * @cap_offset: Offset of the DOE mailbox
> > @@ -453,24 +439,20 @@ static void pci_doe_flush_mb(void *mb)
> > * RETURNS: created mailbox object on success
> > * ERR_PTR(-errno) on failure
> > */
> > -struct pci_doe_mb *pcim_doe_create_mb(struct pci_dev *pdev, u16 cap_offset)
> > +static struct pci_doe_mb *pci_doe_create_mb(struct pci_dev *pdev,
> > + u16 cap_offset)
> > {
> > struct pci_doe_mb *doe_mb;
> > - struct device *dev = &pdev->dev;
> > int rc;
> >
> > - doe_mb = devm_kzalloc(dev, sizeof(*doe_mb), GFP_KERNEL);
> > + doe_mb = kzalloc(sizeof(*doe_mb), GFP_KERNEL);
> > if (!doe_mb)
> > return ERR_PTR(-ENOMEM);
> >
> > doe_mb->pdev = pdev;
> > doe_mb->cap_offset = cap_offset;
> > init_waitqueue_head(&doe_mb->wq);
> > -
> > xa_init(&doe_mb->prots);
> See below - I'd move xa_init() down to just above the pci_doe_cache_protocols()
> call.
>
> > - rc = devm_add_action(dev, pci_doe_xa_destroy, doe_mb);
> > - if (rc)
> > - return ERR_PTR(rc);
> >
> > doe_mb->work_queue = alloc_ordered_workqueue("%s %s DOE [%x]", 0,
> > dev_driver_string(&pdev->dev),
> > @@ -479,35 +461,90 @@ struct pci_doe_mb *pcim_doe_create_mb(struct pci_dev *pdev, u16 cap_offset)
> > if (!doe_mb->work_queue) {
> > pci_err(pdev, "[%x] failed to allocate work queue\n",
> > doe_mb->cap_offset);
> > - return ERR_PTR(-ENOMEM);
> > + rc = -ENOMEM;
> > + goto err_free;
> > }
> > - rc = devm_add_action_or_reset(dev, pci_doe_destroy_workqueue, doe_mb);
> > - if (rc)
> > - return ERR_PTR(rc);
> >
> > /* Reset the mailbox by issuing an abort */
> > rc = pci_doe_abort(doe_mb);
> > if (rc) {
> > pci_err(pdev, "[%x] failed to reset mailbox with abort command : %d\n",
> > doe_mb->cap_offset, rc);
> > - return ERR_PTR(rc);
> > + goto err_destroy_wq;
> > }
> >
> > /*
> > * The state machine and the mailbox should be in sync now;
> > - * Set up mailbox flush prior to using the mailbox to query protocols.
> > + * Use the mailbox to query protocols.
> > */
> > - rc = devm_add_action_or_reset(dev, pci_doe_flush_mb, doe_mb);
> > - if (rc)
> > - return ERR_PTR(rc);
> > -
> > rc = pci_doe_cache_protocols(doe_mb);
> > if (rc) {
> > pci_err(pdev, "[%x] failed to cache protocols : %d\n",
> > doe_mb->cap_offset, rc);
> > + goto err_flush;
> > + }
> > +
> > + return doe_mb;
> > +
> > +err_flush:
> > + pci_doe_flush_mb(doe_mb);
> > + xa_destroy(&doe_mb->prots);
>
> Why the reorder wrt to the original devm managed cleanup?
> I'd expect this to happen on any error path after the xa_init.
>
> It doesn't matter in practice because there isn't anything to
> do until after pci_doe_cache_protocols though. Maybe
> simplest option would be move xa_init() down to just above
> the call to pci_doe_cache_protocols()? That way the order
> you have here would meet the 'obviously correct' test.
>
>
> > +err_destroy_wq:
> > + destroy_workqueue(doe_mb->work_queue);
> > +err_free:
> > + kfree(doe_mb);
> > + return ERR_PTR(rc);
> > +}
> > +
> > +/**
> > + * pci_doe_destroy_mb() - Destroy a DOE mailbox object
> > + *
> > + * @ptr: Pointer to DOE mailbox
> > + *
> > + * Destroy all internal data structures created for the DOE mailbox.
>
> Could you comment on why it doesn't make sense to flush the
> mb on this path? Perhaps add a comment here to say what state
> we should be in before calling this?
>
> Not flushing here means you need more complex handling in
> error paths.
>
> > + */
> > +static void pci_doe_destroy_mb(void *ptr)
> > +{
> > + struct pci_doe_mb *doe_mb = ptr;
> > +
> > + xa_destroy(&doe_mb->prots);
>
> If making the change above, also push the xa_destroy() below
> the destroy_workqueue() here.
>
> > + destroy_workqueue(doe_mb->work_queue);
> > + kfree(doe_mb);
> > +}
> > +
> > +/**
> > + * pcim_doe_create_mb() - Create a DOE mailbox object
> > + *
> > + * @pdev: PCI device to create the DOE mailbox for
> > + * @cap_offset: Offset of the DOE mailbox
> > + *
> > + * Create a single mailbox object to manage the mailbox protocol at the
> > + * cap_offset specified. The mailbox will automatically be destroyed on
> > + * driver unbinding from @pdev.
> > + *
> > + * RETURNS: created mailbox object on success
> > + * ERR_PTR(-errno) on failure
> > + */
> > +struct pci_doe_mb *pcim_doe_create_mb(struct pci_dev *pdev, u16 cap_offset)
> > +{
> > + struct pci_doe_mb *doe_mb;
> > + int rc;
> > +
> > + doe_mb = pci_doe_create_mb(pdev, cap_offset);
> > + if (IS_ERR(doe_mb))
> > + return doe_mb;
> > +
> > + rc = devm_add_action(&pdev->dev, pci_doe_destroy_mb, doe_mb);
> > + if (rc) {
> > + pci_doe_flush_mb(doe_mb);
> > + pci_doe_destroy_mb(doe_mb);
> > return ERR_PTR(rc);
> > }
> >
> > + rc = devm_add_action_or_reset(&pdev->dev, pci_doe_flush_mb, doe_mb);
> > + if (rc)
> > + return ERR_PTR(rc);
> > +
> > return doe_mb;
> > }
> > EXPORT_SYMBOL_GPL(pcim_doe_create_mb);
>
next prev parent reply other threads:[~2023-01-24 12:18 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-23 10:10 [PATCH v2 00/10] Collection of DOE material Lukas Wunner
2023-01-23 10:11 ` [PATCH v2 01/10] PCI/DOE: Silence WARN splat with CONFIG_DEBUG_OBJECTS=y Lukas Wunner
2023-01-24 0:33 ` Ira Weiny
2023-01-24 10:32 ` Jonathan Cameron
2023-01-25 21:05 ` Lukas Wunner
2023-01-24 16:18 ` Gregory Price
2023-02-10 23:50 ` Dan Williams
2023-01-23 10:12 ` [PATCH v2 02/10] PCI/DOE: Fix memory leak " Lukas Wunner
2023-01-24 0:35 ` Ira Weiny
2023-01-24 10:33 ` Jonathan Cameron
2023-02-10 23:52 ` Dan Williams
2023-01-23 10:13 ` [PATCH v2 03/10] PCI/DOE: Provide synchronous API and use it internally Lukas Wunner
2023-01-24 0:48 ` Ira Weiny
2023-01-24 10:40 ` Jonathan Cameron
2023-01-24 20:07 ` Ira Weiny
2023-02-10 23:57 ` Dan Williams
2023-01-23 10:14 ` [PATCH v2 04/10] cxl/pci: Use synchronous API for DOE Lukas Wunner
2023-01-24 0:52 ` Ira Weiny
2023-02-03 8:53 ` Li, Ming
2023-02-03 8:56 ` Li, Ming
2023-02-03 9:54 ` Lukas Wunner
2023-01-24 11:01 ` Jonathan Cameron
2023-02-10 22:17 ` Lukas Wunner
2023-01-23 10:15 ` [PATCH v2 05/10] PCI/DOE: Make asynchronous API private Lukas Wunner
2023-01-24 0:55 ` Ira Weiny
2023-01-24 11:03 ` Jonathan Cameron
2023-01-23 10:16 ` [PATCH v2 06/10] PCI/DOE: Allow mailbox creation without devres management Lukas Wunner
2023-01-24 12:15 ` Jonathan Cameron
2023-01-24 12:18 ` Jonathan Cameron [this message]
2023-02-03 9:06 ` Li, Ming
2023-02-03 9:09 ` Li, Ming
2023-02-03 10:08 ` Lukas Wunner
2023-02-10 22:03 ` Lukas Wunner
2023-01-23 10:17 ` [PATCH v2 07/10] PCI/DOE: Create mailboxes on device enumeration Lukas Wunner
2023-01-24 1:14 ` Ira Weiny
2023-01-24 12:21 ` Jonathan Cameron
2023-01-23 10:18 ` [PATCH v2 08/10] cxl/pci: Use CDAT DOE mailbox created by PCI core Lukas Wunner
2023-01-24 1:18 ` Ira Weiny
2023-01-24 12:25 ` Jonathan Cameron
2023-01-23 10:19 ` [PATCH v2 09/10] PCI/DOE: Make mailbox creation API private Lukas Wunner
2023-01-24 1:25 ` Ira Weiny
2023-01-24 12:26 ` Jonathan Cameron
2023-01-23 10:20 ` [PATCH v2 10/10] PCI/DOE: Relax restrictions on request and response size Lukas Wunner
2023-01-23 22:29 ` Bjorn Helgaas
2023-01-24 1:43 ` Ira Weiny
2023-02-10 21:47 ` Lukas Wunner
2023-01-24 12:43 ` Jonathan Cameron
2023-01-24 23:51 ` Bjorn Helgaas
2023-01-25 9:47 ` Jonathan Cameron
2023-02-10 22:10 ` Lukas Wunner
2023-01-23 22:30 ` [PATCH v2 00/10] Collection of DOE material Bjorn Helgaas
2023-02-10 21:39 ` Lukas Wunner
2023-02-11 0:04 ` Dan Williams
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=20230124121759.00007793@huawei.com \
--to=jonathan.cameron@huawei.com \
--cc=alison.schofield@intel.com \
--cc=bwidawsk@kernel.org \
--cc=dan.j.williams@intel.com \
--cc=dave.jiang@intel.com \
--cc=gregory.price@memverge.com \
--cc=hdanton@sina.com \
--cc=helgaas@kernel.org \
--cc=ira.weiny@intel.com \
--cc=linux-cxl@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lukas@wunner.de \
--cc=ming4.li@intel.com \
--cc=vishal.l.verma@intel.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;
as well as URLs for NNTP newsgroup(s).