| Message ID | 1594923911-10885-1-git-send-email-jbaron@akamai.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | EDAC/ie31200: fallback if host bridge device is already initialized | expand |
On 7/16/20 2:52 PM, Luck, Tony wrote: > On Thu, Jul 16, 2020 at 02:25:11PM -0400, Jason Baron wrote: >> The Intel uncore driver may claim some of the pci ids from ie31200 which >> means that the ie31200 edac driver will not initialize them as part of >> pci_register_driver(). >> >> Let's add a fallback for this case to 'pci_get_device()' to get a >> reference on the device such that it can still be configured. This is >> similar in approach to other edac drivers. > > What functionality is lost when this happens? There is no difference in functionality when the fallback occurs. It should operate in the same it does when the uncore driver is not loaded. > > Does the user see some message in the console > log to let them know? I don't think its needed as there should be no user-visible difference. Thanks, -Jason
On 7/16/20 4:33 PM, Jason Baron wrote: > > > On 7/16/20 2:52 PM, Luck, Tony wrote: >> On Thu, Jul 16, 2020 at 02:25:11PM -0400, Jason Baron wrote: >>> The Intel uncore driver may claim some of the pci ids from ie31200 which >>> means that the ie31200 edac driver will not initialize them as part of >>> pci_register_driver(). >>> Hi, I just wondering if there is any feedback on this issue, without this patch the ie31200 edac driver doesn't load properly on a number of boxes. Thanks, -Jason
On Thu, Aug 06, 2020 at 05:35:49PM -0400, Jason Baron wrote: > > > On 7/16/20 4:33 PM, Jason Baron wrote: > > > > > > On 7/16/20 2:52 PM, Luck, Tony wrote: > >> On Thu, Jul 16, 2020 at 02:25:11PM -0400, Jason Baron wrote: > >>> The Intel uncore driver may claim some of the pci ids from ie31200 which > >>> means that the ie31200 edac driver will not initialize them as part of > >>> pci_register_driver(). > >>> > > Hi, > > I just wondering if there is any feedback on this issue, without this > patch the ie31200 edac driver doesn't load properly on a number of boxes. Applied it now. I'll see if I can get it merged for v5.9 (since you posted in plenty of time to make this merge window). -Tony
On Thu, Jul 16, 2020 at 02:25:11PM -0400, Jason Baron wrote: > The Intel uncore driver may claim some of the pci ids from ie31200 which > means that the ie31200 edac driver will not initialize them as part of > pci_register_driver(). > > Let's add a fallback for this case to 'pci_get_device()' to get a > reference on the device such that it can still be configured. This is > similar in approach to other edac drivers. > > Signed-off-by: Jason Baron <jbaron@akamai.com> > Cc: Borislav Petkov <bp@suse.de> > Cc: Mauro Carvalho Chehab <mchehab@kernel.org> > Cc: Tony Luck <tony.luck@intel.com> > Cc: linux-edac <linux-edac@vger.kernel.org> > --- > drivers/edac/ie31200_edac.c | 50 ++++++++++++++++++++++++++++++++++++++++++--- > 1 file changed, 47 insertions(+), 3 deletions(-) > > diff --git a/drivers/edac/ie31200_edac.c b/drivers/edac/ie31200_edac.c > index d68346a..ebe5099 100644 > --- a/drivers/edac/ie31200_edac.c > +++ b/drivers/edac/ie31200_edac.c > @@ -170,6 +170,8 @@ > (n << (28 + (2 * skl) - PAGE_SHIFT)) > > static int nr_channels; > +static struct pci_dev *mci_pdev; > +static int ie31200_registered = 1; > > struct ie31200_priv { > void __iomem *window; > @@ -538,12 +540,16 @@ static int ie31200_probe1(struct pci_dev *pdev, int dev_idx) > static int ie31200_init_one(struct pci_dev *pdev, > const struct pci_device_id *ent) > { > - edac_dbg(0, "MC:\n"); > + int rc; > > + edac_dbg(0, "MC:\n"); > if (pci_enable_device(pdev) < 0) > return -EIO; > + rc = ie31200_probe1(pdev, ent->driver_data); > + if (rc == 0 && !mci_pdev) > + mci_pdev = pci_dev_get(pdev); > > - return ie31200_probe1(pdev, ent->driver_data); > + return rc; > } > > static void ie31200_remove_one(struct pci_dev *pdev) > @@ -552,6 +558,8 @@ static void ie31200_remove_one(struct pci_dev *pdev) > struct ie31200_priv *priv; > > edac_dbg(0, "\n"); > + pci_dev_put(mci_pdev); > + mci_pdev = NULL; > mci = edac_mc_del_mc(&pdev->dev); > if (!mci) > return; > @@ -593,17 +601,53 @@ static struct pci_driver ie31200_driver = { > > static int __init ie31200_init(void) > { > + int pci_rc, i; > + > edac_dbg(3, "MC:\n"); > /* Ensure that the OPSTATE is set correctly for POLL or NMI */ > opstate_init(); > > - return pci_register_driver(&ie31200_driver); > + pci_rc = pci_register_driver(&ie31200_driver); > + if (pci_rc < 0) > + goto fail0; > + > + if (!mci_pdev) { > + ie31200_registered = 0; > + for (i = 0; ie31200_pci_tbl[i].vendor != 0; i++) { > + mci_pdev = pci_get_device(ie31200_pci_tbl[i].vendor, > + ie31200_pci_tbl[i].device, > + NULL); > + if (mci_pdev) > + break; > + } > + if (!mci_pdev) { > + edac_dbg(0, "ie31200 pci_get_device fail\n"); > + pci_rc = -ENODEV; > + goto fail1; > + } > + pci_rc = ie31200_init_one(mci_pdev, &ie31200_pci_tbl[i]); > + if (pci_rc < 0) { > + edac_dbg(0, "ie31200 init fail\n"); > + pci_rc = -ENODEV; > + goto fail1; > + } > + } > + return 0; > + > +fail1: > + pci_unregister_driver(&ie31200_driver); > +fail0: > + pci_dev_put(mci_pdev); > + > + return pci_rc; > } > > static void __exit ie31200_exit(void) > { > edac_dbg(3, "MC:\n"); > pci_unregister_driver(&ie31200_driver); > + if (!ie31200_registered) > + ie31200_remove_one(mci_pdev); > } > > module_init(ie31200_init); We tested this inside in machines having this issue and it works. Patch looks good to me. Acked-by: Aristeu Rozanski <aris@redhat.com>
On August 13, 2020 4:44:06 PM GMT+03:00, Aristeu Rozanski <aris@redhat.com> wrote: >We tested this inside in machines having this issue and it works. >Patch looks good to me. > >Acked-by: Aristeu Rozanski <aris@redhat.com> So Tested-by: you ?
On Thu, Aug 13, 2020 at 04:55:50PM +0300, Boris Petkov wrote: > On August 13, 2020 4:44:06 PM GMT+03:00, Aristeu Rozanski <aris@redhat.com> wrote: > >We tested this inside in machines having this issue and it works. > >Patch looks good to me. > > > >Acked-by: Aristeu Rozanski <aris@redhat.com> > > So Tested-by: you ? Not by me, "we" meant as in company. Tested-by: Vishal Agrawal <vagrawal@redhat.com>
On August 13, 2020 5:17:10 PM GMT+03:00, Aristeu Rozanski <aris@redhat.com> wrote: >> So Tested-by: you ? > >Not by me, "we" meant as in company. "you" can also mean you as a company. ;-P >Tested-by: Vishal Agrawal <vagrawal@redhat.com> Thx.
> >Tested-by: Vishal Agrawal <vagrawal@redhat.com>
Boris,
I applied this patch when Jason Baron sent a reminder last week. It is sitting in the
ie31200 topic branch of the RAS tree (and also in the next branch ... so has had a
couple of days in linux-next too).
That copy doesn't have these Acked-by and Tested-by tags, but is otherwise the
same.
I plan to send a second edac pull request to Linus today/romorrow to pick it up.
-Tony
diff --git a/drivers/edac/ie31200_edac.c b/drivers/edac/ie31200_edac.c index d68346a..ebe5099 100644 --- a/drivers/edac/ie31200_edac.c +++ b/drivers/edac/ie31200_edac.c @@ -170,6 +170,8 @@ (n << (28 + (2 * skl) - PAGE_SHIFT)) static int nr_channels; +static struct pci_dev *mci_pdev; +static int ie31200_registered = 1; struct ie31200_priv { void __iomem *window; @@ -538,12 +540,16 @@ static int ie31200_probe1(struct pci_dev *pdev, int dev_idx) static int ie31200_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) { - edac_dbg(0, "MC:\n"); + int rc; + edac_dbg(0, "MC:\n"); if (pci_enable_device(pdev) < 0) return -EIO; + rc = ie31200_probe1(pdev, ent->driver_data); + if (rc == 0 && !mci_pdev) + mci_pdev = pci_dev_get(pdev); - return ie31200_probe1(pdev, ent->driver_data); + return rc; } static void ie31200_remove_one(struct pci_dev *pdev) @@ -552,6 +558,8 @@ static void ie31200_remove_one(struct pci_dev *pdev) struct ie31200_priv *priv; edac_dbg(0, "\n"); + pci_dev_put(mci_pdev); + mci_pdev = NULL; mci = edac_mc_del_mc(&pdev->dev); if (!mci) return; @@ -593,17 +601,53 @@ static struct pci_driver ie31200_driver = { static int __init ie31200_init(void) { + int pci_rc, i; + edac_dbg(3, "MC:\n"); /* Ensure that the OPSTATE is set correctly for POLL or NMI */ opstate_init(); - return pci_register_driver(&ie31200_driver); + pci_rc = pci_register_driver(&ie31200_driver); + if (pci_rc < 0) + goto fail0; + + if (!mci_pdev) { + ie31200_registered = 0; + for (i = 0; ie31200_pci_tbl[i].vendor != 0; i++) { + mci_pdev = pci_get_device(ie31200_pci_tbl[i].vendor, + ie31200_pci_tbl[i].device, + NULL); + if (mci_pdev) + break; + } + if (!mci_pdev) { + edac_dbg(0, "ie31200 pci_get_device fail\n"); + pci_rc = -ENODEV; + goto fail1; + } + pci_rc = ie31200_init_one(mci_pdev, &ie31200_pci_tbl[i]); + if (pci_rc < 0) { + edac_dbg(0, "ie31200 init fail\n"); + pci_rc = -ENODEV; + goto fail1; + } + } + return 0; + +fail1: + pci_unregister_driver(&ie31200_driver); +fail0: + pci_dev_put(mci_pdev); + + return pci_rc; } static void __exit ie31200_exit(void) { edac_dbg(3, "MC:\n"); pci_unregister_driver(&ie31200_driver); + if (!ie31200_registered) + ie31200_remove_one(mci_pdev); } module_init(ie31200_init);
The Intel uncore driver may claim some of the pci ids from ie31200 which means that the ie31200 edac driver will not initialize them as part of pci_register_driver(). Let's add a fallback for this case to 'pci_get_device()' to get a reference on the device such that it can still be configured. This is similar in approach to other edac drivers. Signed-off-by: Jason Baron <jbaron@akamai.com> Cc: Borislav Petkov <bp@suse.de> Cc: Mauro Carvalho Chehab <mchehab@kernel.org> Cc: Tony Luck <tony.luck@intel.com> Cc: linux-edac <linux-edac@vger.kernel.org> --- drivers/edac/ie31200_edac.c | 50 ++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 47 insertions(+), 3 deletions(-)