From: Wei Yang <weiyang@linux.vnet.ibm.com>
To: Bjorn Helgaas <bhelgaas@google.com>
Cc: Gavin Shan <gwshan@linux.vnet.ibm.com>,
Wei Yang <weiyang@linux.vnet.ibm.com>,
linuxppc-dev@lists.ozlabs.org, linux-pci@vger.kernel.org
Subject: Re: [PATCH V7 06/10] powerpc/eeh: Create PE for VFs
Date: Tue, 16 Jun 2015 16:50:24 +0800 [thread overview]
Message-ID: <20150616085024.GA1230@richard> (raw)
In-Reply-To: <20150603154638.GL3631@google.com>
On Wed, Jun 03, 2015 at 10:46:38AM -0500, Bjorn Helgaas wrote:
>On Wed, Jun 03, 2015 at 03:10:23PM +1000, Gavin Shan wrote:
>> It's correct. "The following operations" refers to eeh_add_device_late()
>> and eeh_sysfs_add_device(). The former one requires the resources for
>> one particular PCI device (VF here) are finalized (assigned). eeh_sysfs_add_device()
>> will fail if the sysfs entry for the PCI device isn't populated yet.
>
>eeh_add_device_late() contains several things that read config space:
>eeh_save_bars() caches the entire config header, and
>eeh_addr_cache_insert_dev() looks at the device resources (which are
>determined by BARs in config space). I think this is an error-prone
>approach. I think it would be simpler and safer for you to capture what
>you need in your PCI config accessors.
>
>eeh_add_device_late() also contains code to deal with an EEH cache that
>"might not be removed correctly because of unbalanced kref to the device
>during unplug time." That's unrelated to this patch series, but it sounds
>... like a hacky workaround for some bug in the unplug path.
>
>> >>> + eeh_add_device_early(pdn);
>> >>> + eeh_add_device_late(pdev);
>> >>> + eeh_sysfs_add_device(pdev);
>> >>> +}
>> >>> +DECLARE_PCI_FIXUP_FINAL(PCI_ANY_ID, PCI_ANY_ID, pnv_eeh_vf_final_fixup);
>> >>
>> >>Ugh. This is powerpc code, but I don't like using fixups as a hook like
>> >>this. There is a pcibios_add_device() -- could this be done there?
>> >>
>> >
>> >I don't like it neither :-) But looks we can't put it in the
>> >pcibios_add_device().
>> >
>> >>If not, what happens after pcibios_add_device() that is required for this
>> >>code? Maybe we need a pcibios_bus_add_device() hook?
>> >
>> >The pnv_eeh_vf_final_fixup() will try to create sysfs for VFs. This requires
>> >the VF sysfs(kobj) is initialized properly. If we put these into
>> >pcibios_add_device(), the eeh_sysfs_add_device() would fail.
>> >
>> >Below is the call flow for your reference:
>> >
>> >pci_device_add()
>> > pcibios_add_device()
>> > device_add() <--- kobj initialized here
>> >
>>
>> We can put it into pcibios_bus_add_device(), but we don't it currently. If
>> Bjorn agree to add pcibios_bus_add_device(), I'm fine to move the block code
>> there.
>
>I think I'm OK with adding a pcibios_bus_add_device(). I think that would
>be better than using the fixup mechanism for this.
>
Hi, Bjorn, Gavin,
Been working on some bug recently, just got a chance to this one.
Would you mind giving me some hint, where you suggest to put the
pcibios_bus_add_device()?
>Bjorn
--
Richard Yang
Help you, Help me
next prev parent reply other threads:[~2015-06-16 8:51 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-19 1:35 [PATCH V6 00/10] VF EEH on Power8 Wei Yang
2015-05-19 1:35 ` [PATCH V6 01/10] PCI/IOV: Rename and export virtfn_add/virtfn_remove Wei Yang
2015-05-19 5:24 ` Wei Yang
2015-05-19 1:35 ` [PATCH V6 02/10] powerpc/pci: Cache VF index in pci_dn Wei Yang
2015-05-19 1:35 ` [PATCH V6 03/10] powerpc/pci: Remove VFs prior to PF Wei Yang
2015-05-19 1:35 ` [PATCH V6 04/10] powerpc/eeh: Trace first 7 BARs in address cache Wei Yang
2015-05-19 1:35 ` [PATCH V6 05/10] powerpc/powernv: EEH device for VF Wei Yang
2015-05-19 1:35 ` [PATCH V6 06/10] powerpc/eeh: Create PE for VFs Wei Yang
2015-05-19 1:35 ` [PATCH V6 07/10] powerpc/powernv: Support EEH reset for VF PE Wei Yang
2015-05-19 1:35 ` [PATCH V6 08/10] powerpc/powernv: Support PCI config restore for VFs Wei Yang
2015-05-19 1:35 ` [PATCH V6 09/10] powerpc/eeh: Support error recovery for VF PE Wei Yang
2015-05-19 1:35 ` [PATCH V6 10/10] powerpc/powernv: compound PE for VFs Wei Yang
2015-05-19 10:50 ` [PATCH V7 00/10] VF EEH on Power8 Wei Yang
2015-05-19 10:50 ` [PATCH V7 01/10] PCI/IOV: Rename and export virtfn_add/virtfn_remove Wei Yang
2015-06-02 17:19 ` Bjorn Helgaas
2015-06-03 1:38 ` Wei Yang
2015-05-19 10:50 ` [PATCH V7 02/10] powerpc/pci: Cache VF index in pci_dn Wei Yang
2015-05-19 10:50 ` [PATCH V7 03/10] powerpc/pci: Remove VFs prior to PF Wei Yang
2015-06-01 23:20 ` Bjorn Helgaas
2015-06-02 3:44 ` Wei Yang
2015-05-19 10:50 ` [PATCH V7 04/10] powerpc/eeh: Trace first 7 BARs in address cache Wei Yang
2015-06-01 23:32 ` Bjorn Helgaas
2015-06-02 3:51 ` Wei Yang
2015-06-02 4:11 ` Gavin Shan
2015-06-03 1:47 ` Wei Yang
2015-05-19 10:50 ` [PATCH V7 05/10] powerpc/powernv: EEH device for VF Wei Yang
2015-05-19 10:50 ` [PATCH V7 06/10] powerpc/eeh: Create PE for VFs Wei Yang
2015-06-01 23:46 ` Bjorn Helgaas
2015-06-03 3:31 ` Wei Yang
2015-06-03 5:10 ` Gavin Shan
2015-06-03 15:46 ` Bjorn Helgaas
2015-06-04 1:25 ` Gavin Shan
2015-06-04 5:46 ` Wei Yang
2015-06-04 7:10 ` Gavin Shan
2015-06-16 8:50 ` Wei Yang [this message]
2015-06-16 13:22 ` Bjorn Helgaas
2015-06-01 23:49 ` Bjorn Helgaas
2015-06-03 3:39 ` Wei Yang
2015-05-19 10:50 ` [PATCH V7 07/10] powerpc/powernv: Support EEH reset for VF PE Wei Yang
2015-05-19 10:50 ` [PATCH V7 08/10] powerpc/powernv: Support PCI config restore for VFs Wei Yang
2015-06-02 0:01 ` Bjorn Helgaas
2015-06-03 1:37 ` Wei Yang
2015-06-03 5:14 ` Gavin Shan
2015-05-19 10:50 ` [PATCH V7 09/10] powerpc/eeh: Support error recovery for VF PE Wei Yang
2015-05-19 10:50 ` [PATCH V7 10/10] powerpc/powernv: compound PE for VFs Wei Yang
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=20150616085024.GA1230@richard \
--to=weiyang@linux.vnet.ibm.com \
--cc=bhelgaas@google.com \
--cc=gwshan@linux.vnet.ibm.com \
--cc=linux-pci@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.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).