All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Pali Rohár" <pali@kernel.org>
To: Sandor Bodo-Merle <sbodomerle@gmail.com>
Cc: linux-pci@vger.kernel.org, bcm-kernel-feedback-list@broadcom.com
Subject: Re: pcie-iproc-msi.c: Bug in Multi-MSI support?
Date: Thu, 20 May 2021 16:05:29 +0200	[thread overview]
Message-ID: <20210520140529.rczoz3npjoadzfqc@pali> (raw)
In-Reply-To: <CABLWAfQbKy=fpaY6J=gqtJy5L+pqNeqwU6qkVswYaWnVjiwAHw@mail.gmail.com>

Hello!

On Thursday 20 May 2021 15:47:46 Sandor Bodo-Merle wrote:
> Hi Pali,
> 
> thanks for catching this - i dig up the followup fixup commit we have
> for the iproc multi MSI (it was sent to Broadcom - but unfortunately
> we missed upstreaming it).
> 
> Commit fc54bae28818 ("PCI: iproc: Allow allocation of multiple MSIs")
> failed to reserve the proper number of bits from the inner domain.
> We need to allocate the proper amount of bits otherwise the domains for
> multiple PCIe endpoints may overlap and freeing one of them will result
> in freeing unrelated MSI vectors.
> 
> Fixes: fc54bae28818 ("PCI: iproc: Allow allocation of multiple MSIs")
> ---
>  drivers/pci/host/pcie-iproc-msi.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git drivers/pci/host/pcie-iproc-msi.c drivers/pci/host/pcie-iproc-msi.c
> index 708fdb1065f8..a00492dccb74 100644
> --- drivers/pci/host/pcie-iproc-msi.c
> +++ drivers/pci/host/pcie-iproc-msi.c
> @@ -260,11 +260,11 @@ static int iproc_msi_irq_domain_alloc(struct
> irq_domain *domain,
> 
>         mutex_lock(&msi->bitmap_lock);
> 
> -       /* Allocate 'nr_cpus' number of MSI vectors each time */
> +       /* Allocate 'nr_irqs' multiplied by 'nr_cpus' number of MSI
> vectors each time */
>         hwirq = bitmap_find_next_zero_area(msi->bitmap, msi->nr_msi_vecs, 0,
> -                                          msi->nr_cpus, 0);
> +                                          msi->nr_cpus * nr_irqs, 0);

I'm not sure if this construction is correct. Multi-MSI interrupts needs
to be aligned to number of requested interrupts. So if wifi driver asks
for 32 Multi-MSI interrupts then first allocated interrupt number must
be dividable by 32.

>         if (hwirq < msi->nr_msi_vecs) {
> -               bitmap_set(msi->bitmap, hwirq, msi->nr_cpus);
> +               bitmap_set(msi->bitmap, hwirq, msi->nr_cpus * nr_irqs);

And another issue is that only power of 2 interrupts for Multi-MSI can
be allocated. Otherwise one number may be allocated to more devices.

But I'm not sure how number of CPUs affects it as other PCIe controller
drivers do not use number of CPUs.

Other drivers are using bitmap_find_free_region() function with
order_base_2(nr_irqs) as argument.

I hope that somebody else more skilled with MSI interrupts look at these
constructions if are correct or needs more rework.

>         } else {
>                 mutex_unlock(&msi->bitmap_lock);
>                 return -ENOSPC;
> @@ -292,7 +292,7 @@ static void iproc_msi_irq_domain_free(struct
> irq_domain *domain,
>         mutex_lock(&msi->bitmap_lock);
> 
>         hwirq = hwirq_to_canonical_hwirq(msi, data->hwirq);
> -       bitmap_clear(msi->bitmap, hwirq, msi->nr_cpus);
> +       bitmap_clear(msi->bitmap, hwirq, msi->nr_cpus * nr_irqs);
> 
>         mutex_unlock(&msi->bitmap_lock);
> 
> 
> On Thu, May 20, 2021 at 2:04 PM Pali Rohár <pali@kernel.org> wrote:
> >
> > Hello!
> >
> > I think there is a bug in pcie-iproc-msi.c driver. It declares
> > Multi MSI support via MSI_FLAG_MULTI_PCI_MSI flag, see:
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/pci/controller/pcie-iproc-msi.c?h=v5.12#n174
> >
> > but its iproc_msi_irq_domain_alloc() function completely ignores nr_irqs
> > argument when allocating interrupt numbers from bitmap, see:
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/pci/controller/pcie-iproc-msi.c?h=v5.12#n246
> >
> > I think this this is incorrect as alloc callback should allocate nr_irqs
> > multi interrupts as caller requested. All other drivers with Multi MSI
> > support are doing it.
> >
> > Could you look at it?

  reply	other threads:[~2021-05-20 14:07 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-20 12:00 pcie-iproc-msi.c: Bug in Multi-MSI support? Pali Rohár
2021-05-20 13:47 ` Sandor Bodo-Merle
2021-05-20 14:05   ` Pali Rohár [this message]
2021-05-20 14:22     ` Sandor Bodo-Merle
2021-05-20 17:11       ` Ray Jui
2021-05-20 19:31         ` Sandor Bodo-Merle
2021-05-21 11:39         ` Sandor Bodo-Merle
2021-05-21 14:00           ` Sandor Bodo-Merle
2021-05-24 10:37         ` Marc Zyngier
2021-05-25 17:27           ` Ray Jui
2021-05-25 17:35             ` Pali Rohár
2021-05-25 17:39               ` Ray Jui
2021-05-26  7:57             ` Marc Zyngier
2021-05-26 16:10               ` Sandor Bodo-Merle
2021-06-02  8:34                 ` Marc Zyngier
2021-06-03 16:59                   ` Ray Jui
2021-06-04  9:17                     ` Sandor Bodo-Merle
2021-06-04 16:47                       ` Ray Jui
2021-05-26 17:34               ` Ray Jui
2021-05-26 20:18                 ` Sandor Bodo-Merle

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=20210520140529.rczoz3npjoadzfqc@pali \
    --to=pali@kernel.org \
    --cc=bcm-kernel-feedback-list@broadcom.com \
    --cc=linux-pci@vger.kernel.org \
    --cc=sbodomerle@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.