RE: [PATCH v2] PCI: hv: Fix the definition of vector in hv_compose_msi_msg()

From: Dexuan Cui
Date: Mon Apr 03 2023 - 14:29:56 EST


> From: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Sent: Monday, April 3, 2023 6:08 AM
> To: Dexuan Cui <decui@xxxxxxxxxxxxx>
> Cc: Boqun Feng <boqun.feng@xxxxxxxxx>; quic_jhugo@xxxxxxxxxxx;
> quic_carlv@xxxxxxxxxxx; wei.liu@xxxxxxxxxx; KY Srinivasan
> <kys@xxxxxxxxxxxxx>; Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>;
> sthemmin@xxxxxxxxxxxxx; lpieralisi@xxxxxxxxxx; bhelgaas@xxxxxxxxxx;
> linux-hyperv@xxxxxxxxxxxxxxx; linux-pci@xxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx; Michael Kelley (LINUX)
> <mikelley@xxxxxxxxxxxxx>; robh@xxxxxxxxxx; kw@xxxxxxxxx;
> helgaas@xxxxxxxxxx; alex.williamson@xxxxxxxxxx; stable@xxxxxxxxxxxxxxx;
> Sasha Levin <sashal@xxxxxxxxxx>
> Subject: Re: [PATCH v2] PCI: hv: Fix the definition of vector in
> hv_compose_msi_msg()
>
> On Thu, Mar 30, 2023 at 03:23:45AM +0000, Dexuan Cui wrote:
> > > From: Boqun Feng <boqun.feng@xxxxxxxxx>
> > > Sent: Wednesday, March 29, 2023 7:56 PM
> > > To: Dexuan Cui <decui@xxxxxxxxxxxxx>
> > > ...
> > > On Wed, Mar 29, 2023 at 06:56:12PM -0700, Boqun Feng wrote:
> > > > [Cc stable]
> > > >
> > > > On Thu, Oct 27, 2022 at 01:52:56PM -0700, Dexuan Cui wrote:
> > > > > The local variable 'vector' must be u32 rather than u8: see the
> > > > > struct hv_msi_desc3.
> > > > >
> > > > > 'vector_count' should be u16 rather than u8: see struct hv_msi_desc,
> > > > > hv_msi_desc2 and hv_msi_desc3.
> > > > >
> > > >
> > > > Dexuan, I think this patch should only be in 5.15, because...
> > > >
> > >
> > > Sorry, I meant:
> > >
> > > "this patch should also be backported in 5.15"
> > >
> > > Regards,
> > > Boqun
> > >
> > > > > Fixes: a2bad844a67b ("PCI: hv: Fix interrupt mapping for multi-MSI")
> > > >
> > > > ^^^ this commit is already in 5.15.y (commit id 92dcb50f7f09).
> > > >
> > > > Upstream id e70af8d040d2b7904dca93d942ba23fb722e21b1
> > > > Cc: <stable@xxxxxxxxxxxxxxx> # 5.15.x
> >
> > The faulty commit a2bad844a67b ("PCI: hv: Fix interrupt mapping for
> multi-MSI")
> > is in all the stable branches, even including 4.14.y, so yes, the commit
> > e70af8d040d2 ("PCI: hv: Fix the definition of vector in
> hv_compose_msi_msg()")
> > should be backported to all the stable branches as well, including
> > v5.15.y, v5.10.y, v5.4.y, v4.19.y, v4.14.y.
> >
> > e70af8d040d2 has a Fixes tag. Not sure why it's not automatically
> backported.
>
> Also, the most obvious reason, it does NOT apply there! If this needs
> to go to 5.15.y and older, please send working backports of it.
>
> thanks,
>
> greg k-h

I also found that we need to manually rebase the patch for 5.15.y
and older. +Jose, who will help to do the backport.