Hi Linus,
I've provided the simplified diffstat from a temporary merge branch to
avoid the noise of merging QAT dependencies from a branch provided by
Herbert. The dependencies were already merged in commit 84c7d76b5ab6
("Merge tag 'v6.10-p1' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6")
Thanks,
Alex
The following changes since commit ed30a4a51bb196781c8058073ea720133a65596f:
Linux 6.9-rc5 (2024-04-21 12:35:54 -0700)
are available in the Git repository at:
https://github.com/awilliam/linux-vfio.git tags/vfio-v6.10-rc1
for you to fetch changes up to cbb325e77fbe62a06184175aa98c9eb98736c3e8:
vfio/pci: Restore zero affected bus reset devices warning (2024-05-17 08:00:52 -0600)
----------------------------------------------------------------
VFIO updates for v6.10-rc1
- The vfio fsl-mc bus driver has become orphaned. We'll consider
removing it in future releases if a new maintainer isn't found.
(Alex Williamson)
- Improved usage of opaque data in vfio-pci INTx handling,
avoiding lookups of the eventfd through the interrupt and
irqfd runtime paths. (Alex Williamson)
- Resolve an error path memory leak introduced in vfio-pci
interrupt code. (Ye Bin)
- Addition of interrupt support for vfio devices exposed on the
CDX bus, including a new MSI allocation helper and export of
existing helpers for MSI alloc and free. (Nipun Gupta)
- A new vfio-pci variant driver supporting migration of Intel
QAT VF devices for the GEN4 PFs. (Xin Zeng & Yahui Cao)
- Resolve a possibly circular locking dependency in vfio-pci
by avoiding copy_to_user() from a PCI bus walk callback.
(Alex Williamson)
- Trivial docs update to remove a duplicate semicolon.
(Foryun Ma)
----------------------------------------------------------------
Alex Williamson (6):
MAINTAINERS: Orphan vfio fsl-mc bus driver
vfio/pci: Pass eventfd context to IRQ handler
vfio/pci: Pass eventfd context object through irqfd
Merge branch 'vfio' of https://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6 into v6.10/vfio/qat-v7
vfio/pci: Collect hot-reset devices to local buffer
vfio/pci: Restore zero affected bus reset devices warning
Nipun Gupta (2):
genirq/msi: Add MSI allocation helper and export MSI functions
vfio/cdx: add interrupt support
Xin Zeng (1):
vfio/qat: Add vfio_pci driver for Intel QAT SR-IOV VF devices
Ye Bin (1):
vfio/pci: fix potential memory leak in vfio_intx_enable()
foryun.ma (1):
vfio: remove an extra semicolon
Documentation/driver-api/vfio.rst | 2 +-
MAINTAINERS | 11 +-
drivers/vfio/cdx/Makefile | 2 +-
drivers/vfio/cdx/intr.c | 217 ++++++++++++
drivers/vfio/cdx/main.c | 63 +++-
drivers/vfio/cdx/private.h | 18 +
drivers/vfio/pci/Kconfig | 2 +
drivers/vfio/pci/Makefile | 2 +
drivers/vfio/pci/qat/Kconfig | 12 +
drivers/vfio/pci/qat/Makefile | 3 +
drivers/vfio/pci/qat/main.c | 702 ++++++++++++++++++++++++++++++++++++++
drivers/vfio/pci/vfio_pci_core.c | 81 +++--
drivers/vfio/pci/vfio_pci_intrs.c | 61 ++--
include/linux/msi.h | 6 +
kernel/irq/msi.c | 2 +
15 files changed, 1116 insertions(+), 68 deletions(-)
create mode 100644 drivers/vfio/cdx/intr.c
create mode 100644 drivers/vfio/pci/qat/Kconfig
create mode 100644 drivers/vfio/pci/qat/Makefile
create mode 100644 drivers/vfio/pci/qat/main.c
On Mon, 20 May 2024 at 11:12, Alex Williamson
<[email protected]> wrote:
>
> I've provided the simplified diffstat from a temporary merge branch to
> avoid the noise of merging QAT dependencies from a branch provided by
> Herbert.
The diffstat looks good, but the merge itself sucks.
This is the totality of the "explanation" in the merge commit: "".
Yup. That's it. Nothing. Nada.
If you cannot explain *why* you merged a branch from some other tree,
youi damn well shouldn't have done the merge in the first place.
Merge commits need explanations just like regular commits do. In fact,
because there isn't some obvious diff attached to them, explanations
are arguably even more needed.
I've pulled this, but dammit, why does this keep happening?
Linus
The pull request you sent on Mon, 20 May 2024 12:12:23 -0600:
> https://github.com/awilliam/linux-vfio.git tags/vfio-v6.10-rc1
has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/30aec6e1bb617e1349d7fa5498898d7d4351d71e
Thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html
On Mon, 20 May 2024 15:05:26 -0700
Linus Torvalds <[email protected]> wrote:
> On Mon, 20 May 2024 at 11:12, Alex Williamson
> <[email protected]> wrote:
> >
> > I've provided the simplified diffstat from a temporary merge branch to
> > avoid the noise of merging QAT dependencies from a branch provided by
> > Herbert.
>
> The diffstat looks good, but the merge itself sucks.
>
> This is the totality of the "explanation" in the merge commit: "".
>
> Yup. That's it. Nothing. Nada.
>
> If you cannot explain *why* you merged a branch from some other tree,
> youi damn well shouldn't have done the merge in the first place.
>
> Merge commits need explanations just like regular commits do. In fact,
> because there isn't some obvious diff attached to them, explanations
> are arguably even more needed.
>
> I've pulled this, but dammit, why does this keep happening?
Sorry. In my case I've looked through logs and I've seen bare merges in
the past and I guess I assumed the reasoning here would be more obvious.
Clearly that's wrong. I'll do better. Thanks,
Alex
On Mon, 20 May 2024 at 16:03, Alex Williamson
<[email protected]> wrote:
>
> Sorry. In my case I've looked through logs and I've seen bare merges in
> the past and I guess I assumed the reasoning here would be more obvious.
Yeah, the bare merges in the past is why I'm so frustrated about this
all. I don't understand why this keeps happening so much.
Who do people keep doing this thing where they just think "I don't
need to explain this thing".
Yes, git made merges easy. It was one of the design goals, since (a) I
do a lot of them and (b) it's what EVERY OTHER SCM historically
absolutely sucked at.
But just because merging used to be hard, and git made it so easy,
doesn't mean that people should then not even explain them.
I complained to Andrew this merge window about one of his pull
requests that had _seven_ pointless and totally unexplained merges.
It's like people do this operation in their sleep or something, and
don't think about how big an operation a merge is.
Linus