Message ID | 20201116044529.1028783-1-dkadashev@gmail.com (mailing list archive) |
---|---|
Headers | show |
Series | io_uring: add mkdirat support | expand |
On Mon, Nov 16, 2020 at 11:45:27AM +0700, Dmitry Kadashev wrote: > This adds mkdirat support to io_uring and is heavily based on recently > added renameat() / unlinkat() support. > > The first patch is preparation with no functional changes, makes > do_mkdirat accept struct filename pointer rather than the user string. > > The second one leverages that to implement mkdirat in io_uring. > > Based on for-5.11/io_uring. > > Dmitry Kadashev (2): > fs: make do_mkdirat() take struct filename > io_uring: add support for IORING_OP_MKDIRAT > > fs/internal.h | 1 + > fs/io_uring.c | 58 +++++++++++++++++++++++++++++++++++ > fs/namei.c | 20 ++++++++---- > include/uapi/linux/io_uring.h | 1 + > 4 files changed, 74 insertions(+), 6 deletions(-) > > -- > 2.28.0 > Hi Al Viro, Ping. Jens mentioned before that this looks fine by him, but you or someone from fsdevel should approve the namei.c part first. Thanks, Dmitry
On Fri, Dec 4, 2020 at 5:57 PM Dmitry Kadashev <dkadashev@gmail.com> wrote: > > On Mon, Nov 16, 2020 at 11:45:27AM +0700, Dmitry Kadashev wrote: > > This adds mkdirat support to io_uring and is heavily based on recently > > added renameat() / unlinkat() support. > > > > The first patch is preparation with no functional changes, makes > > do_mkdirat accept struct filename pointer rather than the user string. > > > > The second one leverages that to implement mkdirat in io_uring. > > > > Based on for-5.11/io_uring. > > > > Dmitry Kadashev (2): > > fs: make do_mkdirat() take struct filename > > io_uring: add support for IORING_OP_MKDIRAT > > > > fs/internal.h | 1 + > > fs/io_uring.c | 58 +++++++++++++++++++++++++++++++++++ > > fs/namei.c | 20 ++++++++---- > > include/uapi/linux/io_uring.h | 1 + > > 4 files changed, 74 insertions(+), 6 deletions(-) > > > > -- > > 2.28.0 > > > > Hi Al Viro, > > Ping. Jens mentioned before that this looks fine by him, but you or > someone from fsdevel should approve the namei.c part first. Another ping. Jens, you've mentioned the patch looks good to you, and with quite similar changes (unlinkat, renameat) being sent for 5.11 is there anything that I can do to help this to be accepted (not necessarily for 5.11 at this point)?
On 12/15/20 4:43 AM, Dmitry Kadashev wrote: > On Fri, Dec 4, 2020 at 5:57 PM Dmitry Kadashev <dkadashev@gmail.com> wrote: >> >> On Mon, Nov 16, 2020 at 11:45:27AM +0700, Dmitry Kadashev wrote: >>> This adds mkdirat support to io_uring and is heavily based on recently >>> added renameat() / unlinkat() support. >>> >>> The first patch is preparation with no functional changes, makes >>> do_mkdirat accept struct filename pointer rather than the user string. >>> >>> The second one leverages that to implement mkdirat in io_uring. >>> >>> Based on for-5.11/io_uring. >>> >>> Dmitry Kadashev (2): >>> fs: make do_mkdirat() take struct filename >>> io_uring: add support for IORING_OP_MKDIRAT >>> >>> fs/internal.h | 1 + >>> fs/io_uring.c | 58 +++++++++++++++++++++++++++++++++++ >>> fs/namei.c | 20 ++++++++---- >>> include/uapi/linux/io_uring.h | 1 + >>> 4 files changed, 74 insertions(+), 6 deletions(-) >>> >>> -- >>> 2.28.0 >>> >> >> Hi Al Viro, >> >> Ping. Jens mentioned before that this looks fine by him, but you or >> someone from fsdevel should approve the namei.c part first. > > Another ping. > > Jens, you've mentioned the patch looks good to you, and with quite > similar changes (unlinkat, renameat) being sent for 5.11 is there > anything that I can do to help this to be accepted (not necessarily > for 5.11 at this point)? Since we're aiming for 5.12 at this point, let's just hold off a bit and see if Al gets time to ack/review the VFS side of things. There's no immediate rush. It's on my TODO list, so we'll get there eventually.
On Tue, Dec 15, 2020 at 11:20 PM Jens Axboe <axboe@kernel.dk> wrote: > > On 12/15/20 4:43 AM, Dmitry Kadashev wrote: > > On Fri, Dec 4, 2020 at 5:57 PM Dmitry Kadashev <dkadashev@gmail.com> wrote: > >> > >> On Mon, Nov 16, 2020 at 11:45:27AM +0700, Dmitry Kadashev wrote: > >>> This adds mkdirat support to io_uring and is heavily based on recently > >>> added renameat() / unlinkat() support. > >>> > >>> The first patch is preparation with no functional changes, makes > >>> do_mkdirat accept struct filename pointer rather than the user string. > >>> > >>> The second one leverages that to implement mkdirat in io_uring. > >>> > >>> Based on for-5.11/io_uring. > >>> > >>> Dmitry Kadashev (2): > >>> fs: make do_mkdirat() take struct filename > >>> io_uring: add support for IORING_OP_MKDIRAT > >>> > >>> fs/internal.h | 1 + > >>> fs/io_uring.c | 58 +++++++++++++++++++++++++++++++++++ > >>> fs/namei.c | 20 ++++++++---- > >>> include/uapi/linux/io_uring.h | 1 + > >>> 4 files changed, 74 insertions(+), 6 deletions(-) > >>> > >>> -- > >>> 2.28.0 > >>> > >> > >> Hi Al Viro, > >> > >> Ping. Jens mentioned before that this looks fine by him, but you or > >> someone from fsdevel should approve the namei.c part first. > > > > Another ping. > > > > Jens, you've mentioned the patch looks good to you, and with quite > > similar changes (unlinkat, renameat) being sent for 5.11 is there > > anything that I can do to help this to be accepted (not necessarily > > for 5.11 at this point)? > > Since we're aiming for 5.12 at this point, let's just hold off a bit and > see if Al gets time to ack/review the VFS side of things. There's no > immediate rush. OK, sounds good, thanks.
On Tue, Dec 15, 2020 at 11:20 PM Jens Axboe <axboe@kernel.dk> wrote: > > On 12/15/20 4:43 AM, Dmitry Kadashev wrote: > > On Fri, Dec 4, 2020 at 5:57 PM Dmitry Kadashev <dkadashev@gmail.com> wrote: > >> > >> On Mon, Nov 16, 2020 at 11:45:27AM +0700, Dmitry Kadashev wrote: > >>> This adds mkdirat support to io_uring and is heavily based on recently > >>> added renameat() / unlinkat() support. > >>> > >>> The first patch is preparation with no functional changes, makes > >>> do_mkdirat accept struct filename pointer rather than the user string. > >>> > >>> The second one leverages that to implement mkdirat in io_uring. > >>> > >>> Based on for-5.11/io_uring. > >>> > >>> Dmitry Kadashev (2): > >>> fs: make do_mkdirat() take struct filename > >>> io_uring: add support for IORING_OP_MKDIRAT > >>> > >>> fs/internal.h | 1 + > >>> fs/io_uring.c | 58 +++++++++++++++++++++++++++++++++++ > >>> fs/namei.c | 20 ++++++++---- > >>> include/uapi/linux/io_uring.h | 1 + > >>> 4 files changed, 74 insertions(+), 6 deletions(-) > >>> > >>> -- > >>> 2.28.0 > >>> > >> > >> Hi Al Viro, > >> > >> Ping. Jens mentioned before that this looks fine by him, but you or > >> someone from fsdevel should approve the namei.c part first. > > > > Another ping. > > > > Jens, you've mentioned the patch looks good to you, and with quite > > similar changes (unlinkat, renameat) being sent for 5.11 is there > > anything that I can do to help this to be accepted (not necessarily > > for 5.11 at this point)? > > Since we're aiming for 5.12 at this point, let's just hold off a bit and > see if Al gets time to ack/review the VFS side of things. There's no > immediate rush. > > It's on my TODO list, so we'll get there eventually. Another reminder, since afaict 5.12 stuff is being merged now.
On 11/15/20 9:45 PM, Dmitry Kadashev wrote: > This adds mkdirat support to io_uring and is heavily based on recently > added renameat() / unlinkat() support. > > The first patch is preparation with no functional changes, makes > do_mkdirat accept struct filename pointer rather than the user string. > > The second one leverages that to implement mkdirat in io_uring. > > Based on for-5.11/io_uring. I want to tentatively queue this up. Do you have the liburing support and test case(s) for it as well that you can send?
On Wed, Jan 27, 2021 at 5:35 AM Jens Axboe <axboe@kernel.dk> wrote: > > On 11/15/20 9:45 PM, Dmitry Kadashev wrote: > > This adds mkdirat support to io_uring and is heavily based on recently > > added renameat() / unlinkat() support. > > > > The first patch is preparation with no functional changes, makes > > do_mkdirat accept struct filename pointer rather than the user string. > > > > The second one leverages that to implement mkdirat in io_uring. > > > > Based on for-5.11/io_uring. > > I want to tentatively queue this up. Do you have the liburing support > and test case(s) for it as well that you can send? I do, I've sent it in the past, here it is: https://lore.kernel.org/io-uring/20201116051005.1100302-1-dkadashev@gmail.com/ I need to (figure out the way to) fix the kernel / namei side after Al's comments though.
On 1/27/21 4:06 AM, Dmitry Kadashev wrote: > On Wed, Jan 27, 2021 at 5:35 AM Jens Axboe <axboe@kernel.dk> wrote: >> >> On 11/15/20 9:45 PM, Dmitry Kadashev wrote: >>> This adds mkdirat support to io_uring and is heavily based on recently >>> added renameat() / unlinkat() support. >>> >>> The first patch is preparation with no functional changes, makes >>> do_mkdirat accept struct filename pointer rather than the user string. >>> >>> The second one leverages that to implement mkdirat in io_uring. >>> >>> Based on for-5.11/io_uring. >> >> I want to tentatively queue this up. Do you have the liburing support >> and test case(s) for it as well that you can send? > > I do, I've sent it in the past, here it is: > https://lore.kernel.org/io-uring/20201116051005.1100302-1-dkadashev@gmail.com/ I thought so, thanks. I'll queue it up once we have agreement on the kernel side. > I need to (figure out the way to) fix the kernel / namei side after Al's > comments though. Thanks, yes please do and re-post it.