The first patch adds new APIs to support migration of users
from old device coredump related APIs.
The second patch fix sleep in atomic context bugs of mwifiex
caused by dev_coredumpv().
Duoming Zhou (2):
devcoredump: add new APIs to support migration of users from old
device coredump related APIs
mwifiex: fix sleep in atomic context bugs caused by dev_coredumpv
drivers/base/devcoredump.c | 116 ++++++++++++++++++
drivers/net/wireless/marvell/mwifiex/init.c | 9 +-
drivers/net/wireless/marvell/mwifiex/main.h | 3 +-
.../net/wireless/marvell/mwifiex/sta_event.c | 6 +-
include/linux/devcoredump.h | 34 +++++
5 files changed, 160 insertions(+), 8 deletions(-)
--
2.17.1
On Tue, Aug 23, 2022 at 4:21 AM Duoming Zhou <[email protected]> wrote:
>
> The first patch adds new APIs to support migration of users
> from old device coredump related APIs.
>
> The second patch fix sleep in atomic context bugs of mwifiex
> caused by dev_coredumpv().
>
> Duoming Zhou (2):
> devcoredump: add new APIs to support migration of users from old
> device coredump related APIs
> mwifiex: fix sleep in atomic context bugs caused by dev_coredumpv
I would have expected a third patch in here, that actually converts
existing users. Then in the following release cycle, clean up any new
users of the old API that pop up in the meantime and drop the old API.
But I'll defer to the people who would actually be merging your code.
Technically it could also work to simply provide the API this cycle,
and convert everyone in the next.
Brian
Hello,
On Wed, 24 Aug 2022 13:42:09 -0700 Brian Norris wrote:
> On Tue, Aug 23, 2022 at 4:21 AM Duoming Zhou <[email protected]> wrote:
> >
> > The first patch adds new APIs to support migration of users
> > from old device coredump related APIs.
> >
> > The second patch fix sleep in atomic context bugs of mwifiex
> > caused by dev_coredumpv().
> >
> > Duoming Zhou (2):
> > devcoredump: add new APIs to support migration of users from old
> > device coredump related APIs
> > mwifiex: fix sleep in atomic context bugs caused by dev_coredumpv
>
> I would have expected a third patch in here, that actually converts
> existing users. Then in the following release cycle, clean up any new
> users of the old API that pop up in the meantime and drop the old API.
>
> But I'll defer to the people who would actually be merging your code.
> Technically it could also work to simply provide the API this cycle,
> and convert everyone in the next.
Thank your for your time and reply.
If this patch set is merged into the linux-next tree, I will send the
third patch which targets at linux-next tree and converts existing users
at later timer of this release cycle. Because there are new users that
may use the old APIs comes into linux-next tree during the remaining time
of this release cycle.
Best regards,
Duoming Zhou
On Thu, Aug 25, 2022 at 08:44:55AM +0800, [email protected] wrote:
> Hello,
>
> On Wed, 24 Aug 2022 13:42:09 -0700 Brian Norris wrote:
>
> > On Tue, Aug 23, 2022 at 4:21 AM Duoming Zhou <[email protected]> wrote:
> > >
> > > The first patch adds new APIs to support migration of users
> > > from old device coredump related APIs.
> > >
> > > The second patch fix sleep in atomic context bugs of mwifiex
> > > caused by dev_coredumpv().
> > >
> > > Duoming Zhou (2):
> > > devcoredump: add new APIs to support migration of users from old
> > > device coredump related APIs
> > > mwifiex: fix sleep in atomic context bugs caused by dev_coredumpv
> >
> > I would have expected a third patch in here, that actually converts
> > existing users. Then in the following release cycle, clean up any new
> > users of the old API that pop up in the meantime and drop the old API.
> >
> > But I'll defer to the people who would actually be merging your code.
> > Technically it could also work to simply provide the API this cycle,
> > and convert everyone in the next.
>
> Thank your for your time and reply.
>
> If this patch set is merged into the linux-next tree, I will send the
> third patch which targets at linux-next tree and converts existing users
> at later timer of this release cycle. Because there are new users that
> may use the old APIs comes into linux-next tree during the remaining time
> of this release cycle.
No, that's not how this works, we can't add patches with new functions
that no one uses. And it's not how I asked for this api to be migrated
over time properly. I'll try to respond to your patch with more details
in a week or so when I catch up on patch review...
greg k-h
Hello,
On Thu, 25 Aug 2022 07:59:33 +0200 Greg Kroah-Hartman wrote:
> > On Wed, 24 Aug 2022 13:42:09 -0700 Brian Norris wrote:
> >
> > > On Tue, Aug 23, 2022 at 4:21 AM Duoming Zhou <[email protected]> wrote:
> > > >
> > > > The first patch adds new APIs to support migration of users
> > > > from old device coredump related APIs.
> > > >
> > > > The second patch fix sleep in atomic context bugs of mwifiex
> > > > caused by dev_coredumpv().
> > > >
> > > > Duoming Zhou (2):
> > > > devcoredump: add new APIs to support migration of users from old
> > > > device coredump related APIs
> > > > mwifiex: fix sleep in atomic context bugs caused by dev_coredumpv
> > >
> > > I would have expected a third patch in here, that actually converts
> > > existing users. Then in the following release cycle, clean up any new
> > > users of the old API that pop up in the meantime and drop the old API.
> > >
> > > But I'll defer to the people who would actually be merging your code.
> > > Technically it could also work to simply provide the API this cycle,
> > > and convert everyone in the next.
> >
> > Thank your for your time and reply.
> >
> > If this patch set is merged into the linux-next tree, I will send the
> > third patch which targets at linux-next tree and converts existing users
> > at later timer of this release cycle. Because there are new users that
> > may use the old APIs comes into linux-next tree during the remaining time
> > of this release cycle.
>
> No, that's not how this works, we can't add patches with new functions
> that no one uses. And it's not how I asked for this api to be migrated
> over time properly. I'll try to respond to your patch with more details
> in a week or so when I catch up on patch review...
Thank you for your time and I look forward to your reply.
Best regards,
Duoming Zhou