bcache patches for Linux v5.18

bcache patches for Linux v5.18

[Coly Li]

Hi Jens,

I have technical problem to send patches via email this time, please 
consider to pull the bcache changes from my bcache tree. They can be 
applied on top of your for-5.18/drivers branch.

We have 2 patches for Linux v5.18, both of them are from Mingzhe Zou. 
The first patch improves bcache initialization speed by avoid 
unnecessary cost of cache consistency, the second one fixes a potential 
NULL pointer deference in bcache initialization time,

Please take them for Linux v5.18, thanks in advance.

Coly Li


The following changes since commit dcde98da997075053041942ecf97d787855722ec:

   Merge branch 'for-linus' of 
git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input (2022-03-05 
15:49:45 -0800)

are available in the Git repository at:

https://git.kernel.org/pub/scm/linux/kernel/git/colyli/linux-bcache.git/ 
for-next

for you to fetch changes up to 387ac428d163b6fffa923618624505e7afe0b53d:

   bcache: fixup multiple threads crash (2022-03-06 18:14:43 +0800)

----------------------------------------------------------------
Mingzhe Zou (2):
       bcache: fixup bcache_dev_sectors_dirty_add() multithreaded CPU 
false sharing
       bcache: fixup multiple threads crash

  drivers/md/bcache/btree.c     |  6 ++++--
  drivers/md/bcache/writeback.c | 17 +++++++++++------
  2 files changed, 15 insertions(+), 8 deletions(-)

Re: bcache patches for Linux v5.18

[Jens Axboe]

On 3/6/22 3:35 AM, Coly Li wrote:
> Hi Jens,
> 
> I have technical problem to send patches via email this time, please
> consider to pull the bcache changes from my bcache tree. They can be
> applied on top of your for-5.18/drivers branch.
> 
> We have 2 patches for Linux v5.18, both of them are from Mingzhe Zou.
> The first patch improves bcache initialization speed by avoid
> unnecessary cost of cache consistency, the second one fixes a
> potential NULL pointer deference in bcache initialization time,
> 
> Please take them for Linux v5.18, thanks in advance.

I can take a git pull, but don't base it on something that isn't a tree
of mine. If I pull your branch right now, I'll get a ton of unrelated
changes.

If you want to do a git pull vs sending patches, base it on
for-5.18/drivers instead.

-- 
Jens Axboe

Re: bcache patches for Linux v5.18

[Coly Li]

On 3/6/22 10:17 PM, Jens Axboe wrote:
> On 3/6/22 3:35 AM, Coly Li wrote:
>> Hi Jens,
>>
>> I have technical problem to send patches via email this time, please
>> consider to pull the bcache changes from my bcache tree. They can be
>> applied on top of your for-5.18/drivers branch.
>>
>> We have 2 patches for Linux v5.18, both of them are from Mingzhe Zou.
>> The first patch improves bcache initialization speed by avoid
>> unnecessary cost of cache consistency, the second one fixes a
>> potential NULL pointer deference in bcache initialization time,
>>
>> Please take them for Linux v5.18, thanks in advance.
> I can take a git pull, but don't base it on something that isn't a tree
> of mine. If I pull your branch right now, I'll get a ton of unrelated
> changes.

I see, my for-next branch is not 100% clone the for-5.18/drivers branch, 
although the patches are verified on top of it.


> If you want to do a git pull vs sending patches, base it on
> for-5.18/drivers instead.
>
Copied. Now I pull my for-next branch from your for-5.18/drivers branch, 
and they are same, and the two patches are added on top of it.

Here is the updated pull request, could you please try it again? Thanks.


The following changes since commit 13d4ef0f66b7ee9415e101e213acaf94a0cb28ee:

   floppy: use memcpy_{to,from}_bvec (2022-03-04 12:29:21 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/colyli/linux-bcache.git 
for-next

for you to fetch changes up to 887554ab96588de2917b6c8c73e552da082e5368:

   bcache: fixup multiple threads crash (2022-03-06 22:33:45 +0800)

----------------------------------------------------------------
Mingzhe Zou (2):
       bcache: fixup bcache_dev_sectors_dirty_add() multithreaded CPU 
false sharing
       bcache: fixup multiple threads crash

  drivers/md/bcache/btree.c     |  6 ++++--
  drivers/md/bcache/writeback.c | 17 +++++++++++------
  2 files changed, 15 insertions(+), 8 deletions(-)

Re: bcache patches for Linux v5.18

[Jens Axboe]

On 3/6/22 7:55 AM, Coly Li wrote:
> On 3/6/22 10:17 PM, Jens Axboe wrote:
>> On 3/6/22 3:35 AM, Coly Li wrote:
>>> Hi Jens,
>>>
>>> I have technical problem to send patches via email this time, please
>>> consider to pull the bcache changes from my bcache tree. They can be
>>> applied on top of your for-5.18/drivers branch.
>>>
>>> We have 2 patches for Linux v5.18, both of them are from Mingzhe Zou.
>>> The first patch improves bcache initialization speed by avoid
>>> unnecessary cost of cache consistency, the second one fixes a
>>> potential NULL pointer deference in bcache initialization time,
>>>
>>> Please take them for Linux v5.18, thanks in advance.
>> I can take a git pull, but don't base it on something that isn't a tree
>> of mine. If I pull your branch right now, I'll get a ton of unrelated
>> changes.
> 
> I see, my for-next branch is not 100% clone the for-5.18/drivers
> branch, although the patches are verified on top of it.

It doesn't have to be a clone, but you can't sell a pull request that
ends up meaning the person pulling will get a lot more than the stuff
you just committed. Which means it should've been based on
for-5.18/drivers, or some earlier point there depending on when you
pulled it.

>> If you want to do a git pull vs sending patches, base it on
>> for-5.18/drivers instead.
>>
> Copied. Now I pull my for-next branch from your for-5.18/drivers
> branch, and they are same, and the two patches are added on top of it.

Thanks, pulled.

-- 
Jens Axboe

Re: bcache patches for Linux v5.18

[Coly Li]

On 3/6/22 11:15 PM, Jens Axboe wrote:
> On 3/6/22 7:55 AM, Coly Li wrote:
>> On 3/6/22 10:17 PM, Jens Axboe wrote:
>>> On 3/6/22 3:35 AM, Coly Li wrote:
>>>> Hi Jens,
>>>>
>>>> I have technical problem to send patches via email this time, please
>>>> consider to pull the bcache changes from my bcache tree. They can be
>>>> applied on top of your for-5.18/drivers branch.
>>>>
>>>> We have 2 patches for Linux v5.18, both of them are from Mingzhe Zou.
>>>> The first patch improves bcache initialization speed by avoid
>>>> unnecessary cost of cache consistency, the second one fixes a
>>>> potential NULL pointer deference in bcache initialization time,
>>>>
>>>> Please take them for Linux v5.18, thanks in advance.
>>> I can take a git pull, but don't base it on something that isn't a tree
>>> of mine. If I pull your branch right now, I'll get a ton of unrelated
>>> changes.
>> I see, my for-next branch is not 100% clone the for-5.18/drivers
>> branch, although the patches are verified on top of it.
> It doesn't have to be a clone, but you can't sell a pull request that
> ends up meaning the person pulling will get a lot more than the stuff
> you just committed. Which means it should've been based on
> for-5.18/drivers, or some earlier point there depending on when you
> pulled it.


Copied. Keep in mind for the hint.


>
>>> If you want to do a git pull vs sending patches, base it on
>>> for-5.18/drivers instead.
>>>
>> Copied. Now I pull my for-next branch from your for-5.18/drivers
>> branch, and they are same, and the two patches are added on top of it.
> Thanks, pulled.
>
Thanks.


Coly Li