| Message ID | 20240229094112.1154644-1-kunwu.chan@linux.dev (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | nfs: use KMEM_CACHE() to create nfs_commit_data cache | expand |
On Thu, 2024-02-29 at 17:41 +0800, kunwu.chan@linux.dev wrote: > From: Kunwu Chan <chentao@kylinos.cn> > > Use the KMEM_CACHE() macro instead of kmem_cache_create() to simplify > the creation of SLAB caches when the default values are used. > > Signed-off-by: Kunwu Chan <chentao@kylinos.cn> > --- > fs/nfs/write.c | 5 +---- > 1 file changed, 1 insertion(+), 4 deletions(-) > > diff --git a/fs/nfs/write.c b/fs/nfs/write.c > index bb79d3a886ae..6a75772d447f 100644 > --- a/fs/nfs/write.c > +++ b/fs/nfs/write.c > @@ -2148,10 +2148,7 @@ int __init nfs_init_writepagecache(void) > if (nfs_wdata_mempool == NULL) > goto out_destroy_write_cache; > > - nfs_cdata_cachep = kmem_cache_create("nfs_commit_data", > - sizeof(struct > nfs_commit_data), > - 0, SLAB_HWCACHE_ALIGN, > - NULL); > + nfs_cdata_cachep = KMEM_CACHE(nfs_commit_data, > SLAB_HWCACHE_ALIGN); > if (nfs_cdata_cachep == NULL) > goto out_destroy_write_mempool; If this were being done as part of an actual functional code change, then I'd be OK with it, but otherwise it is just unnecessary churn that gets in the way of back porting any future fixes.
Thanks for the reply. On 2024/2/29 21:40, Trond Myklebust wrote: > On Thu, 2024-02-29 at 17:41 +0800, kunwu.chan@linux.dev wrote: >> From: Kunwu Chan <chentao@kylinos.cn> >> >> Use the KMEM_CACHE() macro instead of kmem_cache_create() to simplify >> the creation of SLAB caches when the default values are used. >> >> Signed-off-by: Kunwu Chan <chentao@kylinos.cn> >> --- >> fs/nfs/write.c | 5 +---- >> 1 file changed, 1 insertion(+), 4 deletions(-) >> >> diff --git a/fs/nfs/write.c b/fs/nfs/write.c >> index bb79d3a886ae..6a75772d447f 100644 >> --- a/fs/nfs/write.c >> +++ b/fs/nfs/write.c >> @@ -2148,10 +2148,7 @@ int __init nfs_init_writepagecache(void) >> if (nfs_wdata_mempool == NULL) >> goto out_destroy_write_cache; >> >> - nfs_cdata_cachep = kmem_cache_create("nfs_commit_data", >> - sizeof(struct >> nfs_commit_data), >> - 0, SLAB_HWCACHE_ALIGN, >> - NULL); >> + nfs_cdata_cachep = KMEM_CACHE(nfs_commit_data, >> SLAB_HWCACHE_ALIGN); >> if (nfs_cdata_cachep == NULL) >> goto out_destroy_write_mempool; > > If this were being done as part of an actual functional code change, > then I'd be OK with it, but otherwise it is just unnecessary churn that > gets in the way of back porting any future fixes. It's just my personal opinion, I meant to do some cleanup. It's not entirely necessary either, as everyone prefers a different style of code. It doesn't matter. > >
diff --git a/fs/nfs/write.c b/fs/nfs/write.c index bb79d3a886ae..6a75772d447f 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -2148,10 +2148,7 @@ int __init nfs_init_writepagecache(void) if (nfs_wdata_mempool == NULL) goto out_destroy_write_cache; - nfs_cdata_cachep = kmem_cache_create("nfs_commit_data", - sizeof(struct nfs_commit_data), - 0, SLAB_HWCACHE_ALIGN, - NULL); + nfs_cdata_cachep = KMEM_CACHE(nfs_commit_data, SLAB_HWCACHE_ALIGN); if (nfs_cdata_cachep == NULL) goto out_destroy_write_mempool;