| Message ID | 20200722090518.214624-2-cmaiolino@redhat.com (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Headers | show |
| Series | Continue xfs kmem cleanup - V3 | expand |
On Wed, Jul 22, 2020 at 11:05:14AM +0200, Carlos Maiolino wrote: > - /* > - * if this didn't occur in transactions, we could use > - * KM_MAYFAIL and return NULL here on ENOMEM. Set the > - * code up to do this anyway. > - */ > - ip = kmem_zone_alloc(xfs_inode_zone, 0); > - if (!ip) > - return NULL; > + ip = kmem_cache_alloc(xfs_inode_zone, GFP_KERNEL | __GFP_NOFAIL); > + I would have kept a version of this comment. But if this on your radar for the next merge window anyway this should be ok: Reviewed-by: Christoph Hellwig <hch@lst.de>
On Wed, Jul 22, 2020 at 03:13:46PM +0100, Christoph Hellwig wrote: > On Wed, Jul 22, 2020 at 11:05:14AM +0200, Carlos Maiolino wrote: > > - /* > > - * if this didn't occur in transactions, we could use > > - * KM_MAYFAIL and return NULL here on ENOMEM. Set the > > - * code up to do this anyway. > > - */ > > - ip = kmem_zone_alloc(xfs_inode_zone, 0); > > - if (!ip) > > - return NULL; > > + ip = kmem_cache_alloc(xfs_inode_zone, GFP_KERNEL | __GFP_NOFAIL); > > + > > I would have kept a version of this comment. But if this on your > radar for the next merge window anyway this should be ok: I thought it makes no sense to keep the comment, and yes, I'll keep an eye on it and rework this as soon as we have PF_FSTRANS back. Thanks for the review. > > Reviewed-by: Christoph Hellwig <hch@lst.de> >
On Wed, Jul 22, 2020 at 11:05:14AM +0200, Carlos Maiolino wrote: > Use kmem_cache_alloc() directly. > > All kmem_zone_alloc() users pass 0 as flags, which are translated into: > GFP_KERNEL | __GFP_NOWARN, and kmem_zone_alloc() loops forever until the > allocation succeeds. > > We can use __GFP_NOFAIL to tell the allocator to loop forever rather > than doing it ourself, and because the allocation will never fail, we do > not need to use __GFP_NOWARN anymore. Hence, all callers can be > converted to use GFP_KERNEL | __GFP_NOFAIL > > Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com> > --- > > Changelog: > V2: > - Wire up xfs_inode_alloc to use __GFP_NOFAIL > if it's called inside a transaction > - Rewrite changelog in a more decent way. > V3: > - Use __GFP_NOFAIL unconditionally in xfs_inode_alloc(), > use of PF_FSTRANS will be added when the patch re-adding > it is moved to mainline. > > fs/xfs/libxfs/xfs_alloc.c | 3 ++- > fs/xfs/libxfs/xfs_bmap.c | 3 ++- > fs/xfs/xfs_icache.c | 10 ++-------- > 3 files changed, 6 insertions(+), 10 deletions(-) Looks fine with or without the comment as christoph noted. Reviewed-by: Dave Chinner <dchinner@redhat.com>
On Wed, Jul 22, 2020 at 11:05:14AM +0200, Carlos Maiolino wrote: > Use kmem_cache_alloc() directly. > > All kmem_zone_alloc() users pass 0 as flags, which are translated into: > GFP_KERNEL | __GFP_NOWARN, and kmem_zone_alloc() loops forever until the > allocation succeeds. > > We can use __GFP_NOFAIL to tell the allocator to loop forever rather > than doing it ourself, and because the allocation will never fail, we do > not need to use __GFP_NOWARN anymore. Hence, all callers can be > converted to use GFP_KERNEL | __GFP_NOFAIL > > Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com> Looks ok to me too (I put the comment back in), Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> --D > --- > > Changelog: > V2: > - Wire up xfs_inode_alloc to use __GFP_NOFAIL > if it's called inside a transaction > - Rewrite changelog in a more decent way. > V3: > - Use __GFP_NOFAIL unconditionally in xfs_inode_alloc(), > use of PF_FSTRANS will be added when the patch re-adding > it is moved to mainline. > > fs/xfs/libxfs/xfs_alloc.c | 3 ++- > fs/xfs/libxfs/xfs_bmap.c | 3 ++- > fs/xfs/xfs_icache.c | 10 ++-------- > 3 files changed, 6 insertions(+), 10 deletions(-) > > diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c > index 203e74fa64aa6..583242253c027 100644 > --- a/fs/xfs/libxfs/xfs_alloc.c > +++ b/fs/xfs/libxfs/xfs_alloc.c > @@ -2467,7 +2467,8 @@ xfs_defer_agfl_block( > ASSERT(xfs_bmap_free_item_zone != NULL); > ASSERT(oinfo != NULL); > > - new = kmem_zone_alloc(xfs_bmap_free_item_zone, 0); > + new = kmem_cache_alloc(xfs_bmap_free_item_zone, > + GFP_KERNEL | __GFP_NOFAIL); > new->xefi_startblock = XFS_AGB_TO_FSB(mp, agno, agbno); > new->xefi_blockcount = 1; > new->xefi_oinfo = *oinfo; > diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c > index 667cdd0dfdf4a..fd5c0d669d0d7 100644 > --- a/fs/xfs/libxfs/xfs_bmap.c > +++ b/fs/xfs/libxfs/xfs_bmap.c > @@ -553,7 +553,8 @@ __xfs_bmap_add_free( > #endif > ASSERT(xfs_bmap_free_item_zone != NULL); > > - new = kmem_zone_alloc(xfs_bmap_free_item_zone, 0); > + new = kmem_cache_alloc(xfs_bmap_free_item_zone, > + GFP_KERNEL | __GFP_NOFAIL); > new->xefi_startblock = bno; > new->xefi_blockcount = (xfs_extlen_t)len; > if (oinfo) > diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c > index 58a750ce689c0..c2d97e4f131fb 100644 > --- a/fs/xfs/xfs_icache.c > +++ b/fs/xfs/xfs_icache.c > @@ -36,14 +36,8 @@ xfs_inode_alloc( > { > struct xfs_inode *ip; > > - /* > - * if this didn't occur in transactions, we could use > - * KM_MAYFAIL and return NULL here on ENOMEM. Set the > - * code up to do this anyway. > - */ > - ip = kmem_zone_alloc(xfs_inode_zone, 0); > - if (!ip) > - return NULL; > + ip = kmem_cache_alloc(xfs_inode_zone, GFP_KERNEL | __GFP_NOFAIL); > + > if (inode_init_always(mp->m_super, VFS_I(ip))) { > kmem_cache_free(xfs_inode_zone, ip); > return NULL; > -- > 2.26.2 >
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 203e74fa64aa6..583242253c027 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -2467,7 +2467,8 @@ xfs_defer_agfl_block( ASSERT(xfs_bmap_free_item_zone != NULL); ASSERT(oinfo != NULL); - new = kmem_zone_alloc(xfs_bmap_free_item_zone, 0); + new = kmem_cache_alloc(xfs_bmap_free_item_zone, + GFP_KERNEL | __GFP_NOFAIL); new->xefi_startblock = XFS_AGB_TO_FSB(mp, agno, agbno); new->xefi_blockcount = 1; new->xefi_oinfo = *oinfo; diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 667cdd0dfdf4a..fd5c0d669d0d7 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -553,7 +553,8 @@ __xfs_bmap_add_free( #endif ASSERT(xfs_bmap_free_item_zone != NULL); - new = kmem_zone_alloc(xfs_bmap_free_item_zone, 0); + new = kmem_cache_alloc(xfs_bmap_free_item_zone, + GFP_KERNEL | __GFP_NOFAIL); new->xefi_startblock = bno; new->xefi_blockcount = (xfs_extlen_t)len; if (oinfo) diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index 58a750ce689c0..c2d97e4f131fb 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -36,14 +36,8 @@ xfs_inode_alloc( { struct xfs_inode *ip; - /* - * if this didn't occur in transactions, we could use - * KM_MAYFAIL and return NULL here on ENOMEM. Set the - * code up to do this anyway. - */ - ip = kmem_zone_alloc(xfs_inode_zone, 0); - if (!ip) - return NULL; + ip = kmem_cache_alloc(xfs_inode_zone, GFP_KERNEL | __GFP_NOFAIL); + if (inode_init_always(mp->m_super, VFS_I(ip))) { kmem_cache_free(xfs_inode_zone, ip); return NULL;
Use kmem_cache_alloc() directly. All kmem_zone_alloc() users pass 0 as flags, which are translated into: GFP_KERNEL | __GFP_NOWARN, and kmem_zone_alloc() loops forever until the allocation succeeds. We can use __GFP_NOFAIL to tell the allocator to loop forever rather than doing it ourself, and because the allocation will never fail, we do not need to use __GFP_NOWARN anymore. Hence, all callers can be converted to use GFP_KERNEL | __GFP_NOFAIL Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com> --- Changelog: V2: - Wire up xfs_inode_alloc to use __GFP_NOFAIL if it's called inside a transaction - Rewrite changelog in a more decent way. V3: - Use __GFP_NOFAIL unconditionally in xfs_inode_alloc(), use of PF_FSTRANS will be added when the patch re-adding it is moved to mainline. fs/xfs/libxfs/xfs_alloc.c | 3 ++- fs/xfs/libxfs/xfs_bmap.c | 3 ++- fs/xfs/xfs_icache.c | 10 ++-------- 3 files changed, 6 insertions(+), 10 deletions(-)