This second version:
- add recipients for the Documentation/core-api changes
- add fix for __alloc_pages_bulk() to handle GFP_NOFAIL
- drops the annotations for congestion_wait() as being ineffective
as that isn't really useful until an alternative is available
- changes to GFP_NOFAIL documentation changes to focus on the possible
deadlocks rather than the use of memory reserves
- Improves ext4 and xfs patches based on feedback from Ted and Dave.
The patches are independent, except that the last patch depends on the
first.
As mentioned last time:
These are the easy bits. There are 5 calls to congestion_wait() and
one to wait_iff_congested() in mm/ which need consideration. There
are multiple calls to congestion_wait in fs/, particularly fs/f2fs/
which need to be addressed too. I'll try to form an opinion about
these in coming weeks.
(other interesting comment in original cover letter just duplicates
observations made in the commit messages of individual patches).
NeilBrown
---
NeilBrown (6):
MM: Support __GFP_NOFAIL in alloc_pages_bulk_*() and improve doco
MM: improve documentation for __GFP_NOFAIL
EXT4: Remove ENOMEM/congestion_wait() loops.
EXT4: remove congestion_wait from ext4_bio_write_page, and simplify
XFS: remove congestion_wait() loop from kmem_alloc()
XFS: remove congestion_wait() loop from xfs_buf_alloc_pages()
Documentation/core-api/memory-allocation.rst | 25 ++++++++-
fs/ext4/ext4.h | 2 +-
fs/ext4/ext4_jbd2.c | 4 +-
fs/ext4/ext4_jbd2.h | 14 +++---
fs/ext4/extents.c | 53 ++++++++------------
fs/ext4/extents_status.c | 35 +++++++------
fs/ext4/extents_status.h | 2 +-
fs/ext4/ialloc.c | 3 +-
fs/ext4/indirect.c | 2 +-
fs/ext4/inode.c | 6 +--
fs/ext4/ioctl.c | 4 +-
fs/ext4/page-io.c | 13 ++---
fs/ext4/super.c | 2 +-
fs/jbd2/transaction.c | 8 +--
fs/xfs/kmem.c | 19 +++----
fs/xfs/xfs_buf.c | 14 +++---
include/linux/gfp.h | 6 ++-
17 files changed, 113 insertions(+), 99 deletions(-)
--
Signature
When alloc_pages_bulk_array() is called on an array that is partially
allocated, the level of effort to get a single page is less than when
the array was completely unallocated. This behaviour is inconsistent,
but now fixed. One effect if this is that __GFP_NOFAIL will not ensure
at least one page is allocated.
Also clarify the expected success rate. __alloc_pages_bulk() will
allocated one page according to @gfp, and may allocate more if that can
be done cheaply. It is assumed that the caller values cheap allocation
where possible and may decide to use what it has got, or to call again
to get more.
Acked-by: Mel Gorman <redacted>
Fixes: 0f87d9d30f21 ("mm/page_alloc: add an array-based interface to the bulk page allocator")
Signed-off-by: NeilBrown <redacted>
---
mm/page_alloc.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
@@ -5292,7 +5297,7 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid,pcp,pcp_list);if(unlikely(!page)){/* Try and get at least one page */-if(!nr_populated)+if(!nr_account)gotofailed_irq;break;}
__GFP_NOFAIL is documented both in gfp.h and memory-allocation.rst.
The details are not entirely consistent.
This patch ensures both places state that:
- there is a risk of deadlock with reclaim/writeback/oom-kill
- it should only be used when there is no real alternative
- it is preferable to an endless loop
- it is strongly discourages for costly-order allocations.
Signed-off-by: NeilBrown <redacted>
---
Documentation/core-api/memory-allocation.rst | 25 ++++++++++++++++++++++++-
include/linux/gfp.h | 6 +++++-
2 files changed, 29 insertions(+), 2 deletions(-)
@@ -126,7 +126,30 @@ or another request.*``GFP_KERNEL | __GFP_NOFAIL`` - overrides the default allocator behavior and all allocation requests will loop endlessly until they succeed.- This might be really dangerous especially for larger orders.+ Any attempt to use ``__GFP_NOFAIL`` for allocations larger than+ order-1 (2 pages) will trigger a warning.++ Use of ``__GFP_NOFAIL`` can cause deadlocks so it should only be used+ when there is no alternative, and then should be used with caution.+ Deadlocks can happen if the calling process holds any resources+ (e.g. locks) which might be needed for memory reclaim or write-back,+ or which might prevent a process killed by the OOM killer from+ successfully exiting. Where possible, locks should be released+ before using ``__GFP_NOFAIL``.++ While this flag is best avoided, it is still preferable to endless+ loops around the allocator. Endless loops may still be used when+ there is a need to test for the process being killed+ (fatal_signal_pending(current)).++*``GFP_NOFS | __GFP_NOFAIL`` - Loop endlessly instead of failing+ when performing allocations in file system code. The same guidance+ as for ``GFP_KERNEL | __GFP_NOFAIL`` applies with extra emphasis on+ the possibility of deadlocks. ``GFP_NOFS`` often implies that+ filesystem locks are held which might lead to blocking reclaim.+ Preemptively flushing or reclaiming memory associated with such+ locks might be appropriate before requesting a ``__GFP_NOFAIL``+ allocation. Selecting memory allocator ==========================
Indefinite loops waiting for memory allocation are discouraged by
documentation in gfp.h which says the use of __GFP_NOFAIL that it
is definitely preferable to use the flag rather than opencode endless
loop around allocator.
Such loops that use congestion_wait() are particularly unwise as
congestion_wait() is indistinguishable from
schedule_timeout_uninterruptible() in practice - and should be
deprecated.
So this patch changes the two loops in ext4_ext_truncate() to use
__GFP_NOFAIL instead of looping.
As the allocation is multiple layers deeper in the call stack, this
requires passing the EXT4_EX_NOFAIL flag down and handling it in various
places.
__ext4_journal_start_sb() is given a new 'gfp_mask' argument which is
passed through to jbd2__journal_start() and always receives GFP_NOFS
except when called through ext4_journal_state_with_revoke from
ext4_ext_remove_space(), which now can include __GFP_NOFAIL.
jbd2__journal_start() is enhanced so that the gfp_t flags passed are
used for *all* allocations.
Signed-off-by: NeilBrown <redacted>
---
fs/ext4/ext4.h | 2 +-
fs/ext4/ext4_jbd2.c | 4 ++-
fs/ext4/ext4_jbd2.h | 14 +++++++-----
fs/ext4/extents.c | 53 +++++++++++++++++++---------------------------
fs/ext4/extents_status.c | 35 +++++++++++++++++-------------
fs/ext4/extents_status.h | 2 +-
fs/ext4/ialloc.c | 3 ++-
fs/ext4/indirect.c | 2 +-
fs/ext4/inode.c | 6 +++--
fs/ext4/ioctl.c | 4 ++-
fs/ext4/super.c | 2 +-
fs/jbd2/transaction.c | 8 +++----
12 files changed, 67 insertions(+), 68 deletions(-)
@@ -1488,7 +1488,7 @@ static int ext4_ext_search_left(struct inode *inode,staticintext4_ext_search_right(structinode*inode,structext4_ext_path*path,ext4_lblk_t*logical,ext4_fsblk_t*phys,-structext4_extent*ret_ex)+structext4_extent*ret_ex,intnofail){structbuffer_head*bh=NULL;structext4_extent_header*eh;
@@ -1565,7 +1565,7 @@ static int ext4_ext_search_right(struct inode *inode,while(++depth<path->p_depth){/* subtract from p_depth to get proper eh_depth */bh=read_extent_tree_block(inode,block,-path->p_depth-depth,0);+path->p_depth-depth,nofail);if(IS_ERR(bh))returnPTR_ERR(bh);eh=ext_block_hdr(bh);
@@ -1574,7 +1574,7 @@ static int ext4_ext_search_right(struct inode *inode,put_bh(bh);}-bh=read_extent_tree_block(inode,block,path->p_depth-depth,0);+bh=read_extent_tree_block(inode,block,path->p_depth-depth,nofail);if(IS_ERR(bh))returnPTR_ERR(bh);eh=ext_block_hdr(bh);
@@ -2781,6 +2781,10 @@ int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,structpartial_clusterpartial;handle_t*handle;inti=0,err=0;+gfp_tgfp_mask=GFP_NOFS;++if(nofail)+gfp_mask|=__GFP_NOFAIL;partial.pclu=0;partial.lblk=0;
@@ -2789,7 +2793,8 @@ int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,ext_debug(inode,"truncate since %u to %u\n",start,end);/* probably first extent we're gonna free will be last in block */-handle=ext4_journal_start_with_revoke(inode,EXT4_HT_TRUNCATE,+handle=ext4_journal_start_with_revoke(gfp_mask,+inode,EXT4_HT_TRUNCATE,depth+1,ext4_free_metadata_revoke_credits(inode->i_sb,depth));if(IS_ERR(handle))
@@ -2877,7 +2882,7 @@ int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,*/lblk=ex_end+1;err=ext4_ext_search_right(inode,path,&lblk,&pblk,-NULL);+NULL,nofail);if(err<0)gotoout;if(pblk){
@@ -2899,10 +2904,6 @@ int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,}else{path=kcalloc(depth+1,sizeof(structext4_ext_path),GFP_NOFS|__GFP_NOFAIL);-if(path==NULL){-ext4_journal_stop(handle);-return-ENOMEM;-}path[0].p_maxdepth=path[0].p_depth=depth;path[0].p_hdr=ext_inode_hdr(inode);i=0;
@@ -2955,7 +2956,7 @@ int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,memset(path+i+1,0,sizeof(*path));bh=read_extent_tree_block(inode,ext4_idx_pblock(path[i].p_idx),depth-i-1,-EXT4_EX_NOCACHE);+EXT4_EX_NOCACHE|nofail);if(IS_ERR(bh)){/* should we reset i_size? */err=PTR_ERR(bh);
@@ -4186,7 +4187,7 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,if(err)gotoout;ar.lright=map->m_lblk;-err=ext4_ext_search_right(inode,path,&ar.lright,&ar.pright,&ex2);+err=ext4_ext_search_right(inode,path,&ar.lright,&ar.pright,&ex2,0);if(err<0)gotoout;
@@ -4368,23 +4369,13 @@ int ext4_ext_truncate(handle_t *handle, struct inode *inode)last_block=(inode->i_size+sb->s_blocksize-1)>>EXT4_BLOCK_SIZE_BITS(sb);-retry:err=ext4_es_remove_extent(inode,last_block,-EXT_MAX_BLOCKS-last_block);-if(err==-ENOMEM){-cond_resched();-congestion_wait(BLK_RW_ASYNC,HZ/50);-gotoretry;-}+EXT_MAX_BLOCKS-last_block,+EXT4_EX_NOFAIL);if(err)returnerr;-retry_remove_space:-err=ext4_ext_remove_space(inode,last_block,EXT_MAX_BLOCKS-1);-if(err==-ENOMEM){-cond_resched();-congestion_wait(BLK_RW_ASYNC,HZ/50);-gotoretry_remove_space;-}+err=ext4_ext_remove_space(inode,last_block,EXT_MAX_BLOCKS-1,+EXT4_EX_NOFAIL);returnerr;}
@@ -476,9 +476,9 @@ static int start_this_handle(journal_t *journal, handle_t *handle,}/* Allocate a new handle. This should probably be in a slab... */-statichandle_t*new_handle(intnblocks)+statichandle_t*new_handle(intnblocks,gfp_tgfp){-handle_t*handle=jbd2_alloc_handle(GFP_NOFS);+handle_t*handle=jbd2_alloc_handle(gfp);if(!handle)returnNULL;handle->h_total_credits=nblocks;
@@ -505,13 +505,13 @@ handle_t *jbd2__journal_start(journal_t *journal, int nblocks, int rsv_blocks,nblocks+=DIV_ROUND_UP(revoke_records,journal->j_revoke_records_per_block);-handle=new_handle(nblocks);+handle=new_handle(nblocks,gfp_mask);if(!handle)returnERR_PTR(-ENOMEM);if(rsv_blocks){handle_t*rsv_handle;-rsv_handle=new_handle(rsv_blocks);+rsv_handle=new_handle(rsv_blocks,gfp_mask);if(!rsv_handle){jbd2_free_handle(handle);returnERR_PTR(-ENOMEM);
congestion_wait() is indistinguishable from
schedule_timeout_uninterruptible(). It is best avoided and should be
deprecated.
It is not needed in ext4_bio_write_page(). There are two cases.
If there are no ->io_bio yet, then it is appropriate to use __GFP_NOFAIL
which does the waiting in a better place. The code already uses this
flag on the second attempt. This patch changes to it always use that
flag for this case.
If there *are* ->io_bio (in which case the allocation was non-blocking)
we submit the io and return the first case. No waiting is needed in
this case.
So remove the congestion_wait() call, and simplify the code so that the
two cases are somewhat clearer.
Remove the "if (io->io_bio)" before calling ext4_io_submit() as that
test is performed internally by that function.
Signed-off-by: NeilBrown <redacted>
---
fs/ext4/page-io.c | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
Documentation commment in gfp.h discourages indefinite retry loops on
ENOMEM and says of __GFP_NOFAIL that it
is definitely preferable to use the flag rather than opencode
endless loop around allocator.
So remove the loop, instead specifying __GFP_NOFAIL if KM_MAYFAIL was
not given.
As we no longer have the opportunity to report a warning after some
failures, clear __GFP_NOWARN so that the default warning (rate-limited
to 1 ever 10 seconds) will be reported instead.
Signed-off-by: NeilBrown <redacted>
---
fs/xfs/kmem.c | 19 ++++++-------------
1 file changed, 6 insertions(+), 13 deletions(-)
Documentation commment in gfp.h discourages indefinite retry loops on
ENOMEM and says of __GFP_NOFAIL that it
is definitely preferable to use the flag rather than opencode
endless loop around allocator.
congestion_wait() is indistinguishable from
schedule_timeout_uninterruptible() in practice and it is not a good way
to wait for memory to become available.
So add __GFP_NOFAIL to gfp if failure is not an option, and remove the
congestion_wait(). We now only loop when failure is an option, and
alloc_bulk_pages_array() made some progres, but not enough.
Signed-off-by: NeilBrown <redacted>
---
fs/xfs/xfs_buf.c | 14 ++++++--------
1 file changed, 6 insertions(+), 8 deletions(-)
@@ -352,7 +352,7 @@ xfs_buf_alloc_pages(if(flags&XBF_READ_AHEAD)gfp_mask|=__GFP_NORETRY;else-gfp_mask|=GFP_NOFS;+gfp_mask|=GFP_NOFS|__GFP_NOFAIL;/* Make sure that we have a page list */bp->b_page_count=DIV_ROUND_UP(BBTOB(bp->b_length),PAGE_SIZE);
I'm top-posting to cc Jesper with full context of the patch. I don't
have a problem with this patch other than the Fixes: being a bit
marginal, I should have acked as Mel Gorman [off-list ref] and the
@gfp in the comment should have been @gfp_mask.
However, an assumption the API design made was that it should fail fast
if memory is not quickly available but have at least one page in the
array. I don't think the network use case cares about the situation where
the array is already populated but I'd like Jesper to have the opportunity
to think about it. It's possible he would prefer it's explicit and the
check becomes
(!nr_populated || ((gfp_mask & __GFP_NOFAIL) && !nr_account)) to
state that __GFP_NOFAIL users are willing to take a potential latency
penalty if the array is already partially populated but !__GFP_NOFAIL
users would prefer fail-fast behaviour. I'm on the fence because while
I wrote the implementation, it was based on other peoples requirements.
On Fri, Sep 17, 2021 at 12:56:57PM +1000, NeilBrown wrote:
quoted hunk
When alloc_pages_bulk_array() is called on an array that is partially
allocated, the level of effort to get a single page is less than when
the array was completely unallocated. This behaviour is inconsistent,
but now fixed. One effect if this is that __GFP_NOFAIL will not ensure
at least one page is allocated.
Also clarify the expected success rate. __alloc_pages_bulk() will
allocated one page according to @gfp, and may allocate more if that can
be done cheaply. It is assumed that the caller values cheap allocation
where possible and may decide to use what it has got, or to call again
to get more.
Acked-by: Mel Gorman <redacted>
Fixes: 0f87d9d30f21 ("mm/page_alloc: add an array-based interface to the bulk page allocator")
Signed-off-by: NeilBrown <redacted>
---
mm/page_alloc.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
@@ -5292,7 +5297,7 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid,pcp,pcp_list);if(unlikely(!page)){/* Try and get at least one page */-if(!nr_populated)+if(!nr_account)gotofailed_irq;break;}
From: Dave Chinner <david@fromorbit.com> Date: 2021-09-17 21:45:38
On Fri, Sep 17, 2021 at 12:56:57PM +1000, NeilBrown wrote:
quoted hunk
Documentation commment in gfp.h discourages indefinite retry loops on
ENOMEM and says of __GFP_NOFAIL that it
is definitely preferable to use the flag rather than opencode
endless loop around allocator.
So remove the loop, instead specifying __GFP_NOFAIL if KM_MAYFAIL was
not given.
As we no longer have the opportunity to report a warning after some
failures, clear __GFP_NOWARN so that the default warning (rate-limited
to 1 ever 10 seconds) will be reported instead.
Signed-off-by: NeilBrown <redacted>
---
fs/xfs/kmem.c | 19 ++++++-------------
1 file changed, 6 insertions(+), 13 deletions(-)
@@ -11,21 +11,14 @@void*kmem_alloc(size_tsize,xfs_km_flags_tflags){-intretries=0;gfp_tlflags=kmem_flags_convert(flags);-void*ptr;trace_kmem_alloc(size,flags,_RET_IP_);-do{-ptr=kmalloc(size,lflags);-if(ptr||(flags&KM_MAYFAIL))-returnptr;-if(!(++retries%100))-xfs_err(NULL,-"%s(%u) possible memory allocation deadlock size %u in %s (mode:0x%x)",-current->comm,current->pid,-(unsignedint)size,__func__,lflags);-congestion_wait(BLK_RW_ASYNC,HZ/50);-}while(1);+if(!(flags&KM_MAYFAIL)){+lflags|=__GFP_NOFAIL;+lflags&=~__GFP_NOWARN;+}
This logic should really be in kmem_flags_convert() where the gfp
flags are set up. kmem_flags_convert() is only called by
kmem_alloc() now so you should just be able to hack that logic
to do exactly what is necessary.
FWIW, We've kinda not been caring about warts in this code because
the next step for kmem_alloc is to remove kmem_alloc/kmem_zalloc
completely and replace all the callers with kmalloc/kzalloc being
passed the correct gfp flags. There's about 30 kmem_alloc() callers
and 45 kmem_zalloc() calls left to be converted before kmem.[ch] can
go away completely....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
I'm top-posting to cc Jesper with full context of the patch. I don't
have a problem with this patch other than the Fixes: being a bit
marginal, I should have acked as Mel Gorman [off-list ref] and the
@gfp in the comment should have been @gfp_mask.
However, an assumption the API design made was that it should fail fast
if memory is not quickly available but have at least one page in the
array. I don't think the network use case cares about the situation where
the array is already populated but I'd like Jesper to have the opportunity
to think about it. It's possible he would prefer it's explicit and the
check becomes
(!nr_populated || ((gfp_mask & __GFP_NOFAIL) && !nr_account)) to
state that __GFP_NOFAIL users are willing to take a potential latency
penalty if the array is already partially populated but !__GFP_NOFAIL
users would prefer fail-fast behaviour. I'm on the fence because while
I wrote the implementation, it was based on other peoples requirements.
I can see that it could be desirable to not try too hard when we already
have pages allocated, but maybe the best way to achieve that is for the
called to clear __GFP_RECLAIM in that case.
Alternately, callers that really want the __GFP_RECLAIM and __GFP_NOFAIL
flags to be honoured could ensure that the array passed in is empty.
That wouldn't be difficult (for current callers).
In either case, the documentation should make it clear which flags are
honoured when.
Let's see what Jesper has to say.
Thanks,
NeilBrown
On Fri, Sep 17, 2021 at 12:56:57PM +1000, NeilBrown wrote:
quoted
When alloc_pages_bulk_array() is called on an array that is partially
allocated, the level of effort to get a single page is less than when
the array was completely unallocated. This behaviour is inconsistent,
but now fixed. One effect if this is that __GFP_NOFAIL will not ensure
at least one page is allocated.
Also clarify the expected success rate. __alloc_pages_bulk() will
allocated one page according to @gfp, and may allocate more if that can
be done cheaply. It is assumed that the caller values cheap allocation
where possible and may decide to use what it has got, or to call again
to get more.
Acked-by: Mel Gorman <redacted>
Fixes: 0f87d9d30f21 ("mm/page_alloc: add an array-based interface to the bulk page allocator")
Signed-off-by: NeilBrown <redacted>
---
mm/page_alloc.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
@@ -5292,7 +5297,7 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid,pcp,pcp_list);if(unlikely(!page)){/* Try and get at least one page */-if(!nr_populated)+if(!nr_account)gotofailed_irq;break;}
I'm top-posting to cc Jesper with full context of the patch. I don't
have a problem with this patch other than the Fixes: being a bit
marginal, I should have acked as Mel Gorman [off-list ref] and the
@gfp in the comment should have been @gfp_mask.
However, an assumption the API design made was that it should fail fast
if memory is not quickly available but have at least one page in the
array. I don't think the network use case cares about the situation where
the array is already populated but I'd like Jesper to have the opportunity
to think about it. It's possible he would prefer it's explicit and the
check becomes
(!nr_populated || ((gfp_mask & __GFP_NOFAIL) && !nr_account)) to
Note that AFAICS nr_populated is an incomplete piece of information, as we
initially only count pages in the page_array as nr_populated up to the first
NULL pointer. So even before Neil's patch we could decide to allocate even
if there are pre-existing pages, but placed later in the array. Which could
be rather common if the array consumer starts from index 0? So with Neil's
patch this at least becomes consistent, while the check suggested by Mel
leaves there the weird dependency on where pre-existing pages appear in the
page_array.
__GFP_NOFAIL is documented both in gfp.h and memory-allocation.rst.
The details are not entirely consistent.
This patch ensures both places state that:
- there is a risk of deadlock with reclaim/writeback/oom-kill
- it should only be used when there is no real alternative
- it is preferable to an endless loop
- it is strongly discourages for costly-order allocations.
Signed-off-by: NeilBrown <redacted>
We define costly as 3, not 1. But sure it's best to avoid even order>0 for
__GFP_NOFAIL. Advising order>1 seems arbitrary though?
This is not completely arbitrary. We have a warning for any higher order
allocation.
rmqueue:
WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
I do agree that "Using this flag for higher order allocations is
_highly_ discouraged.
We define costly as 3, not 1. But sure it's best to avoid even order>0 for
__GFP_NOFAIL. Advising order>1 seems arbitrary though?
This is not completely arbitrary. We have a warning for any higher order
allocation.
rmqueue:
WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
Oh, I missed that.
quoted
I do agree that "Using this flag for higher order allocations is
_highly_ discouraged.
Well, with the warning in place this is effectively forbidden, not just
discouraged.
Yup, especially as it doesn't obey __GFP_NOWARN.
See commit de2860f46362 ("mm: Add kvrealloc()") as a direct result
of unwittingly tripping over this warning when adding __GFP_NOFAIL
annotations to replace open coded high-order kmalloc loops that have
been in place for a couple of decades without issues.
Personally I think that the way __GFP_NOFAIL is first of all
recommended over open coded loops and then only later found to be
effectively forbidden and needing to be replaced with open coded
loops to be a complete mess.
Not to mention on the impossibility of using __GFP_NOFAIL with
kvmalloc() calls. Just what do we expect kmalloc_node(__GFP_NORETRY
| __GFP_NOFAIL) to do, exactly?
So, effectively, we have to open-code around kvmalloc() in
situations where failure is not an option. Even if we pass
__GFP_NOFAIL to __vmalloc(), it isn't guaranteed to succeed because
of the "we won't honor gfp flags passed to __vmalloc" semantics it
has.
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
That leads to us having to go back to writing extremely custom open
coded loops to avoid awful high-order kmalloc direct reclaim
behaviour and still fall back to vmalloc and to still handle NOFAIL
semantics we need:
https://lore.kernel.org/linux-xfs/20210902095927.911100-8-david@fromorbit.com/
So, really, the problems are much deeper here than just badly
documented, catch-22 rules for __GFP_NOFAIL - we can't even use
__GFP_NOFAIL consistently across the allocation APIs because it
changes allocation behaviours in unusable, self-defeating ways....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
We define costly as 3, not 1. But sure it's best to avoid even order>0 for
__GFP_NOFAIL. Advising order>1 seems arbitrary though?
This is not completely arbitrary. We have a warning for any higher order
allocation.
rmqueue:
WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
Oh, I missed that.
quoted
I do agree that "Using this flag for higher order allocations is
_highly_ discouraged.
Well, with the warning in place this is effectively forbidden, not just
discouraged.
Yup, especially as it doesn't obey __GFP_NOWARN.
See commit de2860f46362 ("mm: Add kvrealloc()") as a direct result
of unwittingly tripping over this warning when adding __GFP_NOFAIL
annotations to replace open coded high-order kmalloc loops that have
been in place for a couple of decades without issues.
Personally I think that the way __GFP_NOFAIL is first of all
recommended over open coded loops and then only later found to be
effectively forbidden and needing to be replaced with open coded
loops to be a complete mess.
Well, there are two things. Opencoding something that _can_ be replaced
by __GFP_NOFAIL and those that cannot because the respective allocator
doesn't really support that semantic. kvmalloc is explicit about that
IIRC. If you have a better way to consolidate the documentation then I
am all for it.
Not to mention on the impossibility of using __GFP_NOFAIL with
kvmalloc() calls. Just what do we expect kmalloc_node(__GFP_NORETRY
| __GFP_NOFAIL) to do, exactly?
This combination doesn't make any sense. Like others. Do you want us to
list all combinations that make sense?
So, effectively, we have to open-code around kvmalloc() in
situations where failure is not an option. Even if we pass
__GFP_NOFAIL to __vmalloc(), it isn't guaranteed to succeed because
of the "we won't honor gfp flags passed to __vmalloc" semantics it
has.
yes vmalloc doesn't support nofail semantic and it is not really trivial
to craft it there.
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
GFP_NOFS (without using the scope API) has the same problem as NOFAIL in
the vmalloc. Hence it is not supported. If you use the scope API then
you can GFP_KERNEL for kvmalloc. This is clumsy but I am not sure how to
define these conditions in a more sensible way. Special case NOFS if the
scope api is in use? Why do you want an explicit NOFS then?
It would be more productive to get to MM people rather than rant on a
xfs specific patchse. Anyway, I can see a kvmalloc mode where the
kmalloc allocation would be really a very optimistic one - like your
effectively GFP_NOWAIT. Nobody has requested such a mode until now and I
am not sure how we would sensibly describe that by a gfp mask.
Btw. your GFP_NOWAIT | __GFP_NORETRY combination doesn't make any sense
in the allocator context as the later is a reclaim mofifier which
doesn't get applied when the reclaim is disabled (in your case by flags
&= ~__GFP_DIRECT_RECLAIM).
GFP flags are not that easy to build a coherent and usable apis.
Something we carry as a baggage for a long time.
So, really, the problems are much deeper here than just badly
documented, catch-22 rules for __GFP_NOFAIL - we can't even use
__GFP_NOFAIL consistently across the allocation APIs because it
changes allocation behaviours in unusable, self-defeating ways....
GFP_NOFAIL sucks. Not all allocator can follow it for practical
reasons. You are welcome to help document those awkward corner cases or
fix them up if you have a good idea how.
Thanks!
--
Michal Hocko
SUSE Labs
We define costly as 3, not 1. But sure it's best to avoid even order>0 for
__GFP_NOFAIL. Advising order>1 seems arbitrary though?
This is not completely arbitrary. We have a warning for any higher order
allocation.
rmqueue:
WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
Oh, I missed that.
quoted
I do agree that "Using this flag for higher order allocations is
_highly_ discouraged.
Well, with the warning in place this is effectively forbidden, not just
discouraged.
Yup, especially as it doesn't obey __GFP_NOWARN.
See commit de2860f46362 ("mm: Add kvrealloc()") as a direct result
of unwittingly tripping over this warning when adding __GFP_NOFAIL
annotations to replace open coded high-order kmalloc loops that have
been in place for a couple of decades without issues.
Personally I think that the way __GFP_NOFAIL is first of all
recommended over open coded loops and then only later found to be
effectively forbidden and needing to be replaced with open coded
loops to be a complete mess.
Well, there are two things. Opencoding something that _can_ be replaced
by __GFP_NOFAIL and those that cannot because the respective allocator
doesn't really support that semantic. kvmalloc is explicit about that
IIRC. If you have a better way to consolidate the documentation then I
am all for it.
I think one thing that might help make the documentation better is to
explicitly state *why* __GFP_NOFAIL is better than a loop.
It occurs to me that
while (!(p = kmalloc(sizeof(*p), GFP_KERNEL));
would behave much the same as adding __GFP_NOFAIL and dropping the
'while'. So why not? I certainly cannot see the need to add any delay
to this loop as kmalloc does a fair bit of sleeping when permitted.
I understand that __GFP_NOFAIL allows page_alloc to dip into reserves,
but Mel holds that up as a reason *not* to use __GFP_NOFAIL as it can
impact on other subsystems. Why not just let the caller decide if they
deserve the boost, but oring in __GFP_ATOMIC or __GFP_MEMALLOC as
appropriate.
I assume there is a good reason. I vaguely remember the conversation
that lead to __GFP_NOFAIL being introduced. I just cannot remember or
deduce what the reason is. So it would be great to have it documented.
quoted
Not to mention on the impossibility of using __GFP_NOFAIL with
kvmalloc() calls. Just what do we expect kmalloc_node(__GFP_NORETRY
| __GFP_NOFAIL) to do, exactly?
This combination doesn't make any sense. Like others. Do you want us to
list all combinations that make sense?
I've been wondering about that. There seem to be sets of flags that are
mutually exclusive. It is as though gfp_t is a struct of a few enums.
0, DMA32, DMA, HIGHMEM
0, FS, IO
0, ATOMIC, MEMALLOC, NOMEMALLOC, HIGH
NORETRY, RETRY_MAYFAIL, 0, NOFAIL
0, KSWAPD_RECLAIM, DIRECT_RECLAIM
0, THISNODE, HARDWALL
In a few cases there seem to be 3 bits where there are only 4 possibly
combinations, so 2 bits would be enough. There is probably no real
value is squeezing these into 2 bits, but clearly documenting the groups
surely wouldn't hurt. Particularly highlighting the difference between
related bits would help.
The set with 'ATOMIC' is hard to wrap my mind around.
They relate to ALLOC_HIGH and ALLOC_HARDER, but also to WMARK_NIN,
WMARK_LOW, WMARK_HIGH ... I think.
I wonder if FS,IO is really in the same set as DIRECT_RECLAIM as they
all affect reclaim. Maybe FS and IO are only relevan if DIRECT_RECLAIM
is set?
I'd love to know that to expect if neither RETRY_MAYFAIL or NOFAIL is
set. I guess it can fail, but it still tries harder than if
RETRY_MAYFAIL is set....
Ahhhh... I found some documentation which mentions that RETRY_MAYFAIL
doesn't trigger the oom killer. Is that it? So RETRY_NOKILLOOM might be
a better name?
quoted
So, effectively, we have to open-code around kvmalloc() in
situations where failure is not an option. Even if we pass
__GFP_NOFAIL to __vmalloc(), it isn't guaranteed to succeed because
of the "we won't honor gfp flags passed to __vmalloc" semantics it
has.
yes vmalloc doesn't support nofail semantic and it is not really trivial
to craft it there.
quoted
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
GFP_NOFS (without using the scope API) has the same problem as NOFAIL in
the vmalloc. Hence it is not supported. If you use the scope API then
you can GFP_KERNEL for kvmalloc. This is clumsy but I am not sure how to
define these conditions in a more sensible way. Special case NOFS if the
scope api is in use? Why do you want an explicit NOFS then?
It would seem to make sense for kvmalloc to WARN_ON if it is passed
flags that does not allow it to use vmalloc.
Such callers could then know they can either change to a direct
kmalloc(), or change flags. Silently ignoring the 'v' in the function
name sees like a poor choice.
Thanks,
NeilBrown
It would be more productive to get to MM people rather than rant on a
xfs specific patchse. Anyway, I can see a kvmalloc mode where the
kmalloc allocation would be really a very optimistic one - like your
effectively GFP_NOWAIT. Nobody has requested such a mode until now and I
am not sure how we would sensibly describe that by a gfp mask.
Btw. your GFP_NOWAIT | __GFP_NORETRY combination doesn't make any sense
in the allocator context as the later is a reclaim mofifier which
doesn't get applied when the reclaim is disabled (in your case by flags
&= ~__GFP_DIRECT_RECLAIM).
GFP flags are not that easy to build a coherent and usable apis.
Something we carry as a baggage for a long time.
quoted
So, really, the problems are much deeper here than just badly
documented, catch-22 rules for __GFP_NOFAIL - we can't even use
__GFP_NOFAIL consistently across the allocation APIs because it
changes allocation behaviours in unusable, self-defeating ways....
GFP_NOFAIL sucks. Not all allocator can follow it for practical
reasons. You are welcome to help document those awkward corner cases or
fix them up if you have a good idea how.
Thanks!
--
Michal Hocko
SUSE Labs
We define costly as 3, not 1. But sure it's best to avoid even order>0 for
__GFP_NOFAIL. Advising order>1 seems arbitrary though?
This is not completely arbitrary. We have a warning for any higher order
allocation.
rmqueue:
WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
Oh, I missed that.
quoted
I do agree that "Using this flag for higher order allocations is
_highly_ discouraged.
Well, with the warning in place this is effectively forbidden, not just
discouraged.
Yup, especially as it doesn't obey __GFP_NOWARN.
See commit de2860f46362 ("mm: Add kvrealloc()") as a direct result
of unwittingly tripping over this warning when adding __GFP_NOFAIL
annotations to replace open coded high-order kmalloc loops that have
been in place for a couple of decades without issues.
Personally I think that the way __GFP_NOFAIL is first of all
recommended over open coded loops and then only later found to be
effectively forbidden and needing to be replaced with open coded
loops to be a complete mess.
Well, there are two things. Opencoding something that _can_ be replaced
by __GFP_NOFAIL and those that cannot because the respective allocator
doesn't really support that semantic. kvmalloc is explicit about that
IIRC. If you have a better way to consolidate the documentation then I
am all for it.
I think one thing that might help make the documentation better is to
explicitly state *why* __GFP_NOFAIL is better than a loop.
It occurs to me that
while (!(p = kmalloc(sizeof(*p), GFP_KERNEL));
would behave much the same as adding __GFP_NOFAIL and dropping the
'while'. So why not? I certainly cannot see the need to add any delay
to this loop as kmalloc does a fair bit of sleeping when permitted.
I understand that __GFP_NOFAIL allows page_alloc to dip into reserves,
but Mel holds that up as a reason *not* to use __GFP_NOFAIL as it can
impact on other subsystems.
__GFP_NOFAIL usage is a risk on its own. It is a hard requirement that
the allocator cannot back off. So it has to absolutely everything to
suceed. Whether it cheats and dips into reserves or not is a mere
implementation detail and a subject to the specific implementation.
Why not just let the caller decide if they
deserve the boost, but oring in __GFP_ATOMIC or __GFP_MEMALLOC as
appropriate.
They can do that. Explicit access to memory reserves is allowed unless
it is explicitly forbidden by NOMEMALLOC flag.
I assume there is a good reason. I vaguely remember the conversation
that lead to __GFP_NOFAIL being introduced. I just cannot remember or
deduce what the reason is. So it would be great to have it documented.
The basic reason is that if the allocator knows this is must suceed
allocation request then it can prioritize it in some way. A dumb kmalloc
loop as you pictured it is likely much less optimal in that sense, isn't
it? Compare that to mempool allocator which is non failing as well but
it has some involved handling and that is certainly not a good fit for
__GFP_NOFAIL in the page allocator.
quoted
quoted
Not to mention on the impossibility of using __GFP_NOFAIL with
kvmalloc() calls. Just what do we expect kmalloc_node(__GFP_NORETRY
| __GFP_NOFAIL) to do, exactly?
This combination doesn't make any sense. Like others. Do you want us to
list all combinations that make sense?
I've been wondering about that. There seem to be sets of flags that are
mutually exclusive. It is as though gfp_t is a struct of a few enums.
0, DMA32, DMA, HIGHMEM
0, FS, IO
0, ATOMIC, MEMALLOC, NOMEMALLOC, HIGH
NORETRY, RETRY_MAYFAIL, 0, NOFAIL
0, KSWAPD_RECLAIM, DIRECT_RECLAIM
0, THISNODE, HARDWALL
In a few cases there seem to be 3 bits where there are only 4 possibly
combinations, so 2 bits would be enough. There is probably no real
value is squeezing these into 2 bits, but clearly documenting the groups
surely wouldn't hurt. Particularly highlighting the difference between
related bits would help.
Don't we have that already? We have them grouped by placement,
watermarks, reclaim and action modifiers. Then we have useful
combinations. I believe we can always improve on that and I am always
ready to listen here.
The set with 'ATOMIC' is hard to wrap my mind around.
They relate to ALLOC_HIGH and ALLOC_HARDER, but also to WMARK_NIN,
WMARK_LOW, WMARK_HIGH ... I think.
ALLOC* and WMARK* is an internal allocator concept and I believe users
of gfp flags shouldn't really care or even know those exist.
I wonder if FS,IO is really in the same set as DIRECT_RECLAIM as they
all affect reclaim. Maybe FS and IO are only relevan if DIRECT_RECLAIM
is set?
yes, this indeed the case. Page allocator doesn't go outside of its
proper without the direct reclaim.
I'd love to know that to expect if neither RETRY_MAYFAIL or NOFAIL is
set. I guess it can fail, but it still tries harder than if
RETRY_MAYFAIL is set....
Ahhhh... I found some documentation which mentions
The reclaim behavior is described along with the respective modifiers. I
believe we can thank you for this structure as you were the primary
driving force to clarify the behavior.
that RETRY_MAYFAIL
doesn't trigger the oom killer. Is that it? So RETRY_NOKILLOOM might be
a better name?
Again the those are implementation details and I am not sure we really
want to bother users with all of them. This wold quickly become hairy
and likely even outdated after some time. The documentation tries to
describe different levels of involvement. NOWAIT - no direct reclaim,
NORETRY - only a light attempt to reclaim, RETRY_MAYFAIL - try as hard
as feasible, NOFAIL - cannot really fail.
If we can improve the wording I am all for it.
quoted
quoted
So, effectively, we have to open-code around kvmalloc() in
situations where failure is not an option. Even if we pass
__GFP_NOFAIL to __vmalloc(), it isn't guaranteed to succeed because
of the "we won't honor gfp flags passed to __vmalloc" semantics it
has.
yes vmalloc doesn't support nofail semantic and it is not really trivial
to craft it there.
quoted
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
GFP_NOFS (without using the scope API) has the same problem as NOFAIL in
the vmalloc. Hence it is not supported. If you use the scope API then
you can GFP_KERNEL for kvmalloc. This is clumsy but I am not sure how to
define these conditions in a more sensible way. Special case NOFS if the
scope api is in use? Why do you want an explicit NOFS then?
It would seem to make sense for kvmalloc to WARN_ON if it is passed
flags that does not allow it to use vmalloc.
vmalloc is certainly not the hottest path in the kernel so I wouldn't be
opposed. One should be careful that WARN_ON is effectively BUG_ON in
some configurations but we are sinners from that perspective all over
the place...
Thanks!
--
Michal Hocko
SUSE Labs
We define costly as 3, not 1. But sure it's best to avoid even order>0 for
__GFP_NOFAIL. Advising order>1 seems arbitrary though?
This is not completely arbitrary. We have a warning for any higher order
allocation.
rmqueue:
WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
Oh, I missed that.
quoted
I do agree that "Using this flag for higher order allocations is
_highly_ discouraged.
Well, with the warning in place this is effectively forbidden, not just
discouraged.
Yup, especially as it doesn't obey __GFP_NOWARN.
See commit de2860f46362 ("mm: Add kvrealloc()") as a direct result
of unwittingly tripping over this warning when adding __GFP_NOFAIL
annotations to replace open coded high-order kmalloc loops that have
been in place for a couple of decades without issues.
Personally I think that the way __GFP_NOFAIL is first of all
recommended over open coded loops and then only later found to be
effectively forbidden and needing to be replaced with open coded
loops to be a complete mess.
Well, there are two things. Opencoding something that _can_ be replaced
by __GFP_NOFAIL and those that cannot because the respective allocator
doesn't really support that semantic. kvmalloc is explicit about that
IIRC. If you have a better way to consolidate the documentation then I
am all for it.
I think one thing that might help make the documentation better is to
explicitly state *why* __GFP_NOFAIL is better than a loop.
It occurs to me that
while (!(p = kmalloc(sizeof(*p), GFP_KERNEL));
would behave much the same as adding __GFP_NOFAIL and dropping the
'while'. So why not? I certainly cannot see the need to add any delay
to this loop as kmalloc does a fair bit of sleeping when permitted.
I understand that __GFP_NOFAIL allows page_alloc to dip into reserves,
but Mel holds that up as a reason *not* to use __GFP_NOFAIL as it can
impact on other subsystems.
__GFP_NOFAIL usage is a risk on its own. It is a hard requirement that
the allocator cannot back off.
No, "allocator cannot back off" isn't a hard requirement for most
GFP_NOFAIL uses. *Not failing the allocation* is the hard
requirement.
How long it takes for the allocation to actually succeed is
irrelevant to most callers, and given that we are replacing loops
that do
while (!(p = kmalloc(sizeof(*p), GFP_KERNEL))
with __GFP_NOFAIL largely indicates that allocation *latency* and/or
deadlocks are not an issue here.
Indeed, if we deadlock in XFS because there is no memory available,
that is *not a problem kmalloc() should be trying to solve*. THe
problem is the caller being unable to handle allocation failure, so
if allocation cannot make progress, that needs to be fixed by the
caller getting rid of the unfailable allocation.
The fact is that we've had these loops in production code for a
couple of decades and these subsystems just aren't failing or
deadlocking with such loops. IOWs, we don't need __GFP_NOFAIL to dig
deep into reserves or drive the system to OOM killing - we just need
to it keep retrying the same allocation until it succeeds.
Put simply, we want "retry forever" semantics to match what
production kernels have been doing for the past couple of decades,
but all we've been given are "never fail" semantics that also do
something different and potentially much more problematic.
Do you see the difference here? __GFP_NOFAIL is not what we
need in the vast majority of cases where it is used. We don't want
the failing allocations to drive the machine hard into critical
reserves, we just want the allocation to -eventually succeed- and if
it doesn't, that's our problem to handle, not kmalloc()....
So it has to absolutely everything to
suceed. Whether it cheats and dips into reserves or not is a mere
implementation detail and a subject to the specific implementation.
My point exactly: that's how the MM interprets __GFP_NOFAIL is supposed to
provide callers with. What we are trying to tell you is that the
semantics associated with __GFP_NOFAIL is not actually what we
require, and it's the current semantics of __GFP_NOFAIL that cause
all the "can't be applied consistently across the entire allocation
APIs" problems....
quoted
quoted
quoted
So, effectively, we have to open-code around kvmalloc() in
situations where failure is not an option. Even if we pass
__GFP_NOFAIL to __vmalloc(), it isn't guaranteed to succeed because
of the "we won't honor gfp flags passed to __vmalloc" semantics it
has.
yes vmalloc doesn't support nofail semantic and it is not really trivial
to craft it there.
Yet retry-forever is trivial to implement across everything:
kvmalloc(size, gfp_mask)
{
gfp_t flags = gfp_mask & ~__GFP_RETRY_FOREVER;
do {
p = __kvmalloc(size, flags)
} while (!p && (gfp_mask & __GFP_RETRY_FOREVER));
return p;
}
That provides "allocation will eventually succeed" semantics just
fine, yes? It doesn't guarantee forwards progress or success, just
that *it won't fail*.
It should be obvious what the difference between "retry forever" and
__GFP_NOFAIL semantics are now, and why we don't actually want
__GFP_NOFAIL. We just want __GFP_RETRY_FOREVER semantics that can be
applied consistently across the entire allocation API regardless of
whatever other flags are passed into the allocation: don't return
until an allocation with the provided semantics succeeds.
quoted
quoted
quoted
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
GFP_NOFS (without using the scope API) has the same problem as NOFAIL in
the vmalloc. Hence it is not supported. If you use the scope API then
you can GFP_KERNEL for kvmalloc. This is clumsy but I am not sure how to
define these conditions in a more sensible way. Special case NOFS if the
scope api is in use? Why do you want an explicit NOFS then?
Exactly my point - this is clumsy and a total mess. I'm not asking
for an explicit GFP_NOFS, just pointing out that the documented
restrictions that "vmalloc can only do GFP_KERNEL allocations" is
completely wrong.
vmalloc()
{
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_save();
p = __vmalloc(gfp_flags | GFP_KERNEL)
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_restore();
}
Yup, that's how simple it is to support GFP_NOFS support in
vmalloc().
This goes along with the argument that "it's impossible to do
GFP_NOFAIL with vmalloc" as I addressed above. These things are not
impossible, but we hide behind "we don't want people to use vmalloc"
as an excuse for having shitty behaviour whilst ignoring that
vmalloc is *heavily used* by core subsystems like filesystems
because they cannot rely on high order allocations succeeding....
It also points out that the scope API is highly deficient.
We can do GFP_NOFS via the scope API, but we can't
do anything else because *there is no scope API for other GFP
flags*.
Why don't we have a GFP_NOFAIL/__GFP_RETRY_FOREVER scope API? That
would save us a lot of bother in XFS. What about GFP_DIRECT_RECLAIM?
I'd really like to turn that off for allocations in the XFS
transaction commit path (as noted already in this thread) because
direct reclaim that can make no progress is actively harmful (as
noted already in this thread)
Like I said - this is more than just bad documentation - the problem
is that the whole allocation API is an inconsistent mess of control
mechanisms to begin with...
quoted
It would seem to make sense for kvmalloc to WARN_ON if it is passed
flags that does not allow it to use vmalloc.
vmalloc is certainly not the hottest path in the kernel so I wouldn't be
opposed.
kvmalloc is most certainly becoming one of the hottest paths in XFS.
IOWs, arguments that "vmalloc is not a hot path" are simply invalid
these days because they are simply untrue. e.g. the profiles I
posted in this thread...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
From: Michal Hocko <mhocko@suse.com> Date: 2021-10-11 11:57:44
On Sat 09-10-21 09:36:49, Dave Chinner wrote:
On Fri, Oct 08, 2021 at 09:48:39AM +0200, Michal Hocko wrote:
quoted
__GFP_NOFAIL usage is a risk on its own. It is a hard requirement that
the allocator cannot back off.
[...]
No, "allocator cannot back off" isn't a hard requirement for most
GFP_NOFAIL uses. *Not failing the allocation* is the hard
requirement.
We are talking about the same thing here I belive. By cannot back off I
really mean cannot fail. Just for the clarification.
How long it takes for the allocation to actually succeed is
irrelevant to most callers, and given that we are replacing loops
that do
while (!(p = kmalloc(sizeof(*p), GFP_KERNEL))
with __GFP_NOFAIL largely indicates that allocation *latency* and/or
deadlocks are not an issue here.
Agreed.
Indeed, if we deadlock in XFS because there is no memory available,
that is *not a problem kmalloc() should be trying to solve*. THe
problem is the caller being unable to handle allocation failure, so
if allocation cannot make progress, that needs to be fixed by the
caller getting rid of the unfailable allocation.
The fact is that we've had these loops in production code for a
couple of decades and these subsystems just aren't failing or
deadlocking with such loops. IOWs, we don't need __GFP_NOFAIL to dig
deep into reserves or drive the system to OOM killing - we just need
to it keep retrying the same allocation until it succeeds.
Put simply, we want "retry forever" semantics to match what
production kernels have been doing for the past couple of decades,
but all we've been given are "never fail" semantics that also do
something different and potentially much more problematic.
Do you see the difference here? __GFP_NOFAIL is not what we
need in the vast majority of cases where it is used. We don't want
the failing allocations to drive the machine hard into critical
reserves, we just want the allocation to -eventually succeed- and if
it doesn't, that's our problem to handle, not kmalloc()....
I can see your point. I do have a recollection that there were some
instance involved where an emergency access to memory reserves helped
in OOM situations.
Anway as I've tried to explain earlier that this all is an
implementation detail users of the flag shouldn't really care about. If
this heuristic is not doing any good then it should be removed.
[...]
quoted
quoted
quoted
quoted
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
GFP_NOFS (without using the scope API) has the same problem as NOFAIL in
the vmalloc. Hence it is not supported. If you use the scope API then
you can GFP_KERNEL for kvmalloc. This is clumsy but I am not sure how to
define these conditions in a more sensible way. Special case NOFS if the
scope api is in use? Why do you want an explicit NOFS then?
Exactly my point - this is clumsy and a total mess. I'm not asking
for an explicit GFP_NOFS, just pointing out that the documented
restrictions that "vmalloc can only do GFP_KERNEL allocations" is
completely wrong.
vmalloc()
{
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_save();
p = __vmalloc(gfp_flags | GFP_KERNEL)
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_restore();
}
Yup, that's how simple it is to support GFP_NOFS support in
vmalloc().
Yes, this would work from the functionality POV but it defeats the
philosophy behind the scope API. Why would you even need this if the
scope was defined by the caller of the allocator? The initial hope was
to get rid of the NOFS abuse that can be seen in many filesystems. All
allocations from the scope would simply inherit the NOFS semantic so
an explicit NOFS shouldn't be really necessary, right?
This goes along with the argument that "it's impossible to do
GFP_NOFAIL with vmalloc" as I addressed above. These things are not
impossible, but we hide behind "we don't want people to use vmalloc"
as an excuse for having shitty behaviour whilst ignoring that
vmalloc is *heavily used* by core subsystems like filesystems
because they cannot rely on high order allocations succeeding....
I do not think there is any reason to discourage anybody from using
vmalloc these days. 32b is dying out and vmalloc space is no longer a
very scarce resource.
It also points out that the scope API is highly deficient.
We can do GFP_NOFS via the scope API, but we can't
do anything else because *there is no scope API for other GFP
flags*.
Why don't we have a GFP_NOFAIL/__GFP_RETRY_FOREVER scope API?
NO{FS,IO} where first flags to start this approach. And I have to admit
the experiment was much less successful then I hoped for. There are
still thousands of direct NOFS users so for some reason defining scopes
is not an easy thing to do.
I am not against NOFAIL scopes in principle but seeing the nofs
"success" I am worried this will not go really well either and it is
much more tricky as NOFAIL has much stronger requirements than NOFS.
Just imagine how tricky this can be if you just call a library code
that is not under your control within a NOFAIL scope. What if that
library code decides to allocate (e.g. printk that would attempt to do
an optimistic NOWAIT allocation).
That
would save us a lot of bother in XFS. What about GFP_DIRECT_RECLAIM?
I'd really like to turn that off for allocations in the XFS
transaction commit path (as noted already in this thread) because
direct reclaim that can make no progress is actively harmful (as
noted already in this thread)
As always if you have reasonable usecases then it is best to bring them
up on the MM list and we can discuss them.
Like I said - this is more than just bad documentation - the problem
is that the whole allocation API is an inconsistent mess of control
mechanisms to begin with...
I am not going to disagree. There is a lot of historical baggage and
it doesn't help that any change is really hard to review because this
interface is used throughout the kernel. I have tried to change some
most obvious inconsistencies and I can tell this has always been a
frustrating experience with a very small "reward" in the end because
there are so many other problems.
That being said, I would more than love to have a consistent and well
defined interface and if you want to spend a lot of time on that then be
my guest.
quoted
quoted
It would seem to make sense for kvmalloc to WARN_ON if it is passed
flags that does not allow it to use vmalloc.
vmalloc is certainly not the hottest path in the kernel so I wouldn't be
opposed.
kvmalloc is most certainly becoming one of the hottest paths in XFS.
IOWs, arguments that "vmalloc is not a hot path" are simply invalid
these days because they are simply untrue. e.g. the profiles I
posted in this thread...
Is it such a hot path that a check for compatible flags would be visible
in profiles though?
--
Michal Hocko
SUSE Labs
From: Dave Chinner <david@fromorbit.com> Date: 2021-10-13 02:32:41
On Mon, Oct 11, 2021 at 01:57:36PM +0200, Michal Hocko wrote:
On Sat 09-10-21 09:36:49, Dave Chinner wrote:
quoted
On Fri, Oct 08, 2021 at 09:48:39AM +0200, Michal Hocko wrote:
quoted
quoted
quoted
quoted
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
GFP_NOFS (without using the scope API) has the same problem as NOFAIL in
the vmalloc. Hence it is not supported. If you use the scope API then
you can GFP_KERNEL for kvmalloc. This is clumsy but I am not sure how to
define these conditions in a more sensible way. Special case NOFS if the
scope api is in use? Why do you want an explicit NOFS then?
Exactly my point - this is clumsy and a total mess. I'm not asking
for an explicit GFP_NOFS, just pointing out that the documented
restrictions that "vmalloc can only do GFP_KERNEL allocations" is
completely wrong.
vmalloc()
{
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_save();
p = __vmalloc(gfp_flags | GFP_KERNEL)
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_restore();
}
Yup, that's how simple it is to support GFP_NOFS support in
vmalloc().
Yes, this would work from the functionality POV but it defeats the
philosophy behind the scope API. Why would you even need this if the
scope was defined by the caller of the allocator?
Who actually cares that vmalloc might be using the scoped API
internally to implement GFP_NOFS or GFP_NOIO? Nobody at all.
It is far more useful (and self documenting!) for one-off allocations
to pass a GFP_NOFS flag than it is to use a scope API...
The initial hope was
to get rid of the NOFS abuse that can be seen in many filesystems. All
allocations from the scope would simply inherit the NOFS semantic so
an explicit NOFS shouldn't be really necessary, right?
Yes, but I think you miss my point entirely: that the vmalloc
restrictions on what gfp flags can be passed without making it
entirely useless are completely arbitrary and non-sensical.
quoted
This goes along with the argument that "it's impossible to do
GFP_NOFAIL with vmalloc" as I addressed above. These things are not
impossible, but we hide behind "we don't want people to use vmalloc"
as an excuse for having shitty behaviour whilst ignoring that
vmalloc is *heavily used* by core subsystems like filesystems
because they cannot rely on high order allocations succeeding....
I do not think there is any reason to discourage anybody from using
vmalloc these days. 32b is dying out and vmalloc space is no longer a
very scarce resource.
We are still discouraged from doing high order allocations and
should only use pages directly. Not to mention that the API doesn't
make it simple to use vmalloc as a direct replacement for high order
kmalloc tends to discourage new users...
quoted
It also points out that the scope API is highly deficient.
We can do GFP_NOFS via the scope API, but we can't
do anything else because *there is no scope API for other GFP
flags*.
Why don't we have a GFP_NOFAIL/__GFP_RETRY_FOREVER scope API?
NO{FS,IO} where first flags to start this approach. And I have to admit
the experiment was much less successful then I hoped for. There are
still thousands of direct NOFS users so for some reason defining scopes
is not an easy thing to do.
I am not against NOFAIL scopes in principle but seeing the nofs
"success" I am worried this will not go really well either and it is
much more tricky as NOFAIL has much stronger requirements than NOFS.
Just imagine how tricky this can be if you just call a library code
that is not under your control within a NOFAIL scope. What if that
library code decides to allocate (e.g. printk that would attempt to do
an optimistic NOWAIT allocation).
I already asked you that _exact_ question earlier in the thread
w.r.t. kvmalloc(GFP_NOFAIL) using optimistic NOWAIT kmalloc
allocation. I asked you as a MM expert to define *and document* the
behaviour that should result, not turn around and use the fact that
it is undefined behaviour as a "this is too hard" excuse for not
changing anything.
THe fact is that the scope APIs are only really useful for certain
contexts where restrictions are set by higher level functionality.
For one-off allocation constraints the API sucks and we end up with
crap like this (found in btrfs):
/*
* We're holding a transaction handle, so use a NOFS memory
* allocation context to avoid deadlock if reclaim happens.
*/
nofs_flag = memalloc_nofs_save();
value = kmalloc(size, GFP_KERNEL);
memalloc_nofs_restore(nofs_flag);
But also from btrfs, this pattern is repeated in several places:
nofs_flag = memalloc_nofs_save();
ctx = kvmalloc(struct_size(ctx, chunks, num_chunks), GFP_KERNEL);
memalloc_nofs_restore(nofs_flag);
This needs to use the scoped API because vmalloc doesn't support
GFP_NOFS. So the poor "vmalloc needs scoped API" pattern is bleeding
over into other code that doesn't have the problems vmalloc does. Do
you see how this leads to poorly written code now?
Or perhaps I should just point at ceph?
/*
* kvmalloc() doesn't fall back to the vmalloc allocator unless flags are
* compatible with (a superset of) GFP_KERNEL. This is because while the
* actual pages are allocated with the specified flags, the page table pages
* are always allocated with GFP_KERNEL.
*
* ceph_kvmalloc() may be called with GFP_KERNEL, GFP_NOFS or GFP_NOIO.
*/
void *ceph_kvmalloc(size_t size, gfp_t flags)
{
void *p;
if ((flags & (__GFP_IO | __GFP_FS)) == (__GFP_IO | __GFP_FS)) {
p = kvmalloc(size, flags);
} else if ((flags & (__GFP_IO | __GFP_FS)) == __GFP_IO) {
unsigned int nofs_flag = memalloc_nofs_save();
p = kvmalloc(size, GFP_KERNEL);
memalloc_nofs_restore(nofs_flag);
} else {
unsigned int noio_flag = memalloc_noio_save();
p = kvmalloc(size, GFP_KERNEL);
memalloc_noio_restore(noio_flag);
}
return p;
}
IOWs, a large number of the users of the scope API simply make
[k]vmalloc() provide GFP_NOFS behaviour. ceph_kvmalloc() is pretty
much a wrapper that indicates how all vmalloc functions should
behave. Honour GFP_NOFS and GFP_NOIO by using the scope API
internally.
quoted
That
would save us a lot of bother in XFS. What about GFP_DIRECT_RECLAIM?
I'd really like to turn that off for allocations in the XFS
transaction commit path (as noted already in this thread) because
direct reclaim that can make no progress is actively harmful (as
noted already in this thread)
As always if you have reasonable usecases then it is best to bring them
up on the MM list and we can discuss them.
They've been pointed out many times in the past, and I've pointed
them out again in this thread. Telling me to "bring them up on the
mm list" when that's exactly what I'm doing right now is not a
helpful response.
quoted
Like I said - this is more than just bad documentation - the problem
is that the whole allocation API is an inconsistent mess of control
mechanisms to begin with...
I am not going to disagree. There is a lot of historical baggage and
it doesn't help that any change is really hard to review because this
interface is used throughout the kernel. I have tried to change some
most obvious inconsistencies and I can tell this has always been a
frustrating experience with a very small "reward" in the end because
there are so many other problems.
Technical debt in the mm APIs is something the mm developers need to
address, not the people who tell you it's a problem for them.
Telling the messenger "do my job for me because I find it too
frustrating to make progress myself" doesn't help anyone make
progress. If you find it frustrating trying to get mm code changed,
imagine what it feels like for someone on the outside asking for
relatively basic things like a consistent control API....
That being said, I would more than love to have a consistent and well
defined interface and if you want to spend a lot of time on that then be
my guest.
My point exactly: saying "fix it yourself" is not a good response....
quoted
quoted
quoted
It would seem to make sense for kvmalloc to WARN_ON if it is passed
flags that does not allow it to use vmalloc.
vmalloc is certainly not the hottest path in the kernel so I wouldn't be
opposed.
kvmalloc is most certainly becoming one of the hottest paths in XFS.
IOWs, arguments that "vmalloc is not a hot path" are simply invalid
these days because they are simply untrue. e.g. the profiles I
posted in this thread...
Is it such a hot path that a check for compatible flags would be visible
in profiles though?
No, that doesn't even show up as noise - the overhead of global
spinlock contention and direct reclaim are the elephants that
profiles point to, not a couple of flag checks on function
parameters...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
From: Michal Hocko <mhocko@suse.com> Date: 2021-10-13 08:27:05
On Wed 13-10-21 13:32:31, Dave Chinner wrote:
On Mon, Oct 11, 2021 at 01:57:36PM +0200, Michal Hocko wrote:
quoted
On Sat 09-10-21 09:36:49, Dave Chinner wrote:
quoted
On Fri, Oct 08, 2021 at 09:48:39AM +0200, Michal Hocko wrote:
quoted
quoted
quoted
quoted
Even the API constaints of kvmalloc() w.r.t. only doing the vmalloc
fallback if the gfp context is GFP_KERNEL - we already do GFP_NOFS
kvmalloc via memalloc_nofs_save/restore(), so this behavioural
restriction w.r.t. gfp flags just makes no sense at all.
GFP_NOFS (without using the scope API) has the same problem as NOFAIL in
the vmalloc. Hence it is not supported. If you use the scope API then
you can GFP_KERNEL for kvmalloc. This is clumsy but I am not sure how to
define these conditions in a more sensible way. Special case NOFS if the
scope api is in use? Why do you want an explicit NOFS then?
Exactly my point - this is clumsy and a total mess. I'm not asking
for an explicit GFP_NOFS, just pointing out that the documented
restrictions that "vmalloc can only do GFP_KERNEL allocations" is
completely wrong.
vmalloc()
{
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_save();
p = __vmalloc(gfp_flags | GFP_KERNEL)
if (!(gfp_flags & __GFP_FS))
memalloc_nofs_restore();
}
Yup, that's how simple it is to support GFP_NOFS support in
vmalloc().
Yes, this would work from the functionality POV but it defeats the
philosophy behind the scope API. Why would you even need this if the
scope was defined by the caller of the allocator?
Who actually cares that vmalloc might be using the scoped API
internally to implement GFP_NOFS or GFP_NOIO? Nobody at all.
It is far more useful (and self documenting!) for one-off allocations
to pass a GFP_NOFS flag than it is to use a scope API...
I would agree with you if the explicit GFP_NOFS usage was consistent
and actually justified in the majority cases. My experience tells me
otherwise though. Many filesystems use the flag just because that is
easier. That leads to a huge overuse of the flag that leads to practical
problems.
I was hoping that if we offer an API that would define problematic
reclaim recursion scopes then it would reduce the abuse. I haven't
expected this to happen overnight but it is few years and it seems
it will not happen soon either.
[...]
quoted
quoted
It also points out that the scope API is highly deficient.
We can do GFP_NOFS via the scope API, but we can't
do anything else because *there is no scope API for other GFP
flags*.
Why don't we have a GFP_NOFAIL/__GFP_RETRY_FOREVER scope API?
NO{FS,IO} where first flags to start this approach. And I have to admit
the experiment was much less successful then I hoped for. There are
still thousands of direct NOFS users so for some reason defining scopes
is not an easy thing to do.
I am not against NOFAIL scopes in principle but seeing the nofs
"success" I am worried this will not go really well either and it is
much more tricky as NOFAIL has much stronger requirements than NOFS.
Just imagine how tricky this can be if you just call a library code
that is not under your control within a NOFAIL scope. What if that
library code decides to allocate (e.g. printk that would attempt to do
an optimistic NOWAIT allocation).
I already asked you that _exact_ question earlier in the thread
w.r.t. kvmalloc(GFP_NOFAIL) using optimistic NOWAIT kmalloc
allocation. I asked you as a MM expert to define *and document* the
behaviour that should result, not turn around and use the fact that
it is undefined behaviour as a "this is too hard" excuse for not
changing anything.
Dave, you have "thrown" a lot of complains in previous emails and it is
hard to tell rants from features requests apart. I am sorry but I
believe it would be much more productive to continue this discussion if
you could mild your tone.
Can I ask you to break down your feature requests into separate emails
so that we can discuss and track them separately rather in this quite a
long thread which has IMHO diverghed from the initial topic. Thanks!
THe fact is that the scope APIs are only really useful for certain
contexts where restrictions are set by higher level functionality.
For one-off allocation constraints the API sucks and we end up with
Could you be more specific about these one-off allocation constrains?
What would be the reason to define one-off NO{FS,IO} allocation
constrain? Or did you have your NOFAIL example in mind?
crap like this (found in btrfs):
/*
* We're holding a transaction handle, so use a NOFS memory
* allocation context to avoid deadlock if reclaim happens.
*/
nofs_flag = memalloc_nofs_save();
value = kmalloc(size, GFP_KERNEL);
memalloc_nofs_restore(nofs_flag);
Yes this looks wrong indeed! If I were to review such a code I would ask
why the scope cannot match the transaction handle context. IIRC jbd does
that.
I am aware of these patterns. I was pulled in some discussions in the
past and in some it turned out that the constrain is not needed at all
and in some cases that has led to a proper scope definition. As you
point out in your other examples it just happens that it is easier to
go an easy path and define scopes ad-hoc to work around allocation
API limitations.
[...]
IOWs, a large number of the users of the scope API simply make
[k]vmalloc() provide GFP_NOFS behaviour. ceph_kvmalloc() is pretty
much a wrapper that indicates how all vmalloc functions should
behave. Honour GFP_NOFS and GFP_NOIO by using the scope API
internally.
I was discouraging from this behavior at vmalloc level to push people
to use scopes properly - aka at the level where the reclaim recursion is
really a problem. If that is infeasible in practice then we can
re-evaluate of course. I was really hoping we can get rid of cargo cult
GFP_NOFS usage this way but the reality often disagrees with hopes.
All that being said, let's discuss [k]vmalloc constrains and usecases
that need changes in a separate email thread.
Thanks!
--
Michal Hocko
SUSE Labs
From: David Sterba <hidden> Date: 2021-10-14 11:32:31
On Wed, Oct 13, 2021 at 10:26:58AM +0200, Michal Hocko wrote:
quoted
crap like this (found in btrfs):
/*
* We're holding a transaction handle, so use a NOFS memory
* allocation context to avoid deadlock if reclaim happens.
*/
nofs_flag = memalloc_nofs_save();
value = kmalloc(size, GFP_KERNEL);
memalloc_nofs_restore(nofs_flag);
Yes this looks wrong indeed! If I were to review such a code I would ask
why the scope cannot match the transaction handle context. IIRC jbd does
that.
Adding the transaction start/end as the NOFS scope is a long term plan
and going on for years, because it's not a change we would need in
btrfs, but rather a favor to MM to switch away from "GFP_NOFS everywhere
because it's easy".
The first step was to convert the easy cases. Almost all safe cases
switching GFP_NOFS to GFP_KERNEL have happened. Another step is to
convert GFP_NOFS to memalloc_nofs_save/GFP_KERNEL/memalloc_nofs_restore
in contexts where we know we'd rely on the transaction NOFS scope in the
future. Once this is implemented, the memalloc_nofs_* calls are deleted
and it works as expected. Now you may argue that the switch could be
changing GFP_NOFS to GFP_KERNEL at that time but that is not that easy
to review or reason about in the whole transaction context in all
allocations.
This leads to code that was found in __btrfs_set_acl and called crap
or wrong, because perhaps the background and the bigger plan is not
immediately obvious. I hope the explanation above it puts it to the
right perspective.
The other class of scoped NOFS protection is around vmalloc-based
allocations but that's for a different reason, would be solved by the
same transaction start/end conversion as well.
I'm working on that from time to time but this usually gets pushed down
in the todo list. It's changing a lot of code, from what I've researched
so far cannot be done at once and would probably introduce bugs hard to
hit because of the external conditions (allocator, system load, ...).
I have a plan to do that incrementally, adding assertions and converting
functions in small batches to be able to catch bugs early, but I'm not
exactly thrilled to start such endeavour in addition to normal
development bug hunting.
To get things moving again, I've refreshed the patch adding stubs and
will try to find the best timing for merg to avoid patch conflicts, but
no promises.
From: Michal Hocko <mhocko@suse.com> Date: 2021-10-14 11:46:53
On Thu 14-10-21 13:32:01, David Sterba wrote:
On Wed, Oct 13, 2021 at 10:26:58AM +0200, Michal Hocko wrote:
quoted
quoted
crap like this (found in btrfs):
/*
* We're holding a transaction handle, so use a NOFS memory
* allocation context to avoid deadlock if reclaim happens.
*/
nofs_flag = memalloc_nofs_save();
value = kmalloc(size, GFP_KERNEL);
memalloc_nofs_restore(nofs_flag);
Yes this looks wrong indeed! If I were to review such a code I would ask
why the scope cannot match the transaction handle context. IIRC jbd does
that.
Adding the transaction start/end as the NOFS scope is a long term plan
and going on for years, because it's not a change we would need in
btrfs, but rather a favor to MM to switch away from "GFP_NOFS everywhere
because it's easy".
The first step was to convert the easy cases. Almost all safe cases
switching GFP_NOFS to GFP_KERNEL have happened. Another step is to
convert GFP_NOFS to memalloc_nofs_save/GFP_KERNEL/memalloc_nofs_restore
in contexts where we know we'd rely on the transaction NOFS scope in the
future. Once this is implemented, the memalloc_nofs_* calls are deleted
and it works as expected. Now you may argue that the switch could be
changing GFP_NOFS to GFP_KERNEL at that time but that is not that easy
to review or reason about in the whole transaction context in all
allocations.
This leads to code that was found in __btrfs_set_acl and called crap
or wrong, because perhaps the background and the bigger plan is not
immediately obvious. I hope the explanation above it puts it to the
right perspective.
Yes it helps. Thanks for the clarification because this is far from
obvious and changelogs I've checked do not mention this high level plan.
I would have gone with a /* TODO: remove me once transactions use scopes... */
but this is obviously your call.
The other class of scoped NOFS protection is around vmalloc-based
allocations but that's for a different reason, would be solved by the
same transaction start/end conversion as well.
I'm working on that from time to time but this usually gets pushed down
in the todo list. It's changing a lot of code, from what I've researched
so far cannot be done at once and would probably introduce bugs hard to
hit because of the external conditions (allocator, system load, ...).
I have a plan to do that incrementally, adding assertions and converting
functions in small batches to be able to catch bugs early, but I'm not
exactly thrilled to start such endeavour in addition to normal
development bug hunting.
To get things moving again, I've refreshed the patch adding stubs and
will try to find the best timing for merg to avoid patch conflicts, but
no promises.
Put simply, we want "retry forever" semantics to match what
production kernels have been doing for the past couple of decades,
but all we've been given are "never fail" semantics that also do
something different and potentially much more problematic.
Do you see the difference here? __GFP_NOFAIL is not what we
need in the vast majority of cases where it is used. We don't want
the failing allocations to drive the machine hard into critical
reserves, we just want the allocation to -eventually succeed- and if
it doesn't, that's our problem to handle, not kmalloc()....
I can see your point. I do have a recollection that there were some
instance involved where an emergency access to memory reserves helped
in OOM situations.
It might have been better to annotate those particular calls with
__GFP_ATOMIC or similar rather then change GFP_NOFAIL for everyone.
Too late to fix that now though I think. Maybe the best way forward is
to discourage new uses of GFP_NOFAIL. We would need a well-documented
replacement.
Anway as I've tried to explain earlier that this all is an
implementation detail users of the flag shouldn't really care about. If
this heuristic is not doing any good then it should be removed.
Maybe users shouldn't care about implementation details, but they do
need to care about semantics and costs.
We need to know when it is appropriate to use GFP_NOFAIL, and when it is
not. And what alternatives there are when it is not appropriate.
Just saying "try to avoid using it" and "requires careful analysis"
isn't acceptable. Sometimes it is unavoidable and analysis can only be
done with a clear understanding of costs. Possibly analysis can only be
done with a clear understanding of the internal implementation details.
quoted
It also points out that the scope API is highly deficient.
We can do GFP_NOFS via the scope API, but we can't
do anything else because *there is no scope API for other GFP
flags*.
Why don't we have a GFP_NOFAIL/__GFP_RETRY_FOREVER scope API?
NO{FS,IO} where first flags to start this approach. And I have to admit
the experiment was much less successful then I hoped for. There are
still thousands of direct NOFS users so for some reason defining scopes
is not an easy thing to do.
I'm not certain your conclusion is valid. It could be that defining
scopes is easy enough, but no one feels motivated to do it.
We need to do more than provide functionality. We need to tell people.
Repeatedly. And advertise widely. And propose patches to make use of
the functionality. And... and... and...
I think changing to the scope API is a good change, but it is
conceptually a big change. It needs to be driven.
I am not against NOFAIL scopes in principle but seeing the nofs
"success" I am worried this will not go really well either and it is
much more tricky as NOFAIL has much stronger requirements than NOFS.
Just imagine how tricky this can be if you just call a library code
that is not under your control within a NOFAIL scope. What if that
library code decides to allocate (e.g. printk that would attempt to do
an optimistic NOWAIT allocation).
__GFP_NOMEMALLOC holds a lesson worth learning here. PF_MEMALLOC
effectively adds __GFP_MEMALLOC to all allocations, but some call sites
need to over-ride that because there are alternate strategies available.
This need-to-over-ride doesn't apply to NOFS or NOIO as that really is a
thread-wide state. But MEMALLOC and NOFAIL are different. Some call
sites can reasonably handle failure locally.
I imagine the scope-api would say something like "NO_ENOMEM". i.e.
memory allocations can fail as long as ENOMEM is never returned.
Any caller that sets __GFP_RETRY_MAYFAIL or __GFP_NORETRY or maybe some
others which not be affected by the NO_ENOMEM scope. But a plain
GFP_KERNEL would.
Introducing the scope api would be a good opportunity to drop the
priority boost and *just* block until success. Priority boosts could
then be added (possibly as a scope) only where they are measurably needed.
I think we have 28 process flags in use. So we can probably afford one
more for PF_MEMALLOC_NO_ENOMEM. What other scope flags might be useful?
PF_MEMALLOC_BOOST which added __GFP_ATOMIC but not __GFP_MEMALLOC ??
PF_MEMALLOC_NORECLAIM ??
quoted
That
would save us a lot of bother in XFS. What about GFP_DIRECT_RECLAIM?
I'd really like to turn that off for allocations in the XFS
transaction commit path (as noted already in this thread) because
direct reclaim that can make no progress is actively harmful (as
noted already in this thread)
As always if you have reasonable usecases then it is best to bring them
up on the MM list and we can discuss them.
We are on the MM lists now... let's discuss :-)
Dave: How would you feel about an effort to change xfs to stop using
GFP_NOFS, and to use memalloc_nofs_save/restore instead? Having a major
filesystem make the transition would be a good test-case, and could be
used to motivate other filesystems to follow.
We could add and use memalloc_no_enomem_save() too.
Thanks,
NeilBrown
From: Michal Hocko <mhocko@suse.com> Date: 2021-10-18 10:23:51
On Tue 12-10-21 08:49:46, Neil Brown wrote:
On Mon, 11 Oct 2021, Michal Hocko wrote:
quoted
On Sat 09-10-21 09:36:49, Dave Chinner wrote:
quoted
Put simply, we want "retry forever" semantics to match what
production kernels have been doing for the past couple of decades,
but all we've been given are "never fail" semantics that also do
something different and potentially much more problematic.
Do you see the difference here? __GFP_NOFAIL is not what we
need in the vast majority of cases where it is used. We don't want
the failing allocations to drive the machine hard into critical
reserves, we just want the allocation to -eventually succeed- and if
it doesn't, that's our problem to handle, not kmalloc()....
I can see your point. I do have a recollection that there were some
instance involved where an emergency access to memory reserves helped
in OOM situations.
It might have been better to annotate those particular calls with
__GFP_ATOMIC or similar rather then change GFP_NOFAIL for everyone.
For historical reasons __GFP_ATOMIC is reserved for non sleeping
allocations. __GFP_HIGH would be an alternative.
Too late to fix that now though I think. Maybe the best way forward is
to discourage new uses of GFP_NOFAIL. We would need a well-documented
replacement.
I am not sure what that should be. Really if the memory reserves
behavior of GFP_NOFAIL is really problematic then let's just reap it
out. I do not see a new nofail like flag is due.
quoted
Anway as I've tried to explain earlier that this all is an
implementation detail users of the flag shouldn't really care about. If
this heuristic is not doing any good then it should be removed.
Maybe users shouldn't care about implementation details, but they do
need to care about semantics and costs.
We need to know when it is appropriate to use GFP_NOFAIL, and when it is
not. And what alternatives there are when it is not appropriate.
Just saying "try to avoid using it" and "requires careful analysis"
isn't acceptable. Sometimes it is unavoidable and analysis can only be
done with a clear understanding of costs. Possibly analysis can only be
done with a clear understanding of the internal implementation details.
What we document currently is this
* %__GFP_NOFAIL: The VM implementation _must_ retry infinitely: the caller
* cannot handle allocation failures. The allocation could block
* indefinitely but will never return with failure. Testing for
* failure is pointless.
* New users should be evaluated carefully (and the flag should be
* used only when there is no reasonable failure policy) but it is
* definitely preferable to use the flag rather than opencode endless
* loop around allocator.
* Using this flag for costly allocations is _highly_ discouraged.
so we tell when to use it - aka no reasonable failure policy. We put
some discouragind language there. There is some discouraging language
for high order allocations. Maybe we should suggest an alternative
there. It seems there are usecases for those as well so we should
implement a proper NOFAIL kvmalloc and recommend it for that instead.
quoted
quoted
It also points out that the scope API is highly deficient.
We can do GFP_NOFS via the scope API, but we can't
do anything else because *there is no scope API for other GFP
flags*.
Why don't we have a GFP_NOFAIL/__GFP_RETRY_FOREVER scope API?
NO{FS,IO} where first flags to start this approach. And I have to admit
the experiment was much less successful then I hoped for. There are
still thousands of direct NOFS users so for some reason defining scopes
is not an easy thing to do.
I'm not certain your conclusion is valid. It could be that defining
scopes is easy enough, but no one feels motivated to do it.
We need to do more than provide functionality. We need to tell people.
Repeatedly. And advertise widely. And propose patches to make use of
the functionality. And... and... and...
Been there, done that for the low hanging fruit. Others were much more
complex for me to follow up and I had other stuff on my table.
I think changing to the scope API is a good change, but it is
conceptually a big change. It needs to be driven.
Agreed.
quoted
I am not against NOFAIL scopes in principle but seeing the nofs
"success" I am worried this will not go really well either and it is
much more tricky as NOFAIL has much stronger requirements than NOFS.
Just imagine how tricky this can be if you just call a library code
that is not under your control within a NOFAIL scope. What if that
library code decides to allocate (e.g. printk that would attempt to do
an optimistic NOWAIT allocation).
__GFP_NOMEMALLOC holds a lesson worth learning here. PF_MEMALLOC
effectively adds __GFP_MEMALLOC to all allocations, but some call sites
need to over-ride that because there are alternate strategies available.
This need-to-over-ride doesn't apply to NOFS or NOIO as that really is a
thread-wide state. But MEMALLOC and NOFAIL are different. Some call
sites can reasonably handle failure locally.
I imagine the scope-api would say something like "NO_ENOMEM". i.e.
memory allocations can fail as long as ENOMEM is never returned.
Any caller that sets __GFP_RETRY_MAYFAIL or __GFP_NORETRY or maybe some
others which not be affected by the NO_ENOMEM scope. But a plain
GFP_KERNEL would.
Introducing the scope api would be a good opportunity to drop the
priority boost and *just* block until success. Priority boosts could
then be added (possibly as a scope) only where they are measurably needed.
I think we have 28 process flags in use. So we can probably afford one
more for PF_MEMALLOC_NO_ENOMEM. What other scope flags might be useful?
PF_MEMALLOC_BOOST which added __GFP_ATOMIC but not __GFP_MEMALLOC ??
PF_MEMALLOC_NORECLAIM ??
I dunno. PF_MEMALLOC and its GFP_$FOO counterparts are quite hard to
wrap my head around. I have never liked thos much TBH and building more
on top sounds like step backward. I might be wrong but this sounds like
even more work than NOFS scopes.
quoted
quoted
That
would save us a lot of bother in XFS. What about GFP_DIRECT_RECLAIM?
I'd really like to turn that off for allocations in the XFS
transaction commit path (as noted already in this thread) because
direct reclaim that can make no progress is actively harmful (as
noted already in this thread)
As always if you have reasonable usecases then it is best to bring them
up on the MM list and we can discuss them.
We are on the MM lists now... let's discuss :-)
Sure we can but this thread is a mix of so many topics that finding
something useful will turn to be hard from my past experience.
Dave: How would you feel about an effort to change xfs to stop using
GFP_NOFS, and to use memalloc_nofs_save/restore instead?
xfs is an example of a well behaved scope user. In fact the API has been
largely based on xfs previous interface. There are still NOFS usesages
in xfs which would be great to get rid of (e.g. the default mapping NOFS
which was added due to lockdep false positives but that is unrelated).
Having a major
filesystem make the transition would be a good test-case, and could be
used to motivate other filesystems to follow.
We could add and use memalloc_no_enomem_save() too.
ext has converted their transaction context to the scope API as well.
There is still some explicit NOFS usage but I haven't checked details
recently.
--
Michal Hocko
SUSE Labs
Put simply, we want "retry forever" semantics to match what
production kernels have been doing for the past couple of decades,
but all we've been given are "never fail" semantics that also do
something different and potentially much more problematic.
Do you see the difference here? __GFP_NOFAIL is not what we
need in the vast majority of cases where it is used. We don't want
the failing allocations to drive the machine hard into critical
reserves, we just want the allocation to -eventually succeed- and if
it doesn't, that's our problem to handle, not kmalloc()....
I can see your point. I do have a recollection that there were some
instance involved where an emergency access to memory reserves helped
in OOM situations.
It might have been better to annotate those particular calls with
__GFP_ATOMIC or similar rather then change GFP_NOFAIL for everyone.
For historical reasons __GFP_ATOMIC is reserved for non sleeping
allocations. __GFP_HIGH would be an alternative.
Historical reasons certainly shouldn't be ignored. But they can be
questioned.
__GFP_ATOMIC is documented as "the caller cannot reclaim or sleep and is
high priority".
This seems to over-lap with __GFP_DIRECT_RECLAIM (which permits reclaim
and is the only place where page_alloc sleeps ... I think).
The effect of setting __GFP_ATOMIC is:
- triggers WARN_ON if __GFP_DIRECT_RECLAIM is also set.
- bypass memcg limits
- ignore the watermark_boost_factor effect
- clears ALLOC_CPUSET
- sets ALLOC_HARDER which provides:
- access to nr_reserved_highatomic reserves
- access to 1/4 the low-watermark reserves (ALLOC_HIGH gives 1/2)
Combine them and you get access to 5/8 of the reserves.
It is also used by driver/iommu/tegra-smmu.c to decide if a spinlock
should remain held, or should be dropped over the alloc_page(). That's
.... not my favourite code.
So apart from the tegra thing and the WARN_ON, there is nothing about
__GFP_ATOMIC which suggests it should only be used for non-sleeping
allocations.
It *should* only be used for allocations with a high failure cost and
relatively short time before the memory will be returned and that likely
includes many non sleeping allocations. It isn't clear to me why an
allocation that is willing to sleep (if absolutely necessary) shouldn't
be able to benefit from the priority boost of __GFP_ATOMIC. Or at least
of ALLOC_HARDER...
Maybe __GFP_HIGH should get the memcg and watermark_boost benefits too?
Given that we have ALLOC_HARDER and ALLOC_HIGH, it would seem to be
sensible to export those two settings in GFP_foo, and not forbid one of
them to be used with __GFP_DIRECT_RECLAIM.
quoted
Too late to fix that now though I think. Maybe the best way forward is
to discourage new uses of GFP_NOFAIL. We would need a well-documented
replacement.
I am not sure what that should be. Really if the memory reserves
behavior of GFP_NOFAIL is really problematic then let's just reap it
out. I do not see a new nofail like flag is due.
Presumably there is a real risk of deadlock if we just remove the
memory-reserves boosts of __GFP_NOFAIL. Maybe it would be safe to
replace all current users of __GFP_NOFAIL with __GFP_NOFAIL|__GFP_HIGH,
and then remove the __GFP_HIGH where analysis suggests there is no risk
of deadlocks.
Or maybe rename the __GFP_NOFAIL flag and #define __GFP_NOFAIL to
include __GFP_HIGH?
This would certainly be a better result than adding a new flag.
quoted
quoted
Anway as I've tried to explain earlier that this all is an
implementation detail users of the flag shouldn't really care about. If
this heuristic is not doing any good then it should be removed.
Maybe users shouldn't care about implementation details, but they do
need to care about semantics and costs.
We need to know when it is appropriate to use GFP_NOFAIL, and when it is
not. And what alternatives there are when it is not appropriate.
Just saying "try to avoid using it" and "requires careful analysis"
isn't acceptable. Sometimes it is unavoidable and analysis can only be
done with a clear understanding of costs. Possibly analysis can only be
done with a clear understanding of the internal implementation details.
What we document currently is this
* %__GFP_NOFAIL: The VM implementation _must_ retry infinitely: the caller
* cannot handle allocation failures. The allocation could block
* indefinitely but will never return with failure. Testing for
* failure is pointless.
This implies it is incompatible with __GFP_NORETRY and (probably)
requires __GFP_RECLAIM. That is worth documenting, and possibly also a
WARN_ON.
* New users should be evaluated carefully (and the flag should be
* used only when there is no reasonable failure policy) but it is
* definitely preferable to use the flag rather than opencode endless
* loop around allocator.
How do we perform this evaluation? And why is it preferable to a loop?
There are times when a loop makes sense, if there might be some other
event that could provide the needed memory ... or if a SIGKILL might
make it irrelevant.
slab allocators presumably shouldn't pass __GFP_NOFAIL to alloc_page(),
but should instead loop around
1/ check if any existing slabs have space
2/ if not, try to allocate a new page
Providing the latter blocks for a while but not indefinitely that should
be optimal.
Why is __GFP_NOFAIL better?
* Using this flag for costly allocations is _highly_ discouraged.
This is unhelpful. Saying something is "discouraged" carries an implied
threat. This is open source and threats need to be open.
Why is it discouraged? IF it is not forbidden, then it is clearly
permitted. Maybe there are costs - so a clear statement of those costs
would be appropriate.
Also, what is a suitable alternative?
Current code will trigger a WARN_ON, so it is effectively forbidden.
Maybe we should document that __GFP_NOFAIL is forbidden for orders above
1, and that vmalloc() should be used instead (thanks for proposing that
patch!).
But that would mean __GFP_NOFAIL cannot be used for slabs which happen
to use large orders.
Hmmm. it appears that slub.c disables __GFP_NOFAIL when it tries for a
large order allocation, and slob.c never tries large order allocations.
So this only affects slab.c.
xfs makes heavy use of kmem_cache_zalloc with __GFP_NOFAIL. I wonder if
any of these slabs have large order with slab.c.
so we tell when to use it - aka no reasonable failure policy. We put
some discouragind language there. There is some discouraging language
for high order allocations. Maybe we should suggest an alternative
there. It seems there are usecases for those as well so we should
implement a proper NOFAIL kvmalloc and recommend it for that instead.
yes- suggest an alternative and also say what the tradeoffs are.
quoted
quoted
quoted
It also points out that the scope API is highly deficient.
We can do GFP_NOFS via the scope API, but we can't
do anything else because *there is no scope API for other GFP
flags*.
Why don't we have a GFP_NOFAIL/__GFP_RETRY_FOREVER scope API?
NO{FS,IO} where first flags to start this approach. And I have to admit
the experiment was much less successful then I hoped for. There are
still thousands of direct NOFS users so for some reason defining scopes
is not an easy thing to do.
I'm not certain your conclusion is valid. It could be that defining
scopes is easy enough, but no one feels motivated to do it.
We need to do more than provide functionality. We need to tell people.
Repeatedly. And advertise widely. And propose patches to make use of
the functionality. And... and... and...
Been there, done that for the low hanging fruit. Others were much more
complex for me to follow up and I had other stuff on my table.
I have no doubt that is a slow and rather thankless task, with no real
payoff until it is complete. It reminds me a bit of BKL removal and
64-bit time. I think it is worth doing though. Finding the balance
between letting it consume you and just giving up would be a challenge.
quoted
I think changing to the scope API is a good change, but it is
conceptually a big change. It needs to be driven.
Agreed.
quoted
quoted
I am not against NOFAIL scopes in principle but seeing the nofs
"success" I am worried this will not go really well either and it is
much more tricky as NOFAIL has much stronger requirements than NOFS.
Just imagine how tricky this can be if you just call a library code
that is not under your control within a NOFAIL scope. What if that
library code decides to allocate (e.g. printk that would attempt to do
an optimistic NOWAIT allocation).
__GFP_NOMEMALLOC holds a lesson worth learning here. PF_MEMALLOC
effectively adds __GFP_MEMALLOC to all allocations, but some call sites
need to over-ride that because there are alternate strategies available.
This need-to-over-ride doesn't apply to NOFS or NOIO as that really is a
thread-wide state. But MEMALLOC and NOFAIL are different. Some call
sites can reasonably handle failure locally.
I imagine the scope-api would say something like "NO_ENOMEM". i.e.
memory allocations can fail as long as ENOMEM is never returned.
Any caller that sets __GFP_RETRY_MAYFAIL or __GFP_NORETRY or maybe some
others which not be affected by the NO_ENOMEM scope. But a plain
GFP_KERNEL would.
Introducing the scope api would be a good opportunity to drop the
priority boost and *just* block until success. Priority boosts could
then be added (possibly as a scope) only where they are measurably needed.
I think we have 28 process flags in use. So we can probably afford one
more for PF_MEMALLOC_NO_ENOMEM. What other scope flags might be useful?
PF_MEMALLOC_BOOST which added __GFP_ATOMIC but not __GFP_MEMALLOC ??
PF_MEMALLOC_NORECLAIM ??
I dunno. PF_MEMALLOC and its GFP_$FOO counterparts are quite hard to
wrap my head around. I have never liked thos much TBH and building more
on top sounds like step backward. I might be wrong but this sounds like
even more work than NOFS scopes.
quoted
quoted
quoted
That
would save us a lot of bother in XFS. What about GFP_DIRECT_RECLAIM?
I'd really like to turn that off for allocations in the XFS
transaction commit path (as noted already in this thread) because
direct reclaim that can make no progress is actively harmful (as
noted already in this thread)
As always if you have reasonable usecases then it is best to bring them
up on the MM list and we can discuss them.
We are on the MM lists now... let's discuss :-)
Sure we can but this thread is a mix of so many topics that finding
something useful will turn to be hard from my past experience.
Unfortunately life is messy. I just wanted to remove all
congestion_wait() calls. But that lead to __GFP_NOFAIL and to scopes
allocation API, and there are still more twisty passages waiting.
Sometimes you don't know what topic will usefully start a constructive
thread until you've already figured out the answer :-(
quoted
Dave: How would you feel about an effort to change xfs to stop using
GFP_NOFS, and to use memalloc_nofs_save/restore instead?
xfs is an example of a well behaved scope user. In fact the API has been
largely based on xfs previous interface. There are still NOFS usesages
in xfs which would be great to get rid of (e.g. the default mapping NOFS
which was added due to lockdep false positives but that is unrelated).
quoted
Having a major
filesystem make the transition would be a good test-case, and could be
used to motivate other filesystems to follow.
We could add and use memalloc_no_enomem_save() too.
ext has converted their transaction context to the scope API as well.
There is still some explicit NOFS usage but I haven't checked details
recently.
Of the directories in fs/,
42 contain no mention of GFP_NOFS
17 contain fewer than 10
The 10 with most frequent usage (including comments) are:
47 fs/afs/
48 fs/f2fs/
49 fs/nfs/
54 fs/dlm/
59 fs/ceph/
66 fs/ext4/
73 fs/ntfs3/
73 fs/ocfs2/
83 fs/ubifs/
231 fs/btrfs/
xfs is 28 - came in number 12. Though there are 25 KM_NOFS
allocations, which would push it up to 7th place.
A few use GFP_NOIO - nfs(11) and f2fs(9) being the biggest users.
So clearly there is work to do
Maybe we could add something to checkpatch.pl to discourage the addition
of new GFP_NOFS usage.
There is a lot of stuff there.... the bits that are important to me are:
- why is __GFP_NOFAIL preferred? It is a valuable convenience, but I
don't see that it is necessary
- is it reasonable to use __GFP_HIGH when looping if there is a risk of
deadlock?
- Will __GFP_DIRECT_RECLAIM always result in a delay before failure? In
that case it should be safe to loop around allocations using
__GFP_DIRECT_RECLAIM without needing congestion_wait() (so it can
just be removed.
Thanks,
NeilBrown
From: Michal Hocko <mhocko@suse.com> Date: 2021-10-19 13:59:31
On Tue 19-10-21 15:32:27, Neil Brown wrote:
On Mon, 18 Oct 2021, Michal Hocko wrote:
quoted
On Tue 12-10-21 08:49:46, Neil Brown wrote:
quoted
On Mon, 11 Oct 2021, Michal Hocko wrote:
quoted
On Sat 09-10-21 09:36:49, Dave Chinner wrote:
quoted
Put simply, we want "retry forever" semantics to match what
production kernels have been doing for the past couple of decades,
but all we've been given are "never fail" semantics that also do
something different and potentially much more problematic.
Do you see the difference here? __GFP_NOFAIL is not what we
need in the vast majority of cases where it is used. We don't want
the failing allocations to drive the machine hard into critical
reserves, we just want the allocation to -eventually succeed- and if
it doesn't, that's our problem to handle, not kmalloc()....
I can see your point. I do have a recollection that there were some
instance involved where an emergency access to memory reserves helped
in OOM situations.
It might have been better to annotate those particular calls with
__GFP_ATOMIC or similar rather then change GFP_NOFAIL for everyone.
For historical reasons __GFP_ATOMIC is reserved for non sleeping
allocations. __GFP_HIGH would be an alternative.
Historical reasons certainly shouldn't be ignored. But they can be
questioned.
Agreed. Changing them is a more challenging task though. For example I
really dislike how access to memory reserves is bound to "no reclaim"
requirement. Ideally those should be completely orthogonal.
I also do not think we need as many ways to ask for memory reserves as
we have. Can a "regular" kernel developer tell a difference between
__GFP_ATOMIC and __GFP_HIGH?
I do not think so, unless one is willing to do ...
__GFP_ATOMIC is documented as "the caller cannot reclaim or sleep and is
high priority".
This seems to over-lap with __GFP_DIRECT_RECLAIM (which permits reclaim
and is the only place where page_alloc sleeps ... I think).
The effect of setting __GFP_ATOMIC is:
- triggers WARN_ON if __GFP_DIRECT_RECLAIM is also set.
- bypass memcg limits
- ignore the watermark_boost_factor effect
- clears ALLOC_CPUSET
- sets ALLOC_HARDER which provides:
- access to nr_reserved_highatomic reserves
- access to 1/4 the low-watermark reserves (ALLOC_HIGH gives 1/2)
Combine them and you get access to 5/8 of the reserves.
... exactly this. And these are bunch of hacks developed over time and
the baggage which is hard to change as I've said. Somebody with a
sufficient time budget should start questioning all those and eventually
make __GFP_ATOMIC a story of the past.
It is also used by driver/iommu/tegra-smmu.c to decide if a spinlock
should remain held, or should be dropped over the alloc_page(). That's
.... not my favourite code.
Exactly!
So apart from the tegra thing and the WARN_ON, there is nothing about
__GFP_ATOMIC which suggests it should only be used for non-sleeping
allocations.
The warning was added when the original GFP_ATOMIC was untangled from the
reclaim implications to keep the "backward compatibility" IIRC. Mostly
for IRQ handlers where the GFP_ATOMIC was used the most. My memory might
fail me though.
It *should* only be used for allocations with a high failure cost and
relatively short time before the memory will be returned and that likely
includes many non sleeping allocations. It isn't clear to me why an
allocation that is willing to sleep (if absolutely necessary) shouldn't
be able to benefit from the priority boost of __GFP_ATOMIC. Or at least
of ALLOC_HARDER...
I completely agree! As mentioned above memory reserves should be
completely orthogonal. I am not sure we want an API for many different
levels of reserves access. Do we need more than __GFP_HIGH? Maybe with a
more descriptive name.
Maybe __GFP_HIGH should get the memcg and watermark_boost benefits too?
Given that we have ALLOC_HARDER and ALLOC_HIGH, it would seem to be
sensible to export those two settings in GFP_foo, and not forbid one of
them to be used with __GFP_DIRECT_RECLAIM.
I think ALLOC_HARDER should be kept internal implementation detail when
the allocator needs to give somebody a boost on top of requests for
internal balancing between requests.
ALLOC_HIGH already matches __GFP_HIGH and that should be the way to
ask for a boost explicitly IMO. We also have ALLOC_OOM as another level of
internal memory reserves for OOM victims. Again something to be in hands
of the allocator.
quoted
quoted
Too late to fix that now though I think. Maybe the best way forward is
to discourage new uses of GFP_NOFAIL. We would need a well-documented
replacement.
I am not sure what that should be. Really if the memory reserves
behavior of GFP_NOFAIL is really problematic then let's just reap it
out. I do not see a new nofail like flag is due.
Presumably there is a real risk of deadlock if we just remove the
memory-reserves boosts of __GFP_NOFAIL. Maybe it would be safe to
replace all current users of __GFP_NOFAIL with __GFP_NOFAIL|__GFP_HIGH,
and then remove the __GFP_HIGH where analysis suggests there is no risk
of deadlocks.
I would much rather not bind those together and go other way around. If
somebody can actually hit deadlocks (those are quite easy to spot as
they do not go away) then we can talk about how to deal with them.
Memory reserves can help only > < this much.
Or maybe rename the __GFP_NOFAIL flag and #define __GFP_NOFAIL to
include __GFP_HIGH?
Wouldn't that lead to the __GFP_ATOMIC story again?
This would certainly be a better result than adding a new flag.
quoted
quoted
quoted
Anway as I've tried to explain earlier that this all is an
implementation detail users of the flag shouldn't really care about. If
this heuristic is not doing any good then it should be removed.
Maybe users shouldn't care about implementation details, but they do
need to care about semantics and costs.
We need to know when it is appropriate to use GFP_NOFAIL, and when it is
not. And what alternatives there are when it is not appropriate.
Just saying "try to avoid using it" and "requires careful analysis"
isn't acceptable. Sometimes it is unavoidable and analysis can only be
done with a clear understanding of costs. Possibly analysis can only be
done with a clear understanding of the internal implementation details.
What we document currently is this
* %__GFP_NOFAIL: The VM implementation _must_ retry infinitely: the caller
* cannot handle allocation failures. The allocation could block
* indefinitely but will never return with failure. Testing for
* failure is pointless.
This implies it is incompatible with __GFP_NORETRY and (probably)
requires __GFP_RECLAIM. That is worth documenting, and possibly also a
WARN_ON.
Yes, I thought this would be obvious as those are reclaim modifiers so
they require a reclaim. But I do see a point that being explicit here
cannot hurt. Same with combining them together. It just doesn't make
much sense to retry for ever and requesting noretry or retry and fail
at the same time. Again a clarification cannot hurt though.
quoted
* New users should be evaluated carefully (and the flag should be
* used only when there is no reasonable failure policy) but it is
* definitely preferable to use the flag rather than opencode endless
* loop around allocator.
How do we perform this evaluation? And why is it preferable to a loop?
There are times when a loop makes sense, if there might be some other
event that could provide the needed memory ... or if a SIGKILL might
make it irrelevant.
slab allocators presumably shouldn't pass __GFP_NOFAIL to alloc_page(),
but should instead loop around
1/ check if any existing slabs have space
2/ if not, try to allocate a new page
Providing the latter blocks for a while but not indefinitely that should
be optimal.
Why is __GFP_NOFAIL better?
Because the allocator can do something if it knows that the allocation
cannot fail. E.g. give such an allocation a higher priority over those
that are allowed to fail. This is not limited to memory reserves,
although this is the only measure that is implemented currently IIRC.
On the other hand if there is something interesting the caller can do
directly - e.g. do internal object management like mempool does - then
it is better to retry at that level.
quoted
* Using this flag for costly allocations is _highly_ discouraged.
This is unhelpful. Saying something is "discouraged" carries an implied
threat. This is open source and threats need to be open.
Why is it discouraged? IF it is not forbidden, then it is clearly
permitted. Maybe there are costs - so a clear statement of those costs
would be appropriate.
Also, what is a suitable alternative?
Current code will trigger a WARN_ON, so it is effectively forbidden.
Maybe we should document that __GFP_NOFAIL is forbidden for orders above
1, and that vmalloc() should be used instead (thanks for proposing that
patch!).
I think we want to recommend kvmalloc as an alternative once vmalloc is
NOFAIL aware.
I will skip over some of the specific regarding SLAB and NOFS usage if
you do not mind and focus on points that have direct documentation
consequences. Also I do not feel qualified commenting on neither SLAB
nor FS internals.
[...]
There is a lot of stuff there.... the bits that are important to me are:
- why is __GFP_NOFAIL preferred? It is a valuable convenience, but I
don't see that it is necessary
I think it is preferred for one and a half reasons. It tells allocator
that this allocation cannot really fail and the caller doesn't have a
very good/clever retry policy (e.g. like mempools mentioned above). The
half reason would be for tracking purposes (git grep __GFP_NOFAIL) is
easier than trying to catch all sorts of while loops over allocation
which do not do anything really interesting.
- is it reasonable to use __GFP_HIGH when looping if there is a risk of
deadlock?
As I've said above. Memory reserves are a finite resource and as such
they cannot fundamentally solve deadlocks. They can help prioritize
though.
- Will __GFP_DIRECT_RECLAIM always result in a delay before failure? In
that case it should be safe to loop around allocations using
__GFP_DIRECT_RECLAIM without needing congestion_wait() (so it can
just be removed.
This is a good question and I do not think we have that documented
anywhere. We do cond_resched() for sure. I do not think we guarantee a
sleeping point in general. Maybe we should, I am not really sure.
Thanks for good comments and tough questions
--
Michal Hocko
SUSE Labs
[clip looks of discussion where we are largely in agreement - happy to
see that!]
quoted
Presumably there is a real risk of deadlock if we just remove the
memory-reserves boosts of __GFP_NOFAIL. Maybe it would be safe to
replace all current users of __GFP_NOFAIL with __GFP_NOFAIL|__GFP_HIGH,
and then remove the __GFP_HIGH where analysis suggests there is no risk
of deadlocks.
I would much rather not bind those together and go other way around. If
somebody can actually hit deadlocks (those are quite easy to spot as
they do not go away) then we can talk about how to deal with them.
Memory reserves can help only > < this much.
I recall maybe 10 years ago Linus saying that he preferred simplicity to
mathematical provability for handling memory deadlocks (or something
like that). I lean towards provability myself, but I do see the other
perspective.
We have mempools and they can provide strong guarantees (though they are
often over-allocated I think). But they can be a bit clumsy. I believe
that DaveM is strong against anything like that in the network layer, so
we strongly depend on GFP_MEMALLOC functionality for swap-over-NFS. I'm
sure it is important elsewhere too.
Of course __GFP_HIGH and __GFP_ATOMIC provide an intermediate priority
level - more likely to fail than __GFP_MEMALLOC. I suspect they should
not be seen as avoiding deadlock, only as improving service. So using
them when we cannot wait might make sense, but there are probably other
circumstances.
quoted
Why is __GFP_NOFAIL better?
Because the allocator can do something if it knows that the allocation
cannot fail. E.g. give such an allocation a higher priority over those
that are allowed to fail. This is not limited to memory reserves,
although this is the only measure that is implemented currently IIRC.
On the other hand if there is something interesting the caller can do
directly - e.g. do internal object management like mempool does - then
it is better to retry at that level.
It *can* do something, but I don't think it *should* do something - not
if that could have a negative impact on other threads. Just because I
cannot fail, that doesn't mean someone else should fail to help me.
Maybe I should just wait longer.
quoted
quoted
* Using this flag for costly allocations is _highly_ discouraged.
This is unhelpful. Saying something is "discouraged" carries an implied
threat. This is open source and threats need to be open.
Why is it discouraged? IF it is not forbidden, then it is clearly
permitted. Maybe there are costs - so a clear statement of those costs
would be appropriate.
Also, what is a suitable alternative?
Current code will trigger a WARN_ON, so it is effectively forbidden.
Maybe we should document that __GFP_NOFAIL is forbidden for orders above
1, and that vmalloc() should be used instead (thanks for proposing that
patch!).
I think we want to recommend kvmalloc as an alternative once vmalloc is
NOFAIL aware.
I will skip over some of the specific regarding SLAB and NOFS usage if
you do not mind and focus on points that have direct documentation
consequences. Also I do not feel qualified commenting on neither SLAB
nor FS internals.
[...]
quoted
There is a lot of stuff there.... the bits that are important to me are:
- why is __GFP_NOFAIL preferred? It is a valuable convenience, but I
don't see that it is necessary
I think it is preferred for one and a half reasons. It tells allocator
that this allocation cannot really fail and the caller doesn't have a
very good/clever retry policy (e.g. like mempools mentioned above). The
half reason would be for tracking purposes (git grep __GFP_NOFAIL) is
easier than trying to catch all sorts of while loops over allocation
which do not do anything really interesting.
I think the one reason is misguided, as described above.
I think the half reason is good, and that we should introduce
memalloc_retry_wait()
and encourage developers to use that for any memalloc retry loop.
__GFP_NOFAIL would then be a convenience flag which causes the allocator
(slab or alloc_page or whatever) to call memalloc_retry_wait() and do
the loop internally.
What exactly memalloc_retry_wait() does (if anything) can be decided
separately and changed as needed.
quoted
- is it reasonable to use __GFP_HIGH when looping if there is a risk of
deadlock?
As I've said above. Memory reserves are a finite resource and as such
they cannot fundamentally solve deadlocks. They can help prioritize
though.
To be fair, they can solve 1 level of deadlock. i.e. if you only need
to make one allocation to guarantee progress, then allocating from
reserves can help. If you might need to make a second allocation
without freeing the first - then a single reserve pool can provide
guarantees (which is why we use mempool is layered block devices - md
over dm over loop of scsi).
quoted
- Will __GFP_DIRECT_RECLAIM always result in a delay before failure? In
that case it should be safe to loop around allocations using
__GFP_DIRECT_RECLAIM without needing congestion_wait() (so it can
just be removed.
This is a good question and I do not think we have that documented
anywhere. We do cond_resched() for sure. I do not think we guarantee a
sleeping point in general. Maybe we should, I am not really sure.
If we add memalloc_retry_wait(), it wouldn't matter. We would only need
to ensure that memalloc_retry_wait() waited if page_alloc didn't.
I think we should:
- introduce memalloc_retry_wait() and use it for all malloc retry loops
including __GFP_NOFAIL
- drop all the priority boosts added for __GFP_NOFAIL
- drop __GFP_ATOMIC and change all the code that tests for __GFP_ATOMIC
to instead test for __GFP_HIGH. __GFP_ATOMIC is NEVER used without
__GFP_HIGH. This give a slight boost to several sites that use
__GFP_HIGH explicitly.
- choose a consistent order threshold for disallowing __GFP_NOFAIL
(rmqueue uses "order > 1", __alloc_pages_slowpath uses
"order > PAGE_ALLOC_COSTLY_ORDER"), test it once - early - and
document kvmalloc as an alternative. Code can also loop if there
is an alternative strategy for freeing up memory.
Thanks,
NeilBrown
Thanks,
NeilBrown