bpf: Minor fixes for non-preallocated memory
Subsystems (4)
v2 cover bpf: Minor fixes for non-preallocated memory
When I was implementing bpf recharge[1], I found some other issues. These issues are independent, so I send them separately.
[1]. https://lore.kernel.org/bpf/[email protected]/
v2:
- fix GFP_HIGH consistently over the bpf code. (Daniel, Roman)
- get rid of an error patch (Hao)
Yafang Shao (2):
bpf: Make non-preallocated allocation low priority
bpf: Warn on non-preallocated case for missed trace types
kernel/bpf/devmap.c | 3 ++- kernel/bpf/hashtab.c | 8 +++++--- kernel/bpf/local_storage.c | 3 ++- kernel/bpf/lpm_trie.c | 3 ++- kernel/bpf/verifier.c | 2 ++ 5 files changed, 13 insertions(+), 6 deletions(-)
Signature · 1 line
2.17.1
PATCH v2 1/2 bpf: Make non-preallocated allocation low priority
GFP_ATOMIC doesn't cooperate well with memcg pressure so far, especially if we allocate too much GFP_ATOMIC memory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, the GFP_ATOMIC can easily break the memcg limit by force charge. So it is very dangerous to use GFP_ATOMIC in non-preallocated case. One way to make it safe is to remove __GFP_HIGH from GFP_ATOMIC, IOW, use (__GFP_ATOMIC | __GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.
We introduced BPF_F_NO_PREALLOC is because full map pre-allocation is too memory expensive for some cases. That means removing __GFP_HIGH doesn't break the rule of BPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.
The force charge of GFP_ATOMIC was introduced in
commit 869712fd3de5 ("mm: memcontrol: fix network errors from failing __GFP_ATOMIC charges") by checking __GFP_ATOMIC, then got improved in commit 1461e8c2b6af ("memcg: unify force charging conditions") by checking __GFP_HIGH (that is no problem because both __GFP_HIGH and __GFP_ATOMIC are set in GFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIM doesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make __GFP_KSWAPD_RECLAIM work well under memcg pressure if desired.
It also fixes a typo in the comment.
--- kernel/bpf/devmap.c | 3 ++- kernel/bpf/hashtab.c | 8 +++++--- kernel/bpf/local_storage.c | 3 ++- kernel/bpf/lpm_trie.c | 3 ++- 4 files changed, 11 insertions(+), 6 deletions(-)
kernel/bpf/devmap.c +2 -1
@@ -845,7 +845,8 @@ static struct bpf_dtab_netdev *__dev_map_alloc_node(struct net *net,struct bpf_dtab_netdev *dev; dev =bpf_map_kmalloc_node (&dtab->map,sizeof (*dev), -GFP_ATOMIC | __GFP_NOWARN, + __GFP_ATOMIC | __GFP_NOWARN | + __GFP_KSWAPD_RECLAIM, dtab->map.numa_node);if (!dev)return ERR_PTR (-ENOMEM );
kernel/bpf/hashtab.c +5 -3
@@ -61,7 +61,7 @@* * As regular device interrupt handlers and soft interrupts are forced into * thread context, the existing code which does -* spin_lock*(); alloc(GPF_ATOMIC); spin_unlock*(); +* spin_lock*(); alloc(GFP_ATOMIC); spin_unlock*(); * just works. * * In theory the BPF locks could be converted to regular spinlocks as well,
@@ -978,7 +978,8 @@ static struct htab_elem *alloc_htab_elem(struct bpf_htab *htab, void *key, goto dec_count; } l_new =bpf_map_kmalloc_node (&htab->map, htab->elem_size, -GFP_ATOMIC | __GFP_NOWARN, + __GFP_ATOMIC | __GFP_NOWARN | + __GFP_KSWAPD_RECLAIM, htab->map.numa_node);if (!l_new) { l_new =ERR_PTR (-ENOMEM );
@@ -996,7 +997,8 @@ static struct htab_elem *alloc_htab_elem(struct bpf_htab *htab, void *key, }else {/* alloc_percpu zero-fills */ pptr =bpf_map_alloc_percpu (&htab->map, size,8 , -GFP_ATOMIC | __GFP_NOWARN); + __GFP_ATOMIC | __GFP_NOWARN | + __GFP_KSWAPD_RECLAIM);if (!pptr) {kfree (l_new); l_new =ERR_PTR (-ENOMEM );
kernel/bpf/local_storage.c +2 -1
@@ -165,7 +165,8 @@ static int cgroup_storage_update_elem(struct bpf_map *map, void *key, } new =bpf_map_kmalloc_node (map,struct_size (new, data, map->value_size), - __GFP_ZERO |GFP_ATOMIC | __GFP_NOWARN, + __GFP_ZERO | __GFP_ATOMIC | __GFP_NOWARN | + __GFP_KSWAPD_RECLAIM, map->numa_node);if (!new)return -ENOMEM ;
kernel/bpf/lpm_trie.c +2 -1
@@ -285,7 +285,8 @@ static struct lpm_trie_node *lpm_trie_node_alloc(const struct lpm_trie *trie,if (value) size += trie->map.value_size; - node =bpf_map_kmalloc_node (&trie->map, size,GFP_ATOMIC | __GFP_NOWARN, + node =bpf_map_kmalloc_node (&trie->map, size, __GFP_ATOMIC | + __GFP_KSWAPD_RECLAIM | __GFP_NOWARN, trie->map.numa_node);if (!node)return NULL ;
Signature · 1 line
2.17.1
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 24 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.We introduced
BPF_F_NO_PREALLOCis because full map pre-allocation is too memory expensive for some cases. That means removing__GFP_HIGHdoesn't break the rule ofBPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.The force charge of
GFP_ATOMICwas introduced in
commit869712fd3de5("mm: memcontrol: fix network errors from failing__GFP_ATOMICcharges") by checking__GFP_ATOMIC, then got improved in commit1461e8c2b6af("memcg: unify force charging conditions") by checking__GFP_HIGH(that is no problem because both__GFP_HIGHand__GFP_ATOMICare set inGFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.
Quoted 7 lines
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIMdoesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make__GFP_KSWAPD_RECLAIMwork well under memcg pressure if desired.
Could you elaborate ?
Quoted 4 lines
It also fixes a typo in the comment.
Signed-off-by Yafang Shao <[email protected]>Reviewed-by Roman Gushchin <[email protected]>
Roman, do you agree with this change ?
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 33 lines · Alexei Starovoitov
Quoted 24 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.We introduced
BPF_F_NO_PREALLOCis because full map pre-allocation is too memory expensive for some cases. That means removing__GFP_HIGHdoesn't break the rule ofBPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.The force charge of
GFP_ATOMICwas introduced in
commit869712fd3de5("mm: memcontrol: fix network errors from failing__GFP_ATOMICcharges") by checking__GFP_ATOMIC, then got improved in commit1461e8c2b6af("memcg: unify force charging conditions") by checking__GFP_HIGH(that is no problem because both__GFP_HIGHand__GFP_ATOMICare set inGFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.Quoted 7 lines
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIMdoesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make__GFP_KSWAPD_RECLAIMwork well under memcg pressure if desired.
Quoted 8 lines
Could you elaborate ?
Quoted 4 lines
It also fixes a typo in the comment.
Signed-off-by Yafang Shao <[email protected]>Reviewed-by Roman Gushchin <[email protected]>Roman, do you agree with this change ?
Yes, removing __GFP_HIGH makes sense to me. I can imagine we might want it for *some* bpf allocations, but applying it unconditionally looks wrong.
Thanks!
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 39 lines · Roman Gushchin
Quoted 37 lines · Alexei Starovoitov
Quoted 24 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.We introduced
BPF_F_NO_PREALLOCis because full map pre-allocation is too memory expensive for some cases. That means removing__GFP_HIGHdoesn't break the rule ofBPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.The force charge of
GFP_ATOMICwas introduced in
commit869712fd3de5("mm: memcontrol: fix network errors from failing__GFP_ATOMICcharges") by checking__GFP_ATOMIC, then got improved in commit1461e8c2b6af("memcg: unify force charging conditions") by checking__GFP_HIGH(that is no problem because both__GFP_HIGHand__GFP_ATOMICare set inGFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.Quoted 7 lines
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIMdoesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make__GFP_KSWAPD_RECLAIMwork well under memcg pressure if desired.Could you elaborate ?
Quoted 2 lines
It also fixes a typo in the comment.
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines
Signed-off-by Yafang Shao <[email protected]>Reviewed-by Roman Gushchin <[email protected]>Roman, do you agree with this change ?
Yes, removing
__GFP_HIGHmakes sense to me. I can imagine we might want it for *some* bpf allocations, but applying it unconditionally looks wrong.
Yeah. It's a difficult trade-off to make without having the data to decide whether removing __GFP_HIGH can cause issues or not, but do you agree that __GFP_HIGH doesn't cooperate well with memcg ? If so it's a bug on memcg side, right? but we should probably apply this band-aid on bpf side to fix the bleeding. Later we can add a knob to allow __GFP_HIGH usage on demand from bpf prog.
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 43 lines · Alexei Starovoitov
Quoted 42 lines · Roman Gushchin
Quoted 40 lines · Alexei Starovoitov
Quoted 24 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.We introduced
BPF_F_NO_PREALLOCis because full map pre-allocation is too memory expensive for some cases. That means removing__GFP_HIGHdoesn't break the rule ofBPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.The force charge of
GFP_ATOMICwas introduced in
commit869712fd3de5("mm: memcontrol: fix network errors from failing__GFP_ATOMICcharges") by checking__GFP_ATOMIC, then got improved in commit1461e8c2b6af("memcg: unify force charging conditions") by checking__GFP_HIGH(that is no problem because both__GFP_HIGHand__GFP_ATOMICare set inGFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.Quoted 7 lines
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIMdoesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make__GFP_KSWAPD_RECLAIMwork well under memcg pressure if desired.Could you elaborate ?
Quoted 4 lines
It also fixes a typo in the comment.
Signed-off-by Yafang Shao <[email protected]>Reviewed-by Roman Gushchin <[email protected]>
Quoted 7 lines
Quoted 4 lines
Quoted 1 line
Roman, do you agree with this change ?
Yes, removing
__GFP_HIGHmakes sense to me. I can imagine we might want it for *some* bpf allocations, but applying it unconditionally looks wrong.Yeah. It's a difficult trade-off to make without having the data to decide whether removing
__GFP_HIGHcan cause issues or not,
Yeah, the change looks reasonable, but it's hard to say without giving it a good testing in (something close to) a production environment.
Quoted 2 lines
but do you agree that
__GFP_HIGHdoesn't cooperate well with memcg ? If so it's a bug on memcg side, right?
No. Historically we allowed high-prio allocations to exceed the memcg limit because otherwise there were too many stability and performance issues. It's not a memcg bug, it's a way to avoid exposing ENOMEM handling bugs all over the kernel code. Without memory cgroups GFP_ATOMIC allocations rarely fail and a lot of code paths in the kernel are not really ready for it (at least it was the case several years ago, maybe things are better now).
But it was usually thought in the context of small(ish) allocations which do not change the global memory usage picture. Subsequent "normal" allocations are triggering reclaim/OOM, so from a user's POV the limit works as expected.
But with the ownership model and size of bpf maps it's a different story:
if a bpf map belongs to an abandoned cgroup, it might consume a lot of memory and there will be no "normal" allocations. So cgroup memory limit will be never applied. It's a valid issue, I agree with Yafang here.
Quoted 4 lines
but we should probably
apply this band-aid on bpf side to fix the bleeding. Later we can add a knob to allow__GFP_HIGHusage on demand from bpf prog.
Yes, it sounds like a good idea. I have to think what's the best approach here, it's not obvious for me.
Thanks!
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 66 lines · Roman Gushchin
Quoted 43 lines · Alexei Starovoitov
Quoted 42 lines · Roman Gushchin
Quoted 40 lines · Alexei Starovoitov
Quoted 24 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.We introduced
BPF_F_NO_PREALLOCis because full map pre-allocation is too memory expensive for some cases. That means removing__GFP_HIGHdoesn't break the rule ofBPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.The force charge of
GFP_ATOMICwas introduced in
commit869712fd3de5("mm: memcontrol: fix network errors from failing__GFP_ATOMICcharges") by checking__GFP_ATOMIC, then got improved in commit1461e8c2b6af("memcg: unify force charging conditions") by checking__GFP_HIGH(that is no problem because both__GFP_HIGHand__GFP_ATOMICare set inGFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.Quoted 7 lines
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIMdoesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make__GFP_KSWAPD_RECLAIMwork well under memcg pressure if desired.Could you elaborate ?
Quoted 4 lines
It also fixes a typo in the comment.
Signed-off-by Yafang Shao <[email protected]>Reviewed-by Roman Gushchin <[email protected]>Quoted 7 lines
Quoted 4 lines
Quoted 1 line
Roman, do you agree with this change ?
Yes, removing
__GFP_HIGHmakes sense to me. I can imagine we might want it for *some* bpf allocations, but applying it unconditionally looks wrong.Yeah. It's a difficult trade-off to make without having the data to decide whether removing
__GFP_HIGHcan cause issues or not,Yeah, the change looks reasonable, but it's hard to say without giving it a good testing in (something close to) a production environment.
Quoted 2 lines
but do you agree that
__GFP_HIGHdoesn't cooperate well with memcg ? If so it's a bug on memcg side, right?No. Historically we allowed high-prio allocations to exceed the memcg limit because otherwise there were too many stability and performance issues. It's not a memcg bug, it's a way to avoid exposing ENOMEM handling bugs all over the kernel code. Without memory cgroups
GFP_ATOMICallocations rarely fail and a lot of code paths in the kernel are not really ready for it (at least it was the case several years ago, maybe things are better now).
Quoted 8 lines
But it was usually thought in the context of small(ish) allocations which do not change the global memory usage picture. Subsequent "normal" allocations are triggering reclaim/OOM, so from a user's POV the limit works as expected.
But with the ownership model and size of bpf maps it's a different story:
if a bpf map belongs to an abandoned cgroup, it might consume a lot of memory and there will be no "normal" allocations. So cgroup memory limit will be never applied. It's a valid issue, I agree with Yafang here.
Understood.
Quoted 7 lines
Quoted 4 lines
but we should probably
apply this band-aid on bpf side to fix the bleeding. Later we can add a knob to allow__GFP_HIGHusage on demand from bpf prog.Yes, it sounds like a good idea. I have to think what's the best approach here, it's not obvious for me.
Ok. Applied this patch for now.
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
In reply to Yafang ShaoQuoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.
Please use GFP_NOWAIT instead of (__GFP_ATOMIC | __GFP_KSWAPD_RECLAIM). There is already a plan to completely remove __GFP_ATOMIC and mm-tree already have a patch for that.
Quoted 15 lines
We introduced
BPF_F_NO_PREALLOCis because full map pre-allocation is too memory expensive for some cases. That means removing__GFP_HIGHdoesn't break the rule ofBPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.The force charge of
GFP_ATOMICwas introduced in
commit869712fd3de5("mm: memcontrol: fix network errors from failing__GFP_ATOMICcharges") by checking__GFP_ATOMIC, then got improved in commit1461e8c2b6af("memcg: unify force charging conditions") by checking__GFP_HIGH(that is no problem because both__GFP_HIGHand__GFP_ATOMICare set inGFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.
Quoted 8 lines
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIMdoesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make__GFP_KSWAPD_RECLAIMwork well under memcg pressure if desired.
IMO there is no need to give all this detail and background on GFP_ATOMIC and __GFP_KSWAPD_RECLAIM. Just say kernel allows GFP_ATOMIC allocations to exceed memcg limits which we don't want in this case. So, replace with GFP_NOWAIT which obey memcg limits. Both of these flags tell kernel that the caller can not sleep.
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 11 lines · Shakeel Butt
Quoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.
Quoted 3 lines
Please use
GFP_NOWAITinstead of (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM). There is already a plan to completely remove__GFP_ATOMICand mm-tree already have a patch for that.
hmm. ok. reverted.
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
In reply to Shakeel ButtQuoted 10 lines · Shakeel Butt
Quoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.
Quoted 3 lines
Please use
GFP_NOWAITinstead of (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM). There is already a plan to completely remove__GFP_ATOMICand mm-tree already have a patch for that.
Oh, I didn't know this, thanks for heads up!
I agree that GFP_NOWAIT is the best choice then.
Btw, we probably shouldn't even add GFP_NOWAIT if the allocation is performed from the bpf syscall context. Why would we fail to pre-allocate a map if we can easily go into the reclaim? But probably better to leave it for a separate change.
Thanks!
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 16 lines · Roman Gushchin
Quoted 10 lines · Shakeel Butt
Quoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.Quoted 3 lines
Please use
GFP_NOWAITinstead of (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM). There is already a plan to completely remove__GFP_ATOMICand mm-tree already have a patch for that.
Quoted 7 lines
Oh, I didn't know this, thanks for heads up!
I agree thatGFP_NOWAITis the best choice then.Btw, we probably shouldn't even add
GFP_NOWAITif the allocation is performed from the bpf syscall context. Why would we fail to pre-allocate a map if we can easily go into the reclaim? But probably better to leave it for a separate change.
The places affected by this patch are in atomic context. Prealloc path from syscall is using GFP_USER.
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 20 lines · Alexei Starovoitov
Quoted 19 lines · Roman Gushchin
Quoted 10 lines · Shakeel Butt
Quoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.Quoted 3 lines
Please use
GFP_NOWAITinstead of (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM). There is already a plan to completely remove__GFP_ATOMICand mm-tree already have a patch for that.Oh, I didn't know this, thanks for heads up!
I agree thatGFP_NOWAITis the best choice then.
Quoted 7 lines
Quoted 4 lines
Btw, we probably shouldn't even add
GFP_NOWAITif the allocation is performed from the bpf syscall context. Why would we fail to pre-allocate a map if we can easily go into the reclaim? But probably better to leave it for a separate change.The places affected by this patch are in atomic context. Prealloc path from syscall is using
GFP_USER.
Right. Sorry, my bad, for some reason I was under an impression it's a common path for all allocations.
Thanks!
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
In reply to Shakeel ButtQuoted 11 lines · Shakeel Butt
Quoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.
Quoted 4 lines
Please use
GFP_NOWAITinstead of (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM). There is already a plan to completely remove__GFP_ATOMICand mm-tree already have a patch for that.
After reading the discussion[1], it looks good to me to use GFP_NOWAIT instead. I will update it.
[1]. https://lore.kernel.org/linux-mm/[email protected]/
Quoted 24 lines
Quoted 15 lines
We introduced
BPF_F_NO_PREALLOCis because full map pre-allocation is too memory expensive for some cases. That means removing__GFP_HIGHdoesn't break the rule ofBPF_F_NO_PREALLOC, but has the same goal with it-avoiding issues caused by too much memory. So let's remove it.The force charge of
GFP_ATOMICwas introduced in
commit869712fd3de5("mm: memcontrol: fix network errors from failing__GFP_ATOMICcharges") by checking__GFP_ATOMIC, then got improved in commit1461e8c2b6af("memcg: unify force charging conditions") by checking__GFP_HIGH(that is no problem because both__GFP_HIGHand__GFP_ATOMICare set inGFP_AOMIC). So, if we want to fix it in memcg, we have to carefully verify all the callsites. Now that we can fix it in BPF, we'd better not modify the memcg code.Quoted 8 lines
This fix can also apply to other run-time allocations, for example, the allocation in lpm trie, local storage and devmap. So let fix it consistently over the bpf code
__GFP_KSWAPD_RECLAIMdoesn't cooperate well with memcg pressure neither currently. But the memcg code can be improved to make__GFP_KSWAPD_RECLAIMwork well under memcg pressure if desired.
Quoted 6 lines
IMO there is no need to give all this detail and background on
GFP_ATOMICand__GFP_KSWAPD_RECLAIM. Just say kernel allowsGFP_ATOMICallocations to exceed memcg limits which we don't want in this case. So, replace withGFP_NOWAITwhich obey memcg limits. Both of these flags tell kernel that the caller can not sleep.
Sure, thanks.
Signature · 2 lines
Regards Yafang
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 13 lines · Yafang Shao
Quoted 11 lines · Shakeel Butt
Quoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.
Quoted 7 lines
Quoted 4 lines
Please use
GFP_NOWAITinstead of (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM). There is already a plan to completely remove__GFP_ATOMICand mm-tree already have a patch for that.After reading the discussion[1], it looks good to me to use
GFP_NOWAITinstead. I will update it.
Should we use GFP_ATOMIC | __GFP_NOMEMALLOC instead to align with its usage in the networking stack?
Re: [PATCH bpf-next v2 1/2] bpf: Make non-preallocated allocation low priority
Quoted 20 lines · Alexei Starovoitov
Quoted 18 lines · Yafang Shao
Quoted 15 lines · Shakeel Butt
Quoted 8 lines · Yafang Shao
GFP_ATOMICdoesn't cooperate well with memcg pressure so far, especially if we allocate too muchGFP_ATOMICmemory. For example, when we set the memcg limit to limit a non-preallocated bpf memory, theGFP_ATOMICcan easily break the memcg limit by force charge. So it is very dangerous to useGFP_ATOMICin non-preallocated case. One way to make it safe is to remove__GFP_HIGHfromGFP_ATOMIC, IOW, use (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM) instead, then it will be limited if we allocate too much memory.Please use
GFP_NOWAITinstead of (__GFP_ATOMIC|__GFP_KSWAPD_RECLAIM). There is already a plan to completely remove__GFP_ATOMICand mm-tree already have a patch for that.
Quoted 5 lines
Quoted 2 lines
After reading the discussion[1], it looks good to me to use
GFP_NOWAITinstead. I will update it.Should we use
GFP_ATOMIC|__GFP_NOMEMALLOCinstead to align with its usage in the networking stack?
GFP_ATOMIC | __GFP_NOMEMALLOC will continue to break the memcg limit, so we have to modify the try_charge_memcg() code to make sure __GFP_NOMEMALLOC takes precedence over the __GFP_HIGH flag, IOW, if both of them are set we won't allow it to break memcg limit. That will need more verification.
Signature · 2 lines
Regards Yafang
PATCH v2 2/2 bpf: Warn on non-preallocated case for missed trace types
In reply to Yafang ShaoBPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE and BPF_PROG_TYPE_TRACING are trace type as well, which may also cause unexpected memory allocation if we set BPF_F_NO_PREALLOC.
Let's also warn on both of them.
--- kernel/bpf/verifier.c | 2 ++ 1 file changed, 2 insertions(+)
kernel/bpf/verifier.c +2 -0
@@ -12570,6 +12570,8 @@ static bool is_tracing_prog_type(enum bpf_prog_type type)case BPF_PROG_TYPE_TRACEPOINT :case BPF_PROG_TYPE_PERF_EVENT :case BPF_PROG_TYPE_RAW_TRACEPOINT : +case BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE : +case BPF_PROG_TYPE_TRACING :return true;default :return false;
Signature · 1 line
2.17.1
Re: [PATCH bpf-next v2 2/2] bpf: Warn on non-preallocated case for missed trace types
Quoted 16 lines · Yafang Shao
BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLEandBPF_PROG_TYPE_TRACINGare trace type as well, which may also cause unexpected memory allocation if we setBPF_F_NO_PREALLOC.
Let's also warn on both of them.Signed-off-by Yafang Shao <[email protected]>--- kernel/bpf/verifier.c | 2 ++ 1 file changed, 2 insertions(+)kernel/bpf/verifier.c +0
-0
kernel/bpf/verifier.c static bool is_tracing_prog_type(enum bpf_prog_type type) · line 12574
Quoted 5 lines
kernel/bpf/verifier.c +2
-0@@ -12570,6 +12570,8 @@ static bool is_tracing_prog_type(enum bpf_prog_type type)case BPF_PROG_TYPE_TRACEPOINT :case BPF_PROG_TYPE_PERF_EVENT :case BPF_PROG_TYPE_RAW_TRACEPOINT : +case BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE : +case BPF_PROG_TYPE_TRACING :
BPF_TRACE_ITER should probably be excluded.
Re: [PATCH bpf-next v2 2/2] bpf: Warn on non-preallocated case for missed trace types
Quoted 13 lines · Alexei Starovoitov
Quoted 11 lines · Yafang Shao
BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLEandBPF_PROG_TYPE_TRACINGare trace type as well, which may also cause unexpected memory allocation if we setBPF_F_NO_PREALLOC.
Let's also warn on both of them.Signed-off-by Yafang Shao <[email protected]>--- kernel/bpf/verifier.c | 2 ++ 1 file changed, 2 insertions(+)
Quoted 12 lines
kernel/bpf/verifier.c static bool is_tracing_prog_type(enum bpf_prog_type type) · line 12574
Quoted 10 lines
kernel/bpf/verifier.c +2
-0@@ -12570,6 +12570,8 @@ static bool is_tracing_prog_type(enum bpf_prog_type type)case BPF_PROG_TYPE_TRACEPOINT :case BPF_PROG_TYPE_PERF_EVENT :case BPF_PROG_TYPE_RAW_TRACEPOINT : +case BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE : +case BPF_PROG_TYPE_TRACING :
BPF_TRACE_ITERshould probably be excluded.
Right, I have verified that BPF_TRACE_ITER can be excluded. Will change it.
Signature · 2 lines
Regards Yafang
Re: [PATCH bpf-next v2 2/2] bpf: Warn on non-preallocated case for missed trace types
Quoted 29 lines · Yafang Shao
Quoted 13 lines · Alexei Starovoitov
Quoted 11 lines · Yafang Shao
BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLEandBPF_PROG_TYPE_TRACINGare trace type as well, which may also cause unexpected memory allocation if we setBPF_F_NO_PREALLOC.
Let's also warn on both of them.Signed-off-by Yafang Shao <[email protected]>--- kernel/bpf/verifier.c | 2 ++ 1 file changed, 2 insertions(+)Quoted 12 lines
kernel/bpf/verifier.c static bool is_tracing_prog_type(enum bpf_prog_type type) · line 12574
Quoted 10 lines
kernel/bpf/verifier.c +2
-0@@ -12570,6 +12570,8 @@ static bool is_tracing_prog_type(enum bpf_prog_type type)case BPF_PROG_TYPE_TRACEPOINT :case BPF_PROG_TYPE_PERF_EVENT :case BPF_PROG_TYPE_RAW_TRACEPOINT : +case BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE : +case BPF_PROG_TYPE_TRACING :
BPF_TRACE_ITERshould probably be excluded.
Quoted 2 lines
Right, I have verified that
BPF_TRACE_ITERcan be excluded. Will change it.
Probably more than that. See that your change broke BPF CI and selftests are failing.
Which means it breaks existing bpf programs.
Re: [PATCH bpf-next v2 2/2] bpf: Warn on non-preallocated case for missed trace types
Quoted 31 lines · Alexei Starovoitov
Quoted 29 lines · Yafang Shao
Quoted 25 lines · Alexei Starovoitov
Quoted 16 lines · Yafang Shao
BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLEandBPF_PROG_TYPE_TRACINGare trace type as well, which may also cause unexpected memory allocation if we setBPF_F_NO_PREALLOC.
Let's also warn on both of them.Signed-off-by Yafang Shao <[email protected]>--- kernel/bpf/verifier.c | 2 ++ 1 file changed, 2 insertions(+)kernel/bpf/verifier.c +0
-0kernel/bpf/verifier.c static bool is_tracing_prog_type(enum bpf_prog_type type) · line 12574
Quoted 5 lines
kernel/bpf/verifier.c +2
-0@@ -12570,6 +12570,8 @@ static bool is_tracing_prog_type(enum bpf_prog_type type)case BPF_PROG_TYPE_TRACEPOINT :case BPF_PROG_TYPE_PERF_EVENT :case BPF_PROG_TYPE_RAW_TRACEPOINT : +case BPF_PROG_TYPE_RAW_TRACEPOINT_WRITABLE : +case BPF_PROG_TYPE_TRACING :
BPF_TRACE_ITERshould probably be excluded.
Quoted 6 lines
Quoted 2 lines
Right, I have verified that
BPF_TRACE_ITERcan be excluded. Will change it.Probably more than that. See that your change broke BPF CI and selftests are failing.
Which means it breaks existing bpf programs.
Ah, yes, "#194 timer:FAIL" which is caused by this change. I will be more careful here.
Signature · 2 lines
Regards Yafang