mas_preallocate() defaults to requesting 1 node for preallocation and then
,depending on the type of store, will update the request variable. There
isn't a check for a slot store type, so slot stores are preallocating the
default 1 node. Slot stores do not require any additional nodes, so add a
check for the slot store case that will bypass node_count_gfp(). Update
the tests to reflect that slot stores do not require allocations.
User visible effects of this bug include increased memory usage from the
unneeded node that was allocated.
Fixes: 0b8bb544b1a7 ("maple_tree: update mas_preallocate() testing")
Cc: <[email protected]> # 6.6+
Signed-off-by: Sidhartha Kumar <[email protected]>
---
This is a modified backport as the patch to fix this in upstream does not
apply to 6.6 because the node_end field was moved from the ma_wr_state to
the ma_state after 6.6.
lib/maple_tree.c | 6 ++++++
tools/testing/radix-tree/maple.c | 2 +-
2 files changed, 7 insertions(+), 1 deletion(-)
diff --git a/lib/maple_tree.c b/lib/maple_tree.c
index bb24d84a4922f..5950d0c0e0f69 100644
--- a/lib/maple_tree.c
+++ b/lib/maple_tree.c
@@ -5501,6 +5501,12 @@ int mas_preallocate(struct ma_state *mas, void *entry, gfp_t gfp)
mas_wr_end_piv(&wr_mas);
node_size = mas_wr_new_end(&wr_mas);
+
+ /* Slot store, does not require additional nodes */
+ if ((node_size == wr_mas.node_end) && ((!mt_in_rcu(mas->tree))
+ || (wr_mas.offset_end - mas->offset == 1)))
+ return 0;
+
if (node_size >= mt_slots[wr_mas.type]) {
/* Split, worst case for now. */
request = 1 + mas_mt_height(mas) * 2;
diff --git a/tools/testing/radix-tree/maple.c b/tools/testing/radix-tree/maple.c
index e5da1cad70baf..76a8990bb14e8 100644
--- a/tools/testing/radix-tree/maple.c
+++ b/tools/testing/radix-tree/maple.c
@@ -35538,7 +35538,7 @@ static noinline void __init check_prealloc(struct maple_tree *mt)
MT_BUG_ON(mt, mas_preallocate(&mas, ptr, GFP_KERNEL) != 0);
allocated = mas_allocated(&mas);
height = mas_mt_height(&mas);
- MT_BUG_ON(mt, allocated != 1);
+ MT_BUG_ON(mt, allocated != 0);
mas_store_prealloc(&mas, ptr);
MT_BUG_ON(mt, mas_allocated(&mas) != 0);
--
2.42.0
On 12/12/23 11:52 AM, Sidhartha Kumar wrote:
> mas_preallocate() defaults to requesting 1 node for preallocation and then
> ,depending on the type of store, will update the request variable. There
> isn't a check for a slot store type, so slot stores are preallocating the
> default 1 node. Slot stores do not require any additional nodes, so add a
> check for the slot store case that will bypass node_count_gfp(). Update
> the tests to reflect that slot stores do not require allocations.
>
> User visible effects of this bug include increased memory usage from the
> unneeded node that was allocated.
>
> Fixes: 0b8bb544b1a7 ("maple_tree: update mas_preallocate() testing")
> Cc: <[email protected]> # 6.6+
> Signed-off-by: Sidhartha Kumar <[email protected]>
> ---
> This is a modified backport as the patch to fix this in upstream does not
> apply to 6.6 because the node_end field was moved from the ma_wr_state to
> the ma_state after 6.6.
Please disregard this patch, the upstream fix[1] does actually apply correctly
to the 6.6 stable branch.
Sorry for the noise,
Sid
[1]
https://lore.kernel.org/linux-mm/[email protected]/T/#u
> lib/maple_tree.c | 6 ++++++
> tools/testing/radix-tree/maple.c | 2 +-
> 2 files changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/lib/maple_tree.c b/lib/maple_tree.c
> index bb24d84a4922f..5950d0c0e0f69 100644
> --- a/lib/maple_tree.c
> +++ b/lib/maple_tree.c
> @@ -5501,6 +5501,12 @@ int mas_preallocate(struct ma_state *mas, void *entry, gfp_t gfp)
>
> mas_wr_end_piv(&wr_mas);
> node_size = mas_wr_new_end(&wr_mas);
> +
> + /* Slot store, does not require additional nodes */
> + if ((node_size == wr_mas.node_end) && ((!mt_in_rcu(mas->tree))
> + || (wr_mas.offset_end - mas->offset == 1)))
> + return 0;
> +
> if (node_size >= mt_slots[wr_mas.type]) {
> /* Split, worst case for now. */
> request = 1 + mas_mt_height(mas) * 2;
> diff --git a/tools/testing/radix-tree/maple.c b/tools/testing/radix-tree/maple.c
> index e5da1cad70baf..76a8990bb14e8 100644
> --- a/tools/testing/radix-tree/maple.c
> +++ b/tools/testing/radix-tree/maple.c
> @@ -35538,7 +35538,7 @@ static noinline void __init check_prealloc(struct maple_tree *mt)
> MT_BUG_ON(mt, mas_preallocate(&mas, ptr, GFP_KERNEL) != 0);
> allocated = mas_allocated(&mas);
> height = mas_mt_height(&mas);
> - MT_BUG_ON(mt, allocated != 1);
> + MT_BUG_ON(mt, allocated != 0);
> mas_store_prealloc(&mas, ptr);
> MT_BUG_ON(mt, mas_allocated(&mas) != 0);
>
On Tue, Dec 12, 2023 at 11:52:55AM -0800, Sidhartha Kumar wrote:
> mas_preallocate() defaults to requesting 1 node for preallocation and then
> ,depending on the type of store, will update the request variable. There
> isn't a check for a slot store type, so slot stores are preallocating the
> default 1 node. Slot stores do not require any additional nodes, so add a
> check for the slot store case that will bypass node_count_gfp(). Update
> the tests to reflect that slot stores do not require allocations.
>
> User visible effects of this bug include increased memory usage from the
> unneeded node that was allocated.
>
> Fixes: 0b8bb544b1a7 ("maple_tree: update mas_preallocate() testing")
> Cc: <[email protected]> # 6.6+
> Signed-off-by: Sidhartha Kumar <[email protected]>
> ---
> This is a modified backport as the patch to fix this in upstream does not
> apply to 6.6 because the node_end field was moved from the ma_wr_state to
> the ma_state after 6.6.
What is the git commit id of this change in Linus's tree?
thanks,
greg k-h
On 12/18/23 2:59 AM, Greg KH wrote:
> On Tue, Dec 12, 2023 at 11:52:55AM -0800, Sidhartha Kumar wrote:
>> mas_preallocate() defaults to requesting 1 node for preallocation and then
>> ,depending on the type of store, will update the request variable. There
>> isn't a check for a slot store type, so slot stores are preallocating the
>> default 1 node. Slot stores do not require any additional nodes, so add a
>> check for the slot store case that will bypass node_count_gfp(). Update
>> the tests to reflect that slot stores do not require allocations.
>>
>> User visible effects of this bug include increased memory usage from the
>> unneeded node that was allocated.
>>
>> Fixes: 0b8bb544b1a7 ("maple_tree: update mas_preallocate() testing")
>> Cc: <[email protected]> # 6.6+
>> Signed-off-by: Sidhartha Kumar <[email protected]>
>> ---
>> This is a modified backport as the patch to fix this in upstream does not
>> apply to 6.6 because the node_end field was moved from the ma_wr_state to
>> the ma_state after 6.6.
>
> What is the git commit id of this change in Linus's tree?
The patch is in akpm's mm-hotfixes-unstable tree and has not made it to Linus's
tree yet.
Thanks,
Sid
>
> thanks,
>
> greg k-h
On Mon, Dec 18, 2023 at 09:53:53AM -0800, Sidhartha Kumar wrote:
> On 12/18/23 2:59 AM, Greg KH wrote:
> > On Tue, Dec 12, 2023 at 11:52:55AM -0800, Sidhartha Kumar wrote:
> > > mas_preallocate() defaults to requesting 1 node for preallocation and then
> > > ,depending on the type of store, will update the request variable. There
> > > isn't a check for a slot store type, so slot stores are preallocating the
> > > default 1 node. Slot stores do not require any additional nodes, so add a
> > > check for the slot store case that will bypass node_count_gfp(). Update
> > > the tests to reflect that slot stores do not require allocations.
> > >
> > > User visible effects of this bug include increased memory usage from the
> > > unneeded node that was allocated.
> > >
> > > Fixes: 0b8bb544b1a7 ("maple_tree: update mas_preallocate() testing")
> > > Cc: <[email protected]> # 6.6+
> > > Signed-off-by: Sidhartha Kumar <[email protected]>
> > > ---
> > > This is a modified backport as the patch to fix this in upstream does not
> > > apply to 6.6 because the node_end field was moved from the ma_wr_state to
> > > the ma_state after 6.6.
> >
> > What is the git commit id of this change in Linus's tree?
>
> The patch is in akpm's mm-hotfixes-unstable tree and has not made it to
> Linus's tree yet.
Ok, please resend when it hits Linus's tree.