Commit "08778795"(block: Fix nr_vecs for inline integrity vectors) from
Martin introduces the function bip_integrity_vecs(get the useful vectors)
to fix the issue about nr_vecs for inline integrity vectors that reported
by David Milburn.
But it seems that bip_integrity_vecs() will return the wrong number if
the bio is not based on any bio_set for some reason(bio->bi_pool == NULL),
because in that case, the bip_inline_vecs[0] is malloced directly.
So here we add the bip_max_vcnt to record the count of vector slots, and
cleanup the function bip_integrity_vecs().
Signed-off-by: Gu Zheng <[email protected]>
---
block/bio-integrity.c | 12 +++---------
include/linux/bio.h | 1 +
2 files changed, 4 insertions(+), 9 deletions(-)
diff --git a/block/bio-integrity.c b/block/bio-integrity.c
index 9e24106..bc423f7b 100644
--- a/block/bio-integrity.c
+++ b/block/bio-integrity.c
@@ -70,8 +70,10 @@ struct bio_integrity_payload *bio_integrity_alloc(struct bio *bio,
bs->bvec_integrity_pool);
if (!bip->bip_vec)
goto err;
+ bip->bip_max_vcnt = bvec_nr_vecs(idx);
} else {
bip->bip_vec = bip->bip_inline_vecs;
+ bip->bip_max_vcnt = inline_vecs;
}
bip->bip_slab = idx;
@@ -114,14 +116,6 @@ void bio_integrity_free(struct bio *bio)
}
EXPORT_SYMBOL(bio_integrity_free);
-static inline unsigned int bip_integrity_vecs(struct bio_integrity_payload *bip)
-{
- if (bip->bip_slab == BIO_POOL_NONE)
- return BIP_INLINE_VECS;
-
- return bvec_nr_vecs(bip->bip_slab);
-}
-
/**
* bio_integrity_add_page - Attach integrity metadata
* @bio: bio to update
@@ -137,7 +131,7 @@ int bio_integrity_add_page(struct bio *bio, struct page *page,
struct bio_integrity_payload *bip = bio->bi_integrity;
struct bio_vec *iv;
- if (bip->bip_vcnt >= bip_integrity_vecs(bip)) {
+ if (bip->bip_vcnt >= bip->bip_max_vcnt) {
printk(KERN_ERR "%s: bip_vec full\n", __func__);
return 0;
}
diff --git a/include/linux/bio.h b/include/linux/bio.h
index 5a64576..1946d72 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -299,6 +299,7 @@ struct bio_integrity_payload {
unsigned short bip_slab; /* slab the bip came from */
unsigned short bip_vcnt; /* # of integrity bio_vecs */
+ unsigned short bip_max_vcnt; /* integrity bio_vec slots */
unsigned bip_owns_buf:1; /* should free bip_buf */
struct work_struct bip_work; /* I/O completion */
--
1.7.7
>>>>> "Gu" == Gu Zheng <[email protected]> writes:
[Sorry about the delay. I'm on vacation right now.]
Gu> But it seems that bip_integrity_vecs() will return the wrong number
Gu> if the bio is not based on any bio_set for some reason(bio->bi_pool
Gu> == NULL), because in that case, the bip_inline_vecs[0] is malloced
Gu> directly. So here we add the bip_max_vcnt to record the count of
Gu> vector slots, and cleanup the function bip_integrity_vecs().
I'm in agreement with your fix.
However, I'm still not sure what the use case is for bios without an
associated bioset. I do not see any callers that pass in a NULL bioset.
--
Martin K. Petersen Oracle Linux Engineering
Hi Martin,
On 06/30/2014 07:15 AM, Martin K. Petersen wrote:
>>>>>> "Gu" == Gu Zheng <[email protected]> writes:
>
> [Sorry about the delay. I'm on vacation right now.]
>
> Gu> But it seems that bip_integrity_vecs() will return the wrong number
> Gu> if the bio is not based on any bio_set for some reason(bio->bi_pool
> Gu> == NULL), because in that case, the bip_inline_vecs[0] is malloced
> Gu> directly. So here we add the bip_max_vcnt to record the count of
> Gu> vector slots, and cleanup the function bip_integrity_vecs().
>
> I'm in agreement with your fix.
Thanks.
>
> However, I'm still not sure what the use case is for bios without an
> associated bioset. I do not see any callers that pass in a NULL bioset.
Please refer to bio_kmalloc()--alloc bio via kmalloc.
And IMO, the API(e.g. bio_alloc_bioset) is EXPORT, any guys(including
some out of mainline code, some special drivers) can alloc a bio via
kmalloc without based on associated bioset.
Regards,
Gu
>
>>>>> "Gu" == Gu Zheng <[email protected]> writes:
Gu> Please refer to bio_kmalloc()--alloc bio via kmalloc.
OK, missed that one because it was in bio.h.
In that case we absolutely need your patch.
Acked-by: Martin K. Petersen <[email protected]>
--
Martin K. Petersen Oracle Linux Engineering