While the endiannes is being handled correctly as indicated by the comment
above the offending line - sparse was unhappy with the missing annotation
as be64_to_cpu() expects a __be64 argument. To mitigate this annotation
all involved variables are changed to a consistent __le64 and the
conversion to uint64_t delayed to the call to rds_cong_map_updated().
Signed-off-by: Nicholas Mc Guire <[email protected]>
---
Problem located by an experimental coccinelle script to locate
patters that make sparse unhappy (false positives):
net/rds/ib_recv.c:827:23: warning: cast to restricted __le64
V2: Edward Cree <[email protected]> rejected the need for using __force
here - instead solve the sparse issue by updating all of the involved
variables - which results in an identical binary as well without using
the __force "solution" to the sparse warning. Thanks !
Patch was compile-tested with: x86_64_defconfig + INFINIBAND=m, RDS_RDMA=m
Patch was verified not to change the binary by diffing the
generated object code before and after applying the patch.
Patch is against 5.1-rc6 (localversion-next is 20190429)
net/rds/ib_recv.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c
index 7055985..8946c89 100644
--- a/net/rds/ib_recv.c
+++ b/net/rds/ib_recv.c
@@ -772,7 +772,7 @@ static void rds_ib_cong_recv(struct rds_connection *conn,
unsigned long frag_off;
unsigned long to_copy;
unsigned long copied;
- uint64_t uncongested = 0;
+ __le64 uncongested = 0;
void *addr;
/* catch completely corrupt packets */
@@ -789,7 +789,7 @@ static void rds_ib_cong_recv(struct rds_connection *conn,
copied = 0;
while (copied < RDS_CONG_MAP_BYTES) {
- uint64_t *src, *dst;
+ __le64 *src, *dst;
unsigned int k;
to_copy = min(RDS_FRAG_SIZE - frag_off, PAGE_SIZE - map_off);
@@ -824,9 +824,7 @@ static void rds_ib_cong_recv(struct rds_connection *conn,
}
/* the congestion map is in little endian order */
- uncongested = le64_to_cpu(uncongested);
-
- rds_cong_map_updated(map, uncongested);
+ rds_cong_map_updated(map, le64_to_cpu(uncongested));
}
static void rds_ib_process_recv(struct rds_connection *conn,
--
2.1.4
On 4/29/19 8:12 PM, Nicholas Mc Guire wrote:
> While the endiannes is being handled correctly as indicated by the comment
> above the offending line - sparse was unhappy with the missing annotation
> as be64_to_cpu() expects a __be64 argument. To mitigate this annotation
> all involved variables are changed to a consistent __le64 and the
> conversion to uint64_t delayed to the call to rds_cong_map_updated().
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
> ---
>
> Problem located by an experimental coccinelle script to locate
> patters that make sparse unhappy (false positives):
> net/rds/ib_recv.c:827:23: warning: cast to restricted __le64
>
> V2: Edward Cree <[email protected]> rejected the need for using __force
> here - instead solve the sparse issue by updating all of the involved
> variables - which results in an identical binary as well without using
> the __force "solution" to the sparse warning. Thanks !
>
> Patch was compile-tested with: x86_64_defconfig + INFINIBAND=m, RDS_RDMA=m
>
> Patch was verified not to change the binary by diffing the
> generated object code before and after applying the patch.
>
Thanks. I was worried about this macro magic o.w
Acked-by: Santosh Shilimkar <[email protected]>
The patch looks good, but the force in the subject sounds weird now.
On Tue, Apr 30, 2019 at 04:43:21AM -0700, Christoph Hellwig wrote:
> The patch looks good, but the force in the subject sounds weird now.
True - but last time I renamed the subject when seindg a V2 I
wsa told not to do that - the rational was if I rename it and
call it V2 -then there is no V1...
If thats nonsense - then I'll resend with a new commit subject
thx!
hofrat
From: Nicholas Mc Guire <[email protected]>
Date: Tue, 30 Apr 2019 05:12:57 +0200
> While the endiannes is being handled correctly as indicated by the comment
> above the offending line - sparse was unhappy with the missing annotation
> as be64_to_cpu() expects a __be64 argument. To mitigate this annotation
> all involved variables are changed to a consistent __le64 and the
> conversion to uint64_t delayed to the call to rds_cong_map_updated().
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
Applied.