IB/mlx5: Fix linkage failure on 32-bit arches
Subsystems (2)
PATCH 1/1 IB/mlx5: Fix linkage failure on 32-bit arches
Commit f164be8c0366 ("IB/mlx5: Extend caps stage to handle VAR capabilities") introduced a straight "/" division of the u64 variable "bar_size", which emits an __udivdi3() libgcc call on 32-bit arches and certain GCC versions:
error: "__udivdi3" [drivers/infiniband/hw/mlx5/mlx5_ib.ko] undefined! [1]
Replace it with the corresponding div_u64() call. Compile-tested on ARCH=mips 32r2el_defconfig BOARDS=ocelot.
capabilities")
--- drivers/infiniband/hw/mlx5/main.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/infiniband/hw/mlx5/main.c +2 -1
@@ -6545,7 +6545,8 @@ static int mlx5_ib_init_var_table(struct mlx5_ib_dev *dev) doorbell_bar_offset); bar_size = (1ULL << log_doorbell_bar_size) *4096 ; var_table->stride_size =1ULL << log_doorbell_stride; - var_table->num_var_hw_entries = bar_size / var_table->stride_size; + var_table->num_var_hw_entries =div_u64 (bar_size, + var_table->stride_size);mutex_init (&var_table->bitmap_lock); var_table->bitmap =bitmap_zalloc (var_table->num_var_hw_entries,GFP_KERNEL );
Signature · 1 line
2.25.0
Quoted 10 lines · Alexander Lobakin
Commit
f164be8c0366("IB/mlx5: Extend caps stage to handle VAR capabilities") introduced a straight "/" division of the u64 variable "bar_size", which emits an__udivdi3()libgcc call on 32-bit arches and certain GCC versions:error: "
__udivdi3" [drivers/infiniband/hw/mlx5/mlx5_ib.ko] undefined! [1]Replace it with the corresponding
div_u64()call. Compile-tested on ARCH=mips 32r2el_defconfig BOARDS=ocelot.
Quoted 8 lines
Fixes f164be8c0366 ("IB/mlx5: Extend caps stage to handle VARcapabilities")
Signed-off-by Alexander Lobakin <[email protected]>--- drivers/infiniband/hw/mlx5/main.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
Randy beat you too it..
https://lore.kernel.org/linux-rdma/[email protected]/
But it seems patchwork missed this somehow.
Applied now at least
Jason
Jason Gunthorpe wrote 14.02.2020 22:24:
Quoted 20 lines · Alexander Lobakin
Quoted 10 lines
Commit
f164be8c0366("IB/mlx5: Extend caps stage to handle VAR capabilities") introduced a straight "/" division of the u64 variable "bar_size", which emits an__udivdi3()libgcc call on 32-bit arches and certain GCC versions:error: "
__udivdi3" [drivers/infiniband/hw/mlx5/mlx5_ib.ko] undefined! [1]Replace it with the corresponding
div_u64()call. Compile-tested on ARCH=mips 32r2el_defconfig BOARDS=ocelot.Quoted 8 lines
Fixes f164be8c0366 ("IB/mlx5: Extend caps stage to handle VARcapabilities")
Signed-off-by Alexander Lobakin <[email protected]>--- drivers/infiniband/hw/mlx5/main.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
Quoted 3 lines
Randy beat you too it..
Ah, OK. Sorry for missing this one. I didn't see any fix over git.kernel.org and thought it doesn't exist yet.
Quoted 3 lines
But it seems patchwork missed this somehow.
Applied now at least
Thanks!
Quoted 1 line
Jason
Regards,
ᚷ ᛖ ᚢ ᚦ ᚠ ᚱ
Quoted 22 lines · Alexander Lobakin
Jason Gunthorpe wrote 14.02.2020 22:24:
Quoted 21 lines · Alexander Lobakin
Quoted 19 lines
Commit
f164be8c0366("IB/mlx5: Extend caps stage to handle VAR capabilities") introduced a straight "/" division of the u64 variable "bar_size", which emits an__udivdi3()libgcc call on 32-bit arches and certain GCC versions:error: "
__udivdi3" [drivers/infiniband/hw/mlx5/mlx5_ib.ko] undefined! [1]Replace it with the corresponding
div_u64()call. Compile-tested on ARCH=mips 32r2el_defconfig BOARDS=ocelot.Fixes f164be8c0366 ("IB/mlx5: Extend caps stage to handle VARcapabilities")
Signed-off-by Alexander Lobakin <[email protected]>---
drivers/infiniband/hw/mlx5/main.c | 3 ++-1 file changed, 2 insertions(+), 1 deletion(-)
Quoted 10 lines
Quoted 3 lines
Randy beat you too it..
Ah, OK. Sorry for missing this one. I didn't see any fix over git.kernel.org and thought it doesn't exist yet.
Quoted 3 lines
But it seems patchwork missed this somehow.
Applied now at least
Jason, I think Alexander's patch is more correct. It uses div_u64, while yours uses div64_u64. The divider is 32-bit, so div_u64 would be more optimized on most 32-bit architectures.
Quoted 6 lines
Thanks!
Quoted 1 line
Jason
Regards,
ᚷ ᛖ ᚢ ᚦ ᚠ ᚱ
Quoted 30 lines · Maxim Mikityanskiy
Quoted 29 lines · Alexander Lobakin
Jason Gunthorpe wrote 14.02.2020 22:24:
Quoted 21 lines · Alexander Lobakin
Quoted 11 lines
Commit
f164be8c0366("IB/mlx5: Extend caps stage to handle VAR capabilities") introduced a straight "/" division of the u64 variable "bar_size", which emits an__udivdi3()libgcc call on 32-bit arches and certain GCC versions:error: "
__udivdi3" [drivers/infiniband/hw/mlx5/mlx5_ib.ko] undefined! [1]Replace it with the corresponding
div_u64()call. Compile-tested on ARCH=mips 32r2el_defconfig BOARDS=ocelot.Quoted 8 lines
Fixes f164be8c0366 ("IB/mlx5: Extend caps stage to handle VARcapabilities")
Signed-off-by Alexander Lobakin <[email protected]>drivers/infiniband/hw/mlx5/main.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)Quoted 3 lines
Randy beat you too it..
Ah, OK. Sorry for missing this one. I didn't see any fix over git.kernel.org and thought it doesn't exist yet.
Quoted 7 lines
Quoted 3 lines
Quoted 3 lines
But it seems patchwork missed this somehow.
Applied now at least
Jason, I think Alexander's patch is more correct. It uses
div_u64, while yours usesdiv64_u64. The divider is 32-bit, sodiv_u64would be more optimized on most 32-bit architectures.
Seems so, but it is already done. You can send a followup to fix it
Jason