| Message ID | 2307631.X1DYNXxKPZ@wuerfel |
|---|---|
| State | New |
| Headers | show |
On 02/01/2016 11:34 AM, Arnd Bergmann wrote: > The amso1100 driver prints a phys_addr_t by casting it to a pointer, > which causes a warning when phys_addr_t is 64-bit and pointers > are 32-bit: > > drivers/staging/rdma/amso1100/c2.c: In function 'c2_rx_error': > drivers/staging/rdma/amso1100/c2.c:430:4: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] > > This changes the format string to print the address as a 64-bit > number all the time to avoids the warnings. Ideally we would use > the %pap format string for phys_addr_t, but that doesn't work > here easily as it requires passing a pointer to the address. > > Signed-off-by: Arnd Bergmann <arnd@arndb.de> > --- > Found on ARM randconfig testing. This must have been there forever but is > rather hard to trigger even with randconfig builds. > > diff --git a/drivers/staging/rdma/amso1100/c2.c b/drivers/staging/rdma/amso1100/c2.c > index b46ebd1ae15a..170a42a60552 100644 > --- a/drivers/staging/rdma/amso1100/c2.c > +++ b/drivers/staging/rdma/amso1100/c2.c > @@ -426,8 +426,8 @@ static void c2_rx_error(struct c2_port *c2_port, struct c2_element *elem) > pr_debug(" index : %Zu\n", > elem - c2_port->rx_ring.start); > pr_debug(" len : %u\n", rx_desc->len); > - pr_debug(" rxp_hdr : %p [PA %p]\n", rxp_hdr, > - (void *) __pa((unsigned long) rxp_hdr)); > + pr_debug(" rxp_hdr : %p [PA %016llx\n", rxp_hdr, > + (u64)__pa((unsigned long) rxp_hdr)); > pr_debug(" flags : 0x%x\n", rxp_hdr->flags); > pr_debug(" status: 0x%x\n", rxp_hdr->status); > pr_debug(" len : %u\n", rxp_hdr->len); > I've sent off a deletion of this file to Linus already, sorry. -- Doug Ledford <dledford@redhat.com> GPG KeyID: 0E572FDD
On Thursday 04 February 2016 07:15:05 Doug Ledford wrote: > I've sent off a deletion of this file to Linus already, sorry. > Even better, thanks! Arnd
diff --git a/drivers/staging/rdma/amso1100/c2.c b/drivers/staging/rdma/amso1100/c2.c index b46ebd1ae15a..170a42a60552 100644 --- a/drivers/staging/rdma/amso1100/c2.c +++ b/drivers/staging/rdma/amso1100/c2.c @@ -426,8 +426,8 @@ static void c2_rx_error(struct c2_port *c2_port, struct c2_element *elem) pr_debug(" index : %Zu\n", elem - c2_port->rx_ring.start); pr_debug(" len : %u\n", rx_desc->len); - pr_debug(" rxp_hdr : %p [PA %p]\n", rxp_hdr, - (void *) __pa((unsigned long) rxp_hdr)); + pr_debug(" rxp_hdr : %p [PA %016llx\n", rxp_hdr, + (u64)__pa((unsigned long) rxp_hdr)); pr_debug(" flags : 0x%x\n", rxp_hdr->flags); pr_debug(" status: 0x%x\n", rxp_hdr->status); pr_debug(" len : %u\n", rxp_hdr->len);
The amso1100 driver prints a phys_addr_t by casting it to a pointer, which causes a warning when phys_addr_t is 64-bit and pointers are 32-bit: drivers/staging/rdma/amso1100/c2.c: In function 'c2_rx_error': drivers/staging/rdma/amso1100/c2.c:430:4: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast] This changes the format string to print the address as a 64-bit number all the time to avoids the warnings. Ideally we would use the %pap format string for phys_addr_t, but that doesn't work here easily as it requires passing a pointer to the address. Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- Found on ARM randconfig testing. This must have been there forever but is rather hard to trigger even with randconfig builds.