Commit Message
Previously VQ was of type long. Using uint64_t ensures it always matches the
same type as the VG register.
Note that in the Linux kernel, VQ is 16bits. We cast it up to 64bits
immediately after reading to ensure we always use the same type throughout
the code.
This spatch was spun off Pedro's review here:
https://sourceware.org/ml/gdb-patches/2018-05/msg00862.html
Tested with other SVE patches under aarch64 SVE emulation.
2018-06-01 Alan Hayward <alan.hayward@arm.com>
* aarch64-tdep.c (aarch64_read_description): Use long for VQ.
* aarch64-tdep.h (aarch64_read_description): Likewise.
* arch/aarch64.c (aarch64_create_target_description): Likewise.
* arch/aarch64.h (aarch64_create_target_description): Likewise.
* features/aarch64-sve.c (create_feature_aarch64_sve): Likewise.
* nat/aarch64-sve-linux-ptrace.c(aarch64_sve_get_vq): Likewise.
* nat/aarch64-sve-linux-ptrace.h (aarch64_sve_get_vq): Likewise.
---
gdb/aarch64-tdep.c | 2 +-
gdb/aarch64-tdep.h | 2 +-
gdb/arch/aarch64.c | 2 +-
gdb/arch/aarch64.h | 2 +-
gdb/features/aarch64-sve.c | 2 +-
gdb/nat/aarch64-sve-linux-ptrace.c | 2 +-
gdb/nat/aarch64-sve-linux-ptrace.h | 2 +-
7 files changed, 7 insertions(+), 7 deletions(-)
Comments
On 06/01/2018 06:53 PM, Alan Hayward wrote:
> 2018-06-01 Alan Hayward <alan.hayward@arm.com>
>
> * aarch64-tdep.c (aarch64_read_description): Use long for VQ.
s/long/uint64_t/.
> --- a/gdb/aarch64-tdep.c
> +++ b/gdb/aarch64-tdep.c
> @@ -2856,7 +2856,7 @@ aarch64_displaced_step_hw_singlestep (struct gdbarch *gdbarch,
> (It is not possible to set VQ to zero on an SVE system). */
>
> const target_desc *
> -aarch64_read_description (long vq)
> +aarch64_read_description (uint64_t vq)
> {
> if (vq > AARCH64_MAX_SVE_VQ)
> error (_("VQ is %ld, maximum supported value is %d"), vq,
This %ld needs to be updated to either use PRIu64 or pulongest.
There may be others, I haven't looked at the code, just the patch.
Otherwise OK.
Thanks,
Pedro Alves
> On 2 Jun 2018, at 11:39, Pedro Alves <palves@redhat.com> wrote:
>
> On 06/01/2018 06:53 PM, Alan Hayward wrote:
>
>> 2018-06-01 Alan Hayward <alan.hayward@arm.com>
>>
>> * aarch64-tdep.c (aarch64_read_description): Use long for VQ.
>
> s/long/uint64_t/.
Oops.
>
>> --- a/gdb/aarch64-tdep.c
>> +++ b/gdb/aarch64-tdep.c
>> @@ -2856,7 +2856,7 @@ aarch64_displaced_step_hw_singlestep (struct gdbarch *gdbarch,
>> (It is not possible to set VQ to zero on an SVE system). */
>>
>> const target_desc *
>> -aarch64_read_description (long vq)
>> +aarch64_read_description (uint64_t vq)
>> {
>> if (vq > AARCH64_MAX_SVE_VQ)
>> error (_("VQ is %ld, maximum supported value is %d"), vq,
>
> This %ld needs to be updated to either use PRIu64 or pulongest.
> There may be others, I haven't looked at the code, just the patch.
I’m surprised this didn’t cause a warning build failure.
Fixed up with:
error (_("VQ is %" PRIu64 ", maximum supported value is %d"), vq,
Double checked that there were no other prints of VQ/VG etc that
were added in my committed SVE patches.
> Otherwise OK.
>
Pushed with the changes above.
Thanks!
Alan.
@@ -2856,7 +2856,7 @@ aarch64_displaced_step_hw_singlestep (struct gdbarch *gdbarch,
(It is not possible to set VQ to zero on an SVE system). */
const target_desc *
-aarch64_read_description (long vq)
+aarch64_read_description (uint64_t vq)
{
if (vq > AARCH64_MAX_SVE_VQ)
error (_("VQ is %ld, maximum supported value is %d"), vq,
@@ -75,7 +75,7 @@ struct gdbarch_tdep
int (*aarch64_syscall_record) (struct regcache *regcache, unsigned long svc_number);
};
-const target_desc *aarch64_read_description (long vq);
+const target_desc *aarch64_read_description (uint64_t vq);
extern int aarch64_process_record (struct gdbarch *gdbarch,
struct regcache *regcache, CORE_ADDR addr);
@@ -26,7 +26,7 @@
/* See arch/aarch64.h. */
target_desc *
-aarch64_create_target_description (long vq)
+aarch64_create_target_description (uint64_t vq)
{
target_desc *tdesc = allocate_target_description ();
@@ -26,7 +26,7 @@
the presence of SVE and the Vector Quotient - the number of 128bit chunks in
an SVE Z register. */
-target_desc *aarch64_create_target_description (long vq);
+target_desc *aarch64_create_target_description (uint64_t vq);
/* Register numbers of various important registers.
Note that on SVE, the Z registers reuse the V register numbers and the V
@@ -23,7 +23,7 @@
static int
create_feature_aarch64_sve (struct target_desc *result, long regnum,
- int scale)
+ uint64_t scale)
{
struct tdesc_feature *feature;
tdesc_type *element_type, *field_type;
@@ -27,7 +27,7 @@
/* See nat/aarch64-sve-linux-ptrace.h. */
-unsigned long
+uint64_t
aarch64_sve_get_vq (int tid)
{
struct iovec iovec;
@@ -32,7 +32,7 @@
/* Read VQ for the given tid using ptrace. If SVE is not supported then zero
is returned (on a system that supports SVE, then VQ cannot be zero). */
-extern unsigned long aarch64_sve_get_vq (int tid);
+uint64_t aarch64_sve_get_vq (int tid);
/* Structures and defines taken from sigcontext.h. */