eal/x86: fix segfaults in waitpkg power intrinsics
Checks
Commit Message
From: David Hunt <david.hunt@intel.com>
The code was recently enhanced to allow the use of the waitpkg
intrinsics rather than the raw assembly in the rte_power functions.
However, the parameters to the intrinsics, while compiling fine, were
incorrect, and would segfault when run on the appropriate hardware.
This patch fixes the intrinsic parameters. Tested on a system with
tpause and umonitor/umwait instructions.
Fixes: 60943c04f3bc ("eal/x86: use intrinsics for power management")
Signed-off-by: David Hunt <david.hunt@intel.com>
Reviewed-by: Bruce Richardson <bruce.richardson@intel.com>
---
lib/eal/x86/rte_power_intrinsics.c | 14 +++++++-------
1 file changed, 7 insertions(+), 7 deletions(-)
Comments
On Tue, Nov 07, 2023 at 04:19:01PM +0000, Bruce Richardson wrote:
> From: David Hunt <david.hunt@intel.com>
>
> The code was recently enhanced to allow the use of the waitpkg
> intrinsics rather than the raw assembly in the rte_power functions.
> However, the parameters to the intrinsics, while compiling fine, were
> incorrect, and would segfault when run on the appropriate hardware.
> This patch fixes the intrinsic parameters. Tested on a system with
> tpause and umonitor/umwait instructions.
>
> Fixes: 60943c04f3bc ("eal/x86: use intrinsics for power management")
>
> Signed-off-by: David Hunt <david.hunt@intel.com>
> Reviewed-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
Reviewed-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
08/11/2023 04:19, Tyler Retzlaff:
> On Tue, Nov 07, 2023 at 04:19:01PM +0000, Bruce Richardson wrote:
> > From: David Hunt <david.hunt@intel.com>
> >
> > The code was recently enhanced to allow the use of the waitpkg
> > intrinsics rather than the raw assembly in the rte_power functions.
> > However, the parameters to the intrinsics, while compiling fine, were
> > incorrect, and would segfault when run on the appropriate hardware.
> > This patch fixes the intrinsic parameters. Tested on a system with
> > tpause and umonitor/umwait instructions.
> >
> > Fixes: 60943c04f3bc ("eal/x86: use intrinsics for power management")
> >
> > Signed-off-by: David Hunt <david.hunt@intel.com>
> > Reviewed-by: Bruce Richardson <bruce.richardson@intel.com>
>
> Reviewed-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
Applied, thanks.
@@ -40,12 +40,12 @@ static void intel_umonitor(volatile void *addr)
static void intel_umwait(const uint64_t timeout)
{
+#if defined(RTE_TOOLCHAIN_MSVC) || defined(__WAITPKG__)
+ _umwait(0, timeout);
+#else
const uint32_t tsc_l = (uint32_t)timeout;
const uint32_t tsc_h = (uint32_t)(timeout >> 32);
-#if defined(RTE_TOOLCHAIN_MSVC) || defined(__WAITPKG__)
- _umwait(tsc_l, tsc_h);
-#else
asm volatile(".byte 0xf2, 0x0f, 0xae, 0xf7;"
: /* ignore rflags */
: "D"(0), /* enter C0.2 */
@@ -208,17 +208,17 @@ rte_power_monitor(const struct rte_power_monitor_cond *pmc,
int
rte_power_pause(const uint64_t tsc_timestamp)
{
- const uint32_t tsc_l = (uint32_t)tsc_timestamp;
- const uint32_t tsc_h = (uint32_t)(tsc_timestamp >> 32);
-
/* prevent user from running this instruction if it's not supported */
if (!wait_supported)
return -ENOTSUP;
/* execute TPAUSE */
#if defined(RTE_TOOLCHAIN_MSVC) || defined(__WAITPKG__)
- _tpause(tsc_l, tsc_h);
+ _tpause(0, tsc_timestamp);
#else
+ const uint32_t tsc_l = (uint32_t)tsc_timestamp;
+ const uint32_t tsc_h = (uint32_t)(tsc_timestamp >> 32);
+
asm volatile(".byte 0x66, 0x0f, 0xae, 0xf7;"
: /* ignore rflags */
: "D"(0), /* enter C0.2 */