From: Jordan Niethe <jniethe5@gmail.com>
To: linuxppc-dev <linuxppc-dev@lists.ozlabs.org>
Cc: Michael Neuling <mikey@neuling.org>
Subject: Re: [PATCH 2/2] selftests: Skip TM tests on synthetic TM implementations
Date: Tue, 15 Jun 2021 16:06:02 +1000 [thread overview]
Message-ID: <CACzsE9oV9s-5E-_ATwj6_E0dvo+nO=wq8YPFsPV983ACtWT8BA@mail.gmail.com> (raw)
In-Reply-To: <20210607233709.941102-2-jniethe5@gmail.com>
On Tue, Jun 8, 2021 at 9:37 AM Jordan Niethe <jniethe5@gmail.com> wrote:
>
> Transactional Memory was removed from the architecture in ISA v3.1. For
> threads running in P8/P9 compatibility mode on P10 a synthetic TM
> implementation is provided. In this implementation, tbegin. always sets
> cr0 eq meaning the abort handler is always called. This is not an issue
> as users of TM are expected to have a fallback non transactional way to
> make forward progress in the abort handler.
>
> As the TM self tests exist only to test TM, no alternative path forward
> is provided, leading to them timing out and failing on the synthetic TM
> implementation.
>
> The TEXASR indicates if a transaction failure is due to a synthetic
> implementation. Check for a synthetic implementation and skip the TM
> tests if so.
>
> Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
> ---
> .../selftests/powerpc/ptrace/ptrace-tm-gpr.c | 1 +
> .../powerpc/ptrace/ptrace-tm-spd-gpr.c | 1 +
> .../powerpc/ptrace/ptrace-tm-spd-tar.c | 1 +
> .../selftests/powerpc/ptrace/ptrace-tm-tar.c | 1 +
> .../selftests/powerpc/tm/tm-resched-dscr.c | 1 +
> .../selftests/powerpc/tm/tm-signal-stack.c | 1 +
> .../testing/selftests/powerpc/tm/tm-syscall.c | 2 +-
> tools/testing/selftests/powerpc/tm/tm.h | 36 +++++++++++++++++++
> 8 files changed, 43 insertions(+), 1 deletion(-)
>
> diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c
> index 7df7100a29be..67ca297c5cca 100644
> --- a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c
> +++ b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c
> @@ -113,6 +113,7 @@ int ptrace_tm_gpr(void)
> int ret, status;
>
> SKIP_IF(!have_htm());
> + SKIP_IF(htm_is_synthetic());
> shm_id = shmget(IPC_PRIVATE, sizeof(int) * 2, 0777|IPC_CREAT);
> pid = fork();
> if (pid < 0) {
> diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c
> index 8706bea5d015..6f2bce1b6c5d 100644
> --- a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c
> +++ b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c
> @@ -119,6 +119,7 @@ int ptrace_tm_spd_gpr(void)
> int ret, status;
>
> SKIP_IF(!have_htm());
> + SKIP_IF(htm_is_synthetic());
> shm_id = shmget(IPC_PRIVATE, sizeof(int) * 3, 0777|IPC_CREAT);
> pid = fork();
> if (pid < 0) {
> diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-tar.c b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-tar.c
> index 2ecfa1158e2b..e112a34fbe59 100644
> --- a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-tar.c
> +++ b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-tar.c
> @@ -129,6 +129,7 @@ int ptrace_tm_spd_tar(void)
> int ret, status;
>
> SKIP_IF(!have_htm());
> + SKIP_IF(htm_is_synthetic());
> shm_id = shmget(IPC_PRIVATE, sizeof(int) * 3, 0777|IPC_CREAT);
> pid = fork();
> if (pid == 0)
> diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-tar.c b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-tar.c
> index 46ef378a15ec..d0db6df0f0ea 100644
> --- a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-tar.c
> +++ b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-tar.c
> @@ -117,6 +117,7 @@ int ptrace_tm_tar(void)
> int ret, status;
>
> SKIP_IF(!have_htm());
> + SKIP_IF(htm_is_synthetic());
> shm_id = shmget(IPC_PRIVATE, sizeof(int) * 2, 0777|IPC_CREAT);
> pid = fork();
> if (pid == 0)
> diff --git a/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c b/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
> index 4cdb83964bb3..85c940ae6ff8 100644
> --- a/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
> +++ b/tools/testing/selftests/powerpc/tm/tm-resched-dscr.c
> @@ -40,6 +40,7 @@ int test_body(void)
> uint64_t rv, dscr1 = 1, dscr2, texasr;
>
> SKIP_IF(!have_htm());
> + SKIP_IF(htm_is_synthetic());
>
> printf("Check DSCR TM context switch: ");
> fflush(stdout);
> diff --git a/tools/testing/selftests/powerpc/tm/tm-signal-stack.c b/tools/testing/selftests/powerpc/tm/tm-signal-stack.c
> index cdcf8c5bbbc7..68807aac8dd3 100644
> --- a/tools/testing/selftests/powerpc/tm/tm-signal-stack.c
> +++ b/tools/testing/selftests/powerpc/tm/tm-signal-stack.c
> @@ -35,6 +35,7 @@ int tm_signal_stack()
> int pid;
>
> SKIP_IF(!have_htm());
> + SKIP_IF(htm_is_synthetic());
>
> pid = fork();
> if (pid < 0)
> diff --git a/tools/testing/selftests/powerpc/tm/tm-syscall.c b/tools/testing/selftests/powerpc/tm/tm-syscall.c
> index becb8207b432..467a6b3134b2 100644
> --- a/tools/testing/selftests/powerpc/tm/tm-syscall.c
> +++ b/tools/testing/selftests/powerpc/tm/tm-syscall.c
> @@ -25,7 +25,6 @@ extern int getppid_tm_suspended(void);
> unsigned retries = 0;
>
> #define TEST_DURATION 10 /* seconds */
> -#define TM_RETRIES 100
>
> pid_t getppid_tm(bool suspend)
> {
> @@ -67,6 +66,7 @@ int tm_syscall(void)
> struct timeval end, now;
>
> SKIP_IF(!have_htm_nosc());
> + SKIP_IF(htm_is_synthetic());
>
> setbuf(stdout, NULL);
>
> diff --git a/tools/testing/selftests/powerpc/tm/tm.h b/tools/testing/selftests/powerpc/tm/tm.h
> index c5a1e5c163fc..c03c6e778876 100644
> --- a/tools/testing/selftests/powerpc/tm/tm.h
> +++ b/tools/testing/selftests/powerpc/tm/tm.h
> @@ -10,6 +10,9 @@
> #include <asm/tm.h>
>
> #include "utils.h"
> +#include "reg.h"
> +
> +#define TM_RETRIES 100
>
> static inline bool have_htm(void)
> {
> @@ -31,6 +34,39 @@ static inline bool have_htm_nosc(void)
> #endif
> }
>
> +/*
> + * Transactional Memory was removed in ISA 3.1. A synthetic TM implementation
> + * is provided on P10 for threads running in P8/P9 compatibility mode. The
> + * synthetic implementation immediately fails after tbegin. This failure sets
> + * Bit 7 (Failure Persistent) and Bit 15 (Implementation-specific).
> + */
> +static inline bool htm_is_synthetic(void)
> +{
> + int i;
> +
> + /*
> + * Per the ISA, the Failure Persistent bit may be incorrect. Try a few
> + * times in case we got an Implementation-specific failure on a non ISA
> + * v3.1 system. On these systems the Implementation-specific failure
> + * should not be persistent.
> + */
> + for (i = 0; i < TM_RETRIES; i++) {
> + asm volatile(
> + "tbegin.;"
> + "beq 1f;"
> + "tend.;"
> + "1:"
> + :
> + :
> + : "memory");
> +
> + if ((__builtin_get_texasr() & (TEXASR_FP | TEXASR_IC)) !=
> + (TEXASR_FP | TEXASR_IC))
> + break;
> + }
> + return i == TM_RETRIES;
> +}
> +
> static inline long failure_code(void)
> {
> return __builtin_get_texasru() >> 24;
> --
> 2.25.1
>
There's a couple more tests that need the same treatment, Will send a
new version.
prev parent reply other threads:[~2021-06-15 6:06 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-07 23:37 [PATCH 1/2] selftests/powerpc: Add missing clobbered register to to ptrace TM tests Jordan Niethe
2021-06-07 23:37 ` [PATCH 2/2] selftests: Skip TM tests on synthetic TM implementations Jordan Niethe
2021-06-15 6:06 ` Jordan Niethe [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CACzsE9oV9s-5E-_ATwj6_E0dvo+nO=wq8YPFsPV983ACtWT8BA@mail.gmail.com' \
--to=jniethe5@gmail.com \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=mikey@neuling.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).