From: Anshuman Khandual <khandual@linux.vnet.ibm.com>
To: Rashmica Gupta <rashmicy@gmail.com>, linuxppc-dev@lists.ozlabs.org
Cc: mikey@neuling.org
Subject: Re: [PATCH] selftests/powerpc: Add test to check if TAR is corrupted
Date: Wed, 02 Dec 2015 10:42:38 +0530 [thread overview]
Message-ID: <565E7DC6.4080306@linux.vnet.ibm.com> (raw)
In-Reply-To: <1449026569-30094-1-git-send-email-rashmicy@gmail.com>
On 12/02/2015 08:52 AM, Rashmica Gupta wrote:
> If the transaction is aborted, the TAR should be rolled back to the
> checkpointed value before the transaction began.
And if it does not abort, it should continue to hold the changed
(inside the transaction) TAR value after the transaction finishes.
The test here does check for both these cases and hence the commit
message should reflect that as well. This test can also accommodate
checks for PPR and DSCR register in one go.
In the TM ptrace patch series (https://lkml.org/lkml/2015/10/8/522)
which is still under review, there are couple of tests related
to (TAR, PPR, DSCR) registers with (TM running & TM suspended) and
without TM. There, the values are being checked using ptrace interface
instead. Can you please check them and just make sure that we are not
duplicating anything here.
>
> Signed-off-by: Rashmica Gupta <rashmicy@gmail.com>
> ---
> To check this yourself, undo the changes from the patch "powerpc/tm: Fix
> context switching TAR, PPR and DSCR SPRs".
>
> tools/testing/selftests/powerpc/tm/.gitignore | 1 +
> tools/testing/selftests/powerpc/tm/Makefile | 2 +-
> tools/testing/selftests/powerpc/tm/tm-tar.c | 90 +++++++++++++++++++++++++++
> 3 files changed, 92 insertions(+), 1 deletion(-)
> create mode 100644 tools/testing/selftests/powerpc/tm/tm-tar.c
>
> diff --git a/tools/testing/selftests/powerpc/tm/.gitignore b/tools/testing/selftests/powerpc/tm/.gitignore
> index 76eae258feeb..9bf6749550e3 100644
> --- a/tools/testing/selftests/powerpc/tm/.gitignore
> +++ b/tools/testing/selftests/powerpc/tm/.gitignore
> @@ -4,3 +4,4 @@ tm-signal-msr-resv
> tm-signal-stack
> tm-fork
> tm-dscr
> +tm-tar
> diff --git a/tools/testing/selftests/powerpc/tm/Makefile b/tools/testing/selftests/powerpc/tm/Makefile
> index 59eec240339d..ab2992f69b4c 100644
> --- a/tools/testing/selftests/powerpc/tm/Makefile
> +++ b/tools/testing/selftests/powerpc/tm/Makefile
> @@ -1,4 +1,4 @@
> -TEST_PROGS := tm-resched-dscr tm-syscall tm-signal-msr-resv tm-signal-stack tm-fork tm-dscr
> +TEST_PROGS := tm-resched-dscr tm-syscall tm-signal-msr-resv tm-signal-stack tm-fork tm-dscr tm-tar
>
> all: $(TEST_PROGS)
>
> diff --git a/tools/testing/selftests/powerpc/tm/tm-tar.c b/tools/testing/selftests/powerpc/tm/tm-tar.c
> new file mode 100644
> index 000000000000..22a0c3485412
> --- /dev/null
> +++ b/tools/testing/selftests/powerpc/tm/tm-tar.c
> @@ -0,0 +1,90 @@
> +/*
> + * Copyright 2015, Michael Neuling, IBM Corp.
> + * Licensed under GPLv2.
> + * Original: Michael Neuling 19/7/2013
> + * Edited: Rashmica Gupta 01/12/2015
> + *
> + * Do some transactions, see if the tar is corrupted.
It can use a better explanation here.
> + *
> + */
> +
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <unistd.h>
> +#include <string.h>
> +
> +#include "tm.h"
> +#include "utils.h"
> +
> +#define SPRN_TAR 0x32f
> +
> +int num_loops = 10000;
> +
> +int test_tar(void)
> +{
> + int i;
> +
> + SKIP_IF(!have_htm());
> +
> + for (i = 0; i < num_loops; i++)
> + {
> + uint64_t result = 0;
> + asm __volatile__(
> + "li 7, 1;"
> + "mtspr %[tar], 7;" // tar = 1
Please use /* */ kind of comment sections not //.
> + "tbegin.;"
> + "beq 3f;"
> + "li 4, 0x7000;" // Loop lots, to use time
> + "2:;" // Start loop
Here as well.
> + "li 7, 2;"
> + "mtspr %[tar], 7;" // tar = 2
> + "tsuspend.;"
> + "li 7, 3;"
> + "mtspr %[tar], 7;" // tar = 3
> + "tresume.;"
> + "subi 4, 4, 1;"
> + "cmpdi 4, 0;"
> + "bne 2b;"
> + "tend.;"
> +
> + // Transaction sucess! TAR should be 3.
> + "mfspr 7, %[tar];"
> + "ori %[res], 7, 4;" // res = 3|4 = 7
> + "b 4f;"
> +
> + // Abort handler. TAR should be rolled back to 1.
> + "3:;"
> + "mfspr 7, %[tar];"
> + "ori %[res], 7, 8;" // res = 1|8 = 9
> + "4:;"
> +
> + : [res]"=r"(result)
> + : [tar]"i"(SPRN_TAR)
> + : "memory", "r0", "r4", "r7");
> +
> + // If result is anything else other than 7 or 9, the tar
> + // value must have been corrupted.
Here as well.
> + if ((result != 7) && (result != 9))
> + return 1;
> +
> + }
> + return 0;
> +}
> +
> +int main(int argc, char *argv[])
> +
> + // A low number of iterations (eg 100) can cause a false pass.
Here as well.
> + if (argc > 1) {
> + if (strcmp(argv[1], "-h") == 0) {
> + printf("Syntax:\n\t%s [<num loops>]\n",
> + argv[0]);
> + return 0;
> + } else {
> + num_loops = atoi(argv[1]);
> + }
> + }
> +
> + printf("Starting, %d loops\n", num_loops);
> +
> + test_harness(test_tar, "tm_tar");
> +}
>
next prev parent reply other threads:[~2015-12-02 5:12 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-02 3:22 [PATCH] selftests/powerpc: Add test to check if TAR is corrupted Rashmica Gupta
2015-12-02 5:12 ` Anshuman Khandual [this message]
2015-12-02 22:29 ` Michael Neuling
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=565E7DC6.4080306@linux.vnet.ibm.com \
--to=khandual@linux.vnet.ibm.com \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=mikey@neuling.org \
--cc=rashmicy@gmail.com \
/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).