From: Bjorn Andersson <bjorn.andersson@linaro.org> To: Benjamin Gaignard <benjamin.gaignard@linaro.org> Cc: ohad@wizery.com, robh+dt@kernel.org, mark.rutland@arm.com, alexandre.torgue@st.com, linux-remoteproc@vger.kernel.org, devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, Benjamin Gaignard <benjamin.gaignard@st.com> Subject: Re: [PATCH 5/5] hwspinlock: Add test module Date: Wed, 7 Nov 2018 23:06:04 -0800 [thread overview] Message-ID: <20181108070604.GE12063@builder> (raw) In-Reply-To: <20181031093032.20386-6-benjamin.gaignard@st.com> On Wed 31 Oct 02:30 PDT 2018, Benjamin Gaignard wrote: > Create a test module to perform simple unitary tests on hwspinlock. > It doesn't cover all the possibles cases but at least allow to test > that very basic features are working. > I like the idea of making these things testable, but I would like to hear from others how useful this module would be to them before merging it. > Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com> > --- > drivers/hwspinlock/Kconfig | 9 +++ > drivers/hwspinlock/Makefile | 1 + > drivers/hwspinlock/hwspinlock_test.c | 132 +++++++++++++++++++++++++++++++++++ > 3 files changed, 142 insertions(+) > create mode 100644 drivers/hwspinlock/hwspinlock_test.c > > diff --git a/drivers/hwspinlock/Kconfig b/drivers/hwspinlock/Kconfig > index e1a20b460590..f340ebb4d1f7 100644 > --- a/drivers/hwspinlock/Kconfig > +++ b/drivers/hwspinlock/Kconfig > @@ -68,3 +68,12 @@ config HWSPINLOCK_STM32 > Say y here to support the STM32 Hardware Spinlock device. > > If unsure, say N. > + > +config HWSPINLOCK_TEST > + tristate "hwspinlock test module" > + depends on HWSPINLOCK && m > + default n > + help > + Select M here if you want to build hwspinlock test module > + > + If unsure, say N. > diff --git a/drivers/hwspinlock/Makefile b/drivers/hwspinlock/Makefile > index c0a9505b4dcf..e6b3d0212fe0 100644 > --- a/drivers/hwspinlock/Makefile > +++ b/drivers/hwspinlock/Makefile > @@ -10,3 +10,4 @@ obj-$(CONFIG_HWSPINLOCK_SIRF) += sirf_hwspinlock.o > obj-$(CONFIG_HWSPINLOCK_SPRD) += sprd_hwspinlock.o > obj-$(CONFIG_HSEM_U8500) += u8500_hsem.o > obj-$(CONFIG_HWSPINLOCK_STM32) += stm32_hwspinlock.o > +obj-$(CONFIG_HWSPINLOCK_TEST) += hwspinlock_test.o > diff --git a/drivers/hwspinlock/hwspinlock_test.c b/drivers/hwspinlock/hwspinlock_test.c > new file mode 100644 > index 000000000000..75819337e45a > --- /dev/null > +++ b/drivers/hwspinlock/hwspinlock_test.c > @@ -0,0 +1,132 @@ > +// SPDX-License-Identifier: GPL-2.0 > +/* > + * Copyright (C) STMicroelectronics SA 2018 > + * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics. > + * License terms: GNU General Public License (GPL), version 2 > + */ > + > +#include <linux/hwspinlock.h> > +#include <linux/kernel.h> > +#include <linux/module.h> > + > +#define TIMEOUT 50 /* ms */ > + > +static void hwspin_lock_test_twice_request_specific(void) > +{ > + struct hwspinlock *lock0 = hwspin_lock_request_specific(0); > + struct hwspinlock *lock1 = hwspin_lock_request_specific(1); In a platform with more than one hwspinlock this will test the implementation that was first probed and it will assume that there's no issues with messing with the first two locks from this provider. > + > + int ret; > + > + if (!lock0 || !lock1) { > + pr_warn("Can't request twice hwspin_lock\n"); > + return; > + } > + > + ret = hwspin_trylock(lock0); > + if (ret) > + pr_warn("Can't trylock requested hwspin_lock 0 (%d)\n", ret); > + > + ret = hwspin_trylock(lock1); > + if (ret) > + pr_warn("Can't trylock requested hwspin_lock 1 (%d)\n", ret); > + > + hwspin_unlock(lock0); > + hwspin_unlock(lock1); As described in the kerneldoc for these it's a bug to attempt to unlock a unlocked lock. I believe this should be interpreted as it being invalid to unlock unless hwspin_trylock() returned successfully. > + > + ret = hwspin_lock_free(lock0); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > + > + ret = hwspin_lock_free(lock1); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > +} > + > +static void hwspin_lock_test_trylock(void) > +{ > + struct hwspinlock *lock = hwspin_lock_request(); > + int ret; > + > + if (!lock) { > + pr_warn("Can't request hwspin_lock\n"); > + return; > + } > + > + ret = hwspin_trylock(lock); > + if (ret) > + pr_warn("Can't trylock requested hwspin_lock (%d)\n", ret); > + > + /* Try to lock a second time, this should failed */ > + ret = hwspin_trylock(lock); > + if (ret != -EBUSY) > + pr_warn("Getting twice the same lock ! (%d)\n", ret); This shouldn't be a warning, as this is a test failure. > + > + hwspin_unlock(lock); > + ret = hwspin_lock_free(lock); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > +} > + > +static void hwspin_lock_test_request_specific(void) > +{ > + /* > + * Let's assume that any hwspin_lock driver would at least have one > + * lock at index 0 > + */ > + struct hwspinlock *lock = hwspin_lock_request_specific(0); > + int ret; > + > + if (!lock) { > + pr_warn("Can't request hwspin_lock 0\n"); > + return; > + } > + > + ret = hwspin_lock_timeout(lock, TIMEOUT); > + if (ret) > + pr_warn("Can't lock requested hwspin_lock 0 (%d)\n", ret); > + > + hwspin_unlock(lock); > + ret = hwspin_lock_free(lock); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > +} > + > +static void hwspin_lock_test_request(void) I would suggest that your test cases test one thing each; so request and request_specific should test just that. > +{ > + struct hwspinlock *lock = hwspin_lock_request(); > + int ret; > + > + if (!lock) { > + pr_warn("Can't request hwspin_lock\n"); > + return; > + } > + > + ret = hwspin_lock_timeout(lock, TIMEOUT); > + if (ret) > + pr_warn("Can't lock requested hwspin_lock (%d)\n", ret); Testing lock-timeout seems useful, but you probably want to elaborate it to ensure that it both can lock a lock but that it times out properly when trying to grab a taken lock. > + > + hwspin_unlock(lock); > + ret = hwspin_lock_free(lock); > + if (ret) > + pr_warn("Can't free requested hwspin_lock\n"); > +} > + > +/* > + * Test hwspinlock user API when module is loaded > + */ > +static int __init hwspin_lock_test_init(void) > +{ > + pr_warn("Loading hwspinlock test module\n"); > + hwspin_lock_test_request(); > + hwspin_lock_test_request_specific(); > + hwspin_lock_test_trylock(); > + hwspin_lock_test_twice_request_specific(); > + > + return 0; > +} > + Regards, Bjorn
WARNING: multiple messages have this Message-ID (diff)
From: bjorn.andersson@linaro.org (Bjorn Andersson) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH 5/5] hwspinlock: Add test module Date: Wed, 7 Nov 2018 23:06:04 -0800 [thread overview] Message-ID: <20181108070604.GE12063@builder> (raw) In-Reply-To: <20181031093032.20386-6-benjamin.gaignard@st.com> On Wed 31 Oct 02:30 PDT 2018, Benjamin Gaignard wrote: > Create a test module to perform simple unitary tests on hwspinlock. > It doesn't cover all the possibles cases but at least allow to test > that very basic features are working. > I like the idea of making these things testable, but I would like to hear from others how useful this module would be to them before merging it. > Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com> > --- > drivers/hwspinlock/Kconfig | 9 +++ > drivers/hwspinlock/Makefile | 1 + > drivers/hwspinlock/hwspinlock_test.c | 132 +++++++++++++++++++++++++++++++++++ > 3 files changed, 142 insertions(+) > create mode 100644 drivers/hwspinlock/hwspinlock_test.c > > diff --git a/drivers/hwspinlock/Kconfig b/drivers/hwspinlock/Kconfig > index e1a20b460590..f340ebb4d1f7 100644 > --- a/drivers/hwspinlock/Kconfig > +++ b/drivers/hwspinlock/Kconfig > @@ -68,3 +68,12 @@ config HWSPINLOCK_STM32 > Say y here to support the STM32 Hardware Spinlock device. > > If unsure, say N. > + > +config HWSPINLOCK_TEST > + tristate "hwspinlock test module" > + depends on HWSPINLOCK && m > + default n > + help > + Select M here if you want to build hwspinlock test module > + > + If unsure, say N. > diff --git a/drivers/hwspinlock/Makefile b/drivers/hwspinlock/Makefile > index c0a9505b4dcf..e6b3d0212fe0 100644 > --- a/drivers/hwspinlock/Makefile > +++ b/drivers/hwspinlock/Makefile > @@ -10,3 +10,4 @@ obj-$(CONFIG_HWSPINLOCK_SIRF) += sirf_hwspinlock.o > obj-$(CONFIG_HWSPINLOCK_SPRD) += sprd_hwspinlock.o > obj-$(CONFIG_HSEM_U8500) += u8500_hsem.o > obj-$(CONFIG_HWSPINLOCK_STM32) += stm32_hwspinlock.o > +obj-$(CONFIG_HWSPINLOCK_TEST) += hwspinlock_test.o > diff --git a/drivers/hwspinlock/hwspinlock_test.c b/drivers/hwspinlock/hwspinlock_test.c > new file mode 100644 > index 000000000000..75819337e45a > --- /dev/null > +++ b/drivers/hwspinlock/hwspinlock_test.c > @@ -0,0 +1,132 @@ > +// SPDX-License-Identifier: GPL-2.0 > +/* > + * Copyright (C) STMicroelectronics SA 2018 > + * Author: Benjamin Gaignard <benjamin.gaignard@st.com> for STMicroelectronics. > + * License terms: GNU General Public License (GPL), version 2 > + */ > + > +#include <linux/hwspinlock.h> > +#include <linux/kernel.h> > +#include <linux/module.h> > + > +#define TIMEOUT 50 /* ms */ > + > +static void hwspin_lock_test_twice_request_specific(void) > +{ > + struct hwspinlock *lock0 = hwspin_lock_request_specific(0); > + struct hwspinlock *lock1 = hwspin_lock_request_specific(1); In a platform with more than one hwspinlock this will test the implementation that was first probed and it will assume that there's no issues with messing with the first two locks from this provider. > + > + int ret; > + > + if (!lock0 || !lock1) { > + pr_warn("Can't request twice hwspin_lock\n"); > + return; > + } > + > + ret = hwspin_trylock(lock0); > + if (ret) > + pr_warn("Can't trylock requested hwspin_lock 0 (%d)\n", ret); > + > + ret = hwspin_trylock(lock1); > + if (ret) > + pr_warn("Can't trylock requested hwspin_lock 1 (%d)\n", ret); > + > + hwspin_unlock(lock0); > + hwspin_unlock(lock1); As described in the kerneldoc for these it's a bug to attempt to unlock a unlocked lock. I believe this should be interpreted as it being invalid to unlock unless hwspin_trylock() returned successfully. > + > + ret = hwspin_lock_free(lock0); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > + > + ret = hwspin_lock_free(lock1); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > +} > + > +static void hwspin_lock_test_trylock(void) > +{ > + struct hwspinlock *lock = hwspin_lock_request(); > + int ret; > + > + if (!lock) { > + pr_warn("Can't request hwspin_lock\n"); > + return; > + } > + > + ret = hwspin_trylock(lock); > + if (ret) > + pr_warn("Can't trylock requested hwspin_lock (%d)\n", ret); > + > + /* Try to lock a second time, this should failed */ > + ret = hwspin_trylock(lock); > + if (ret != -EBUSY) > + pr_warn("Getting twice the same lock ! (%d)\n", ret); This shouldn't be a warning, as this is a test failure. > + > + hwspin_unlock(lock); > + ret = hwspin_lock_free(lock); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > +} > + > +static void hwspin_lock_test_request_specific(void) > +{ > + /* > + * Let's assume that any hwspin_lock driver would at least have one > + * lock at index 0 > + */ > + struct hwspinlock *lock = hwspin_lock_request_specific(0); > + int ret; > + > + if (!lock) { > + pr_warn("Can't request hwspin_lock 0\n"); > + return; > + } > + > + ret = hwspin_lock_timeout(lock, TIMEOUT); > + if (ret) > + pr_warn("Can't lock requested hwspin_lock 0 (%d)\n", ret); > + > + hwspin_unlock(lock); > + ret = hwspin_lock_free(lock); > + if (ret) > + pr_warn("Can't free requested hwspin_lock 0\n"); > +} > + > +static void hwspin_lock_test_request(void) I would suggest that your test cases test one thing each; so request and request_specific should test just that. > +{ > + struct hwspinlock *lock = hwspin_lock_request(); > + int ret; > + > + if (!lock) { > + pr_warn("Can't request hwspin_lock\n"); > + return; > + } > + > + ret = hwspin_lock_timeout(lock, TIMEOUT); > + if (ret) > + pr_warn("Can't lock requested hwspin_lock (%d)\n", ret); Testing lock-timeout seems useful, but you probably want to elaborate it to ensure that it both can lock a lock but that it times out properly when trying to grab a taken lock. > + > + hwspin_unlock(lock); > + ret = hwspin_lock_free(lock); > + if (ret) > + pr_warn("Can't free requested hwspin_lock\n"); > +} > + > +/* > + * Test hwspinlock user API when module is loaded > + */ > +static int __init hwspin_lock_test_init(void) > +{ > + pr_warn("Loading hwspinlock test module\n"); > + hwspin_lock_test_request(); > + hwspin_lock_test_request_specific(); > + hwspin_lock_test_trylock(); > + hwspin_lock_test_twice_request_specific(); > + > + return 0; > +} > + Regards, Bjorn
next prev parent reply other threads:[~2018-11-08 7:06 UTC|newest] Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top 2018-10-31 9:30 [PATCH 0/5] Add support of STM32 hwspinlock Benjamin Gaignard 2018-10-31 9:30 ` Benjamin Gaignard 2018-10-31 9:30 ` [PATCH 1/5] dt-bindings: hwlock: Document STM32 hwspinlock bindings Benjamin Gaignard 2018-10-31 9:30 ` Benjamin Gaignard 2018-11-08 6:49 ` Bjorn Andersson 2018-11-08 6:49 ` Bjorn Andersson 2018-10-31 9:30 ` [PATCH 2/5] hwspinlock: add STM32 hwspinlock device Benjamin Gaignard 2018-10-31 9:30 ` Benjamin Gaignard 2018-11-08 6:47 ` Bjorn Andersson 2018-11-08 6:47 ` Bjorn Andersson 2018-10-31 9:30 ` [PATCH 3/5] ARM: dts: stm32: Add hwspinlock node for stm32mp157 SoC Benjamin Gaignard 2018-10-31 9:30 ` Benjamin Gaignard 2018-11-08 6:51 ` Bjorn Andersson 2018-11-08 6:51 ` Bjorn Andersson 2018-10-31 9:30 ` [PATCH 4/5] ARM: dts: stm32: enable hwspinlock on stm32mp157c-ed1 Benjamin Gaignard 2018-10-31 9:30 ` Benjamin Gaignard 2018-10-31 9:30 ` [PATCH 5/5] hwspinlock: Add test module Benjamin Gaignard 2018-10-31 9:30 ` Benjamin Gaignard 2018-11-08 7:06 ` Bjorn Andersson [this message] 2018-11-08 7:06 ` Bjorn Andersson
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=20181108070604.GE12063@builder \ --to=bjorn.andersson@linaro.org \ --cc=alexandre.torgue@st.com \ --cc=benjamin.gaignard@linaro.org \ --cc=benjamin.gaignard@st.com \ --cc=devicetree@vger.kernel.org \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-remoteproc@vger.kernel.org \ --cc=linux-stm32@st-md-mailman.stormreply.com \ --cc=mark.rutland@arm.com \ --cc=ohad@wizery.com \ --cc=robh+dt@kernel.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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.