<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux-toradex.git/lib/vdso, branch v5.6-rc6</title>
<subtitle>Linux kernel for Apalis and Colibri modules</subtitle>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/'/>
<entry>
<title>lib/vdso: Only read hrtimer_res when needed in __cvdso_clock_getres()</title>
<updated>2020-01-16T20:24:46+00:00</updated>
<author>
<name>Christophe Leroy</name>
<email>christophe.leroy@c-s.fr</email>
</author>
<published>2020-01-16T17:58:27+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=49a101d7169c7729c7bab6b2f896faae34bd6c3d'/>
<id>49a101d7169c7729c7bab6b2f896faae34bd6c3d</id>
<content type='text'>
Only perform READ_ONCE(vd[CS_HRES_COARSE].hrtimer_res) for
HRES and RAW clocks.

Signed-off-by: Christophe Leroy &lt;christophe.leroy@c-s.fr&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/7ac2f0d21652f95e2bbdfa6bd514ae6c7caf53ab.1579196675.git.christophe.leroy@c-s.fr
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Only perform READ_ONCE(vd[CS_HRES_COARSE].hrtimer_res) for
HRES and RAW clocks.

Signed-off-by: Christophe Leroy &lt;christophe.leroy@c-s.fr&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/7ac2f0d21652f95e2bbdfa6bd514ae6c7caf53ab.1579196675.git.christophe.leroy@c-s.fr
</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Prepare for time namespace support</title>
<updated>2020-01-14T11:20:57+00:00</updated>
<author>
<name>Thomas Gleixner</name>
<email>tglx@linutronix.de</email>
</author>
<published>2019-11-12T01:27:09+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=660fd04f9317172ae90f414c68b18a26ae88a829'/>
<id>660fd04f9317172ae90f414c68b18a26ae88a829</id>
<content type='text'>
To support time namespaces in the vdso with a minimal impact on regular non
time namespace affected tasks, the namespace handling needs to be hidden in
a slow path.

The most obvious place is vdso_seq_begin(). If a task belongs to a time
namespace then the VVAR page which contains the system wide vdso data is
replaced with a namespace specific page which has the same layout as the
VVAR page. That page has vdso_data-&gt;seq set to 1 to enforce the slow path
and vdso_data-&gt;clock_mode set to VCLOCK_TIMENS to enforce the time
namespace handling path.

The extra check in the case that vdso_data-&gt;seq is odd, e.g. a concurrent
update of the vdso data is in progress, is not really affecting regular
tasks which are not part of a time namespace as the task is spin waiting
for the update to finish and vdso_data-&gt;seq to become even again.

If a time namespace task hits that code path, it invokes the corresponding
time getter function which retrieves the real VVAR page, reads host time
and then adds the offset for the requested clock which is stored in the
special VVAR page.

If VDSO time namespace support is disabled the whole magic is compiled out.

Initial testing shows that the disabled case is almost identical to the
host case which does not take the slow timens path. With the special timens
page installed the performance hit is constant time and in the range of
5-7%.

For the vdso functions which are not using the sequence count an
unconditional check for vdso_data-&gt;clock_mode is added which switches to
the real vdso when the clock_mode is VCLOCK_TIMENS.

[avagin: Make do_hres_timens() work with raw clocks too: choose vdso_data
 pointer by CS_RAW offset.]

Suggested-by: Andy Lutomirski &lt;luto@kernel.org&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Signed-off-by: Andrei Vagin &lt;avagin@gmail.com&gt;
Signed-off-by: Dmitry Safonov &lt;dima@arista.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20191112012724.250792-21-dima@arista.com


</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
To support time namespaces in the vdso with a minimal impact on regular non
time namespace affected tasks, the namespace handling needs to be hidden in
a slow path.

The most obvious place is vdso_seq_begin(). If a task belongs to a time
namespace then the VVAR page which contains the system wide vdso data is
replaced with a namespace specific page which has the same layout as the
VVAR page. That page has vdso_data-&gt;seq set to 1 to enforce the slow path
and vdso_data-&gt;clock_mode set to VCLOCK_TIMENS to enforce the time
namespace handling path.

The extra check in the case that vdso_data-&gt;seq is odd, e.g. a concurrent
update of the vdso data is in progress, is not really affecting regular
tasks which are not part of a time namespace as the task is spin waiting
for the update to finish and vdso_data-&gt;seq to become even again.

If a time namespace task hits that code path, it invokes the corresponding
time getter function which retrieves the real VVAR page, reads host time
and then adds the offset for the requested clock which is stored in the
special VVAR page.

If VDSO time namespace support is disabled the whole magic is compiled out.

Initial testing shows that the disabled case is almost identical to the
host case which does not take the slow timens path. With the special timens
page installed the performance hit is constant time and in the range of
5-7%.

For the vdso functions which are not using the sequence count an
unconditional check for vdso_data-&gt;clock_mode is added which switches to
the real vdso when the clock_mode is VCLOCK_TIMENS.

[avagin: Make do_hres_timens() work with raw clocks too: choose vdso_data
 pointer by CS_RAW offset.]

Suggested-by: Andy Lutomirski &lt;luto@kernel.org&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Signed-off-by: Andrei Vagin &lt;avagin@gmail.com&gt;
Signed-off-by: Dmitry Safonov &lt;dima@arista.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20191112012724.250792-21-dima@arista.com


</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Mark do_hres() and do_coarse() as __always_inline</title>
<updated>2020-01-14T11:20:48+00:00</updated>
<author>
<name>Andrei Vagin</name>
<email>avagin@gmail.com</email>
</author>
<published>2019-11-12T01:26:51+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=c966533f8c6c45f93c52599f8460e7695f0b7eaa'/>
<id>c966533f8c6c45f93c52599f8460e7695f0b7eaa</id>
<content type='text'>
Performance numbers for Intel(R) Core(TM) i5-6300U CPU @ 2.40GHz
(more clock_gettime() cycles - the better):

clock            | before     | after      | diff
----------------------------------------------------------
monotonic        |  153222105 |  166775025 | 8.8%
monotonic-coarse |  671557054 |  691513017 | 3.0%
monotonic-raw    |  147116067 |  161057395 | 9.5%
boottime         |  153446224 |  166962668 | 9.1%

The improvement for arm64 for monotonic and boottime is around 3.5%.

clock            | before     | after      | diff
==================================================
monotonic          17326692     17951770     3.6%
monotonic-coarse   43624027     44215292     1.3%
monotonic-raw      17541809     17554932     0.1%
boottime           17334982     17954361     3.5%

[ tglx: Avoid the goto ]

Signed-off-by: Andrei Vagin &lt;avagin@gmail.com&gt;
Signed-off-by: Dmitry Safonov &lt;dima@arista.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20191112012724.250792-3-dima@arista.com


</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Performance numbers for Intel(R) Core(TM) i5-6300U CPU @ 2.40GHz
(more clock_gettime() cycles - the better):

clock            | before     | after      | diff
----------------------------------------------------------
monotonic        |  153222105 |  166775025 | 8.8%
monotonic-coarse |  671557054 |  691513017 | 3.0%
monotonic-raw    |  147116067 |  161057395 | 9.5%
boottime         |  153446224 |  166962668 | 9.1%

The improvement for arm64 for monotonic and boottime is around 3.5%.

clock            | before     | after      | diff
==================================================
monotonic          17326692     17951770     3.6%
monotonic-coarse   43624027     44215292     1.3%
monotonic-raw      17541809     17554932     0.1%
boottime           17334982     17954361     3.5%

[ tglx: Avoid the goto ]

Signed-off-by: Andrei Vagin &lt;avagin@gmail.com&gt;
Signed-off-by: Dmitry Safonov &lt;dima@arista.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20191112012724.250792-3-dima@arista.com


</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Avoid duplication in __cvdso_clock_getres()</title>
<updated>2020-01-14T11:20:47+00:00</updated>
<author>
<name>Christophe Leroy</name>
<email>christophe.leroy@c-s.fr</email>
</author>
<published>2019-12-23T14:31:09+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=cdb7c5a9c897ab2e5c56df647dd84c84e150e925'/>
<id>cdb7c5a9c897ab2e5c56df647dd84c84e150e925</id>
<content type='text'>
VDSO_HRES and VDSO_RAW clocks are handled the same way.

Avoid the code duplication.

Signed-off-by: Christophe Leroy &lt;christophe.leroy@c-s.fr&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Reviewed-by: Andy Lutomirski &lt;luto@kernel.org&gt;
Link: https://lore.kernel.org/r/fdf1a968a8f7edd61456f1689ac44082ebb19c15.1577111367.git.christophe.leroy@c-s.fr


</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
VDSO_HRES and VDSO_RAW clocks are handled the same way.

Avoid the code duplication.

Signed-off-by: Christophe Leroy &lt;christophe.leroy@c-s.fr&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Reviewed-by: Andy Lutomirski &lt;luto@kernel.org&gt;
Link: https://lore.kernel.org/r/fdf1a968a8f7edd61456f1689ac44082ebb19c15.1577111367.git.christophe.leroy@c-s.fr


</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Let do_coarse() return 0 to simplify the callsite</title>
<updated>2020-01-14T11:20:46+00:00</updated>
<author>
<name>Christophe Leroy</name>
<email>christophe.leroy@c-s.fr</email>
</author>
<published>2019-12-23T14:31:07+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=8463cf80529d0fd80b84cd5ab8b9b952b01c7eb9'/>
<id>8463cf80529d0fd80b84cd5ab8b9b952b01c7eb9</id>
<content type='text'>
do_coarse() is similar to do_hres() except that it never fails.

Change its type to int instead of void and let it always return success (0)
to simplify the call site.

Signed-off-by: Christophe Leroy &lt;christophe.leroy@c-s.fr&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/21e8afa38c02ca8672c2690307383507fe63b454.1577111367.git.christophe.leroy@c-s.fr


</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
do_coarse() is similar to do_hres() except that it never fails.

Change its type to int instead of void and let it always return success (0)
to simplify the call site.

Signed-off-by: Christophe Leroy &lt;christophe.leroy@c-s.fr&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/21e8afa38c02ca8672c2690307383507fe63b454.1577111367.git.christophe.leroy@c-s.fr


</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Remove checks on return value for 32 bit vDSO</title>
<updated>2020-01-14T11:20:45+00:00</updated>
<author>
<name>Vincenzo Frascino</name>
<email>vincenzo.frascino@arm.com</email>
</author>
<published>2019-08-30T13:58:59+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=a279235ddbe975670afe2267162028ec0a312293'/>
<id>a279235ddbe975670afe2267162028ec0a312293</id>
<content type='text'>
Since all the architectures that support the generic vDSO library have
been converted to support the 32 bit fallbacks it is not required
anymore to check the return value of __cvdso_clock_get*time32_common()
before updating the old_timespec fields.

Remove the related checks from the generic vdso library.

References: c60a32ea4f45 ("lib/vdso/32: Provide legacy syscall fallbacks")
Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20190830135902.20861-6-vincenzo.frascino@arm.com


</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Since all the architectures that support the generic vDSO library have
been converted to support the 32 bit fallbacks it is not required
anymore to check the return value of __cvdso_clock_get*time32_common()
before updating the old_timespec fields.

Remove the related checks from the generic vdso library.

References: c60a32ea4f45 ("lib/vdso/32: Provide legacy syscall fallbacks")
Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20190830135902.20861-6-vincenzo.frascino@arm.com


</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Remove VDSO_HAS_32BIT_FALLBACK</title>
<updated>2020-01-14T11:20:44+00:00</updated>
<author>
<name>Vincenzo Frascino</name>
<email>vincenzo.frascino@arm.com</email>
</author>
<published>2019-08-30T13:58:58+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=b767081c07a400ff1c6f95b87639a9405886e7a6'/>
<id>b767081c07a400ff1c6f95b87639a9405886e7a6</id>
<content type='text'>
VDSO_HAS_32BIT_FALLBACK was introduced to address a regression which
caused seccomp to deny access to the applications to clock_gettime64()
and clock_getres64() because they are not enabled in the existing
filters.

The purpose of VDSO_HAS_32BIT_FALLBACK was to simplify the conditional
implementation of __cvdso_clock_get*time32() variants.

Now that all the architectures that support the generic vDSO library
have been converted to support the 32 bit fallbacks the conditional
can be removed.

Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20190830135902.20861-5-vincenzo.frascino@arm.com

References: c60a32ea4f45 ("lib/vdso/32: Provide legacy syscall fallbacks")

</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
VDSO_HAS_32BIT_FALLBACK was introduced to address a regression which
caused seccomp to deny access to the applications to clock_gettime64()
and clock_getres64() because they are not enabled in the existing
filters.

The purpose of VDSO_HAS_32BIT_FALLBACK was to simplify the conditional
implementation of __cvdso_clock_get*time32() variants.

Now that all the architectures that support the generic vDSO library
have been converted to support the 32 bit fallbacks the conditional
can be removed.

Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20190830135902.20861-5-vincenzo.frascino@arm.com

References: c60a32ea4f45 ("lib/vdso/32: Provide legacy syscall fallbacks")

</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Build 32 bit specific functions in the right context</title>
<updated>2020-01-14T11:20:44+00:00</updated>
<author>
<name>Vincenzo Frascino</name>
<email>vincenzo.frascino@arm.com</email>
</author>
<published>2019-08-30T13:58:56+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=bf279849ad59538a1518c667c0795ec1fe9dbd66'/>
<id>bf279849ad59538a1518c667c0795ec1fe9dbd66</id>
<content type='text'>
clock_gettime32 and clock_getres_time32 should be compiled only with a
32 bit vdso library.

Exclude these symbols when BUILD_VDSO32 is not defined.

Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Reviewed-by: Andy Lutomirski &lt;luto@kernel.org&gt;
Link: https://lore.kernel.org/r/20190830135902.20861-3-vincenzo.frascino@arm.com


</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
clock_gettime32 and clock_getres_time32 should be compiled only with a
32 bit vdso library.

Exclude these symbols when BUILD_VDSO32 is not defined.

Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Reviewed-by: Andy Lutomirski &lt;luto@kernel.org&gt;
Link: https://lore.kernel.org/r/20190830135902.20861-3-vincenzo.frascino@arm.com


</pre>
</div>
</content>
</entry>
<entry>
<title>lib/vdso: Make __cvdso_clock_getres() static</title>
<updated>2020-01-10T18:29:01+00:00</updated>
<author>
<name>Vincenzo Frascino</name>
<email>vincenzo.frascino@arm.com</email>
</author>
<published>2019-11-28T11:17:19+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=ffd08731b2d632459428612431060cf902324a8d'/>
<id>ffd08731b2d632459428612431060cf902324a8d</id>
<content type='text'>
Fix the following sparse warning in the generic vDSO library:

  linux/lib/vdso/gettimeofday.c:224:5: warning: symbol
  '__cvdso_clock_getres' was not declared. Should it be static?

Make it static and also mark it __maybe_unsed.

Fixes: 502a590a170b ("lib/vdso: Move fallback invocation to the callers")
Reported-by: Marc Gonzalez &lt;marc.w.gonzalez@free.fr&gt;
Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20191128111719.8282-1-vincenzo.frascino@arm.com
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Fix the following sparse warning in the generic vDSO library:

  linux/lib/vdso/gettimeofday.c:224:5: warning: symbol
  '__cvdso_clock_getres' was not declared. Should it be static?

Make it static and also mark it __maybe_unsed.

Fixes: 502a590a170b ("lib/vdso: Move fallback invocation to the callers")
Reported-by: Marc Gonzalez &lt;marc.w.gonzalez@free.fr&gt;
Signed-off-by: Vincenzo Frascino &lt;vincenzo.frascino@arm.com&gt;
Signed-off-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Link: https://lore.kernel.org/r/20191128111719.8282-1-vincenzo.frascino@arm.com
</pre>
</div>
</content>
</entry>
<entry>
<title>y2038: vdso: change time_t to __kernel_old_time_t</title>
<updated>2019-11-15T13:38:28+00:00</updated>
<author>
<name>Arnd Bergmann</name>
<email>arnd@arndb.de</email>
</author>
<published>2019-11-05T10:10:01+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=21346564ccad17b928cf0d51584608531d91d298'/>
<id>21346564ccad17b928cf0d51584608531d91d298</id>
<content type='text'>
Only x86 uses the 'time' syscall in vdso, so change that to
__kernel_old_time_t as a preparation for removing 'time_t' and
'__kernel_time_t' later.

Reviewed-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Signed-off-by: Arnd Bergmann &lt;arnd@arndb.de&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Only x86 uses the 'time' syscall in vdso, so change that to
__kernel_old_time_t as a preparation for removing 'time_t' and
'__kernel_time_t' later.

Reviewed-by: Thomas Gleixner &lt;tglx@linutronix.de&gt;
Signed-off-by: Arnd Bergmann &lt;arnd@arndb.de&gt;
</pre>
</div>
</content>
</entry>
</feed>
