[Bug nptl/11588] pthread condvars are not priority inheritance aware

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[jan.kiszka at siemens dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

Jan Kiszka <jan.kiszka at siemens dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jan.kiszka at siemens dot
                   |                            |com

--- Comment #17 from Jan Kiszka <jan.kiszka at siemens dot com> 2011-02-14 17:28:26 UTC ---
(In reply to comment #16)
> Is there any chance that this issue is going to make any progress ? It affects
> real users and we have a working solution for it.

I can only confirm this. We just saw that unfortunate deadlock again in real
application (something to be shipped in products). We ran into it before, but
this time we finally realized that it's not due to an outdated glibc but rather
a still unfixed bug.

Our applications consist of dozens of real-time threads at different prio
levels and practically can't be reworked to avoid cond variables in RT threads.
So we really need a solution, and that preferably without having to carry our
own patched glibc versions.

Thanks,
Jan

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhart at linux dot intel.com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

Darren Hart <dvhart at linux dot intel.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dvhart at linux dot
                   |                            |intel.com

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[jan.kiszka at siemens dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #18 from Jan Kiszka <jan.kiszka at siemens dot com> 2012-06-05 16:36:45 UTC ---
Ping on this unfixed bug. Is there a chance to reconsider the extension Darren
proposed? We are stuck with unhandy and inefficient PRIO_PROTECT here due to
this issue. And I'm sure we aren't alone...

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[triegel at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

Torvald Riegel <triegel at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |triegel at redhat dot com

--- Comment #19 from Torvald Riegel <triegel at redhat dot com> 2012-10-25 11:03:02 UTC ---
Note that the condvar implementation will have to change due to a recent POSIX
change (see bug 13165).  That means that how to solve this bug might also
depend on how we fix bug 13165.  It would be good to fix both in one go.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[schwab@linux-m68k.org]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

Andreas Schwab <schwab@linux-m68k.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|drepper.fsp at gmail dot    |unassigned at sourceware
                   |com                         |dot org

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[scot4spam at yahoo dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

Scot Salmon <scot4spam at yahoo dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |scot4spam at yahoo dot com

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[davidh at realtimeint dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

davidh at realtimeint dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |davidh at realtimeint dot
                   |                            |com

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[stano at meduna dot org]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

Stanislav Meduna <stano at meduna dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |stano at meduna dot org

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #4786|0                           |1
        is obsolete|                            |
                 CC|                            |gratian.crisan at ni dot com

--- Comment #20 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7238
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7238&action=edit
Ported to glibc 2.19

Changes since V2:
* Ported to glibc 2.19
* Dropped the changes related to making the external mutex associated with the
condvar PI-aware since they have been already committed.
* Moved lll_pi_lock()/lll_pi_unlock() to lowlevelpilock.c
* Changed the value bitmap in pthread_condattr to use only 1 bit for clock ID
This matches the equivalent bitmap in the __nwaiters field.
* Used the new condattr masks and shifts where appropriate

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #7238|Ported to glibc 2.19        |[PATCH 1/3 V3] pi-condvars:
        description|                            |add protocol support to
                   |                            |pthread_condattr_t

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #4787|0                           |1
        is obsolete|                            |

--- Comment #21 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7239
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7239&action=edit
[PATCH 2/3 V3] pi-condvars: add tst-condpi1.c, basic API test

Changes since V2:
  * updated copyright to 2013 and tested on glibc 2.19
  * functional equivalent to V2

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #4788|0                           |1
        is obsolete|                            |

--- Comment #22 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7240
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7240&action=edit
[PATCH 3/3 V3] pi-condvars: add tst-condpi2.c, priority inversion avoidance
test

Changes since V2:
  * updated copyright to 2013 and tested on glibc 2.19
  * functional equivalent to V2

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #7238|0                           |1
        is obsolete|                            |

--- Comment #23 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7282
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7282&action=edit
[PATCH 1/3 V3] pi-condvars: add protocol support to pthread_condattr_t

Updated with lates changes suggested on libc-alpha:
https://sourceware.org/ml/libc-alpha/2013-10/msg00731.html

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[davidh at realtimeint dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #24 from davidh at realtimeint dot com ---
GLibC v2.19 was just released and no fix for this issue. Why is this? What's
involved in getting this done (seems like there's a patch available for v2.19
that works)?
-David

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[allan at archlinux dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Allan McRae <allan at archlinux dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |allan at archlinux dot org

--- Comment #25 from Allan McRae <allan at archlinux dot org> ---
It looks like the updated patches were posted to the mailing list were never
pinged once the copyright assignment was confirmed.  Also, it looks like there
were comments to be addressed.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[joseph at codesourcery dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #26 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
I don't see the copyright assignments yet on file according to 
copyright.list, though both contributors said they were working on them.  
If sent in, assign@gnu.org should be chased up about getting 
copyright.list updated, and once the assignments are ready, the 
contributors should keep pinging the patches weekly on libc-alpha for as 
long as needed as per the contribution checklist.

We still need someone to take charge of working through glibc's open nptl 
bugs and reviewing / fixing them (there are a few people with significant 
expertise in the area, but not so much systematic work through all the 
bugs).

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[triegel at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #27 from Torvald Riegel <triegel at redhat dot com> ---
I'm also working on a new condvar implementation, motivated by a POSIX
clarification that requires stronger signaler--waiter guarantees than what's
guaranteed by the current implementation (IOW, the current implementation does
not fulfill POSIX requirements).  The new implementation also fixes an ABA
issue in the current implementation. The PI support in there is incomplete yet,
however, and it might require futex changes/additions.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #28 from Gratian Crisan <gratian.crisan at ni dot com> ---
(In reply to joseph@codesourcery.com from comment #26)
> I don't see the copyright assignments yet on file according to 
> copyright.list, though both contributors said they were working on them.  
> If sent in, assign@gnu.org should be chased up about getting 
> copyright.list updated, and once the assignments are ready, the 
> contributors should keep pinging the patches weekly on libc-alpha for as 
> long as needed as per the contribution checklist.
> 
> We still need someone to take charge of working through glibc's open nptl 
> bugs and reviewing / fixing them (there are a few people with significant 
> expertise in the area, but not so much systematic work through all the 
> bugs).

Last I have heard the right people signed the required papers on our side and
they were sent to assign@gnu.org sometime mid-December. I haven't heard
anything back. Let me see if I can get a status update. Once the copyright
assignment is confirmed I can provide updated patches.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[darren.hart at intel dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #29 from Darren Hart <darren.hart at intel dot com> ---
Gratian, right, as far as I understand it, Paul has provided all the necessary
information for the IBM assignment for my and Dinakar's contributions. If there
is anything I can do to help here, please let me know.

Torvalds, I believe in our discussions we agreed that the work you are doing is
orthogonal to this, and while it may impact the same code paths, it was best to
deal with them as separate efforts and not block one on the other. I would be
very interested in continuing to follow along on that effort as well though.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[triegel at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #30 from Torvald Riegel <triegel at redhat dot com> ---
(In reply to Darren Hart from comment #29)
> Torvalds, I believe in our discussions we agreed that the work you are doing
> is orthogonal to this, and while it may impact the same code paths, it was
> best to deal with them as separate efforts and not block one on the other. I
> would be very interested in continuing to follow along on that effort as
> well though.

Yes, it is orthogonal.  It's just that the current implementation, with or
without your fixes, does not fulfill the stronger waiter-signaler requirements
due to the POSIX clarification.  Thus, we'll need to do something about the
latter; but it's no reason to not include your fixes while the latter isn't
ready.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[jackie.rosen at hushmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Jackie Rosen <jackie.rosen at hushmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jackie.rosen at hushmail dot com

--- Comment #31 from Jackie Rosen <jackie.rosen at hushmail dot com> ---
*** Bug 260998 has been marked as a duplicate of this bug. ***
Seen from the domain http://volichat.com
Page where seen: http://volichat.com/adult-chat-rooms
Marked for reference. Resolved as fixed @bugzilla.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[john.ogness at linutronix dot de]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

John Ogness <john.ogness at linutronix dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |john.ogness at linutronix dot de

--- Comment #32 from John Ogness <john.ogness at linutronix dot de> ---
I believe there is an error in [PATCH 1/3 V3] where COND_PRIO_PROTECT will
clobber the CLOCKID in cond->__data.__nwaiters. This is because the
COND_PROTOCOL values are not shifted by COND_PROTOCOL_SHIFT before reading and
writing to __nwaiters.

I believe the switch cases in pthread_cond_init.c:__pthread_cond_init() should
look like this:

    case PTHREAD_PRIO_INHERIT:
      cond->__data.__nwaiters |= COND_PRIO_INHERIT << COND_PROTOCOL_SHIFT;
      break;

    case PTHREAD_PRIO_PROTECT:
      cond->__data.__nwaiters |= COND_PRIO_PROTECT << COND_PROTOCOL_SHIFT;
      break;


And the cond_lock() and cond_unlock() in cond-lock.h should look like this:

static inline void
cond_lock(pthread_cond_t *cond, int pshared)
{
  if (pshared == LLL_PRIVATE
      && (((cond->__data.__nwaiters & COND_PROTOCOL_MASK) >>
COND_PROTOCOL_SHIFT)
         == COND_PRIO_INHERIT))
    lll_pi_lock (cond->__data.__lock, pshared);
  else
    lll_lock (cond->__data.__lock, pshared);
}

static inline void
cond_unlock(pthread_cond_t *cond, int pshared)
{
  if (pshared == LLL_PRIVATE
      && (((cond->__data.__nwaiters & COND_PROTOCOL_MASK) >>
COND_PROTOCOL_SHIFT)
         == COND_PRIO_INHERIT))
    lll_pi_unlock (cond->__data.__lock, pshared);
  else
    lll_unlock (cond->__data.__lock, pshared);
}

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[darren.hart at intel dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #33 from Darren Hart <darren.hart at intel dot com> ---
I looked through this, and you appear to be correct. We didn't catch it in
testing as the test case only depended on reading and writing the
COND_PROTOCOL_INHERIT, which works as the shift was missing from both
locations.

Oddly, I had observed what I thought was a similar error with the CLOCK_ID, but
after going through it realized it was just an optimization as the CLOCK_ID is
bit 0. Sadly, I didn't catch this while staring at the same code :-(

Nice catch John!

Who wants to respin the patches?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[darren.hart at intel dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Darren Hart <darren.hart at intel dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|dvhart at linux dot intel.com      |

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #34 from Gratian Crisan <gratian.crisan at ni dot com> ---
(In reply to Darren Hart from comment #33)
> I looked through this, and you appear to be correct. We didn't catch it in
> testing as the test case only depended on reading and writing the
> COND_PROTOCOL_INHERIT, which works as the shift was missing from both
> locations.
> 
> Oddly, I had observed what I thought was a similar error with the CLOCK_ID,
> but after going through it realized it was just an optimization as the
> CLOCK_ID is bit 0. Sadly, I didn't catch this while staring at the same code
> :-(
> 
> Nice catch John!
> 
> Who wants to respin the patches?

I can do it but it's going to take me a couple of days due to current workload.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[john.ogness at linutronix dot de]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #35 from John Ogness <john.ogness at linutronix dot de> ---
@Gratian: Maybe while you are at it, you can change the protocol enum in
internaltypes.h to:

enum {
  COND_PRIO_INHERIT = 1,
  COND_PRIO_PROTECT
};

Although technically irrelevant, I found it confusing why it begins with 2.
This was probably the reason for Darren's oversight in the protocol shift bug.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #7282|0                           |1
        is obsolete|                            |

--- Comment #36 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7442
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7442&action=edit
[PATCH 1/3 V4] pi-condvars: add protocol support to pthread_condattr_t

Incorporated suggestions made by John Ogness and updated the version to 2.20
now that 2.19 is released.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[darren.hart at intel dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #37 from Darren Hart <darren.hart at intel dot com> ---
I just checked libc git master and do not see these changes merged. Is there
some pending review or process yet to be completed? Anything needed that I can
help with?

Is the copyright assignment issue closed?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #38 from Gratian Crisan <gratian.crisan at ni dot com> ---
(In reply to Darren Hart from comment #37)

> Is the copyright assignment issue closed?

I apologize for the (lawyer induced) delay. I finally got confirmation that the
assignment was signed on the FSF side.
I will dust off the patches in the next few days, add benchmarks for x86 (to
hopefully prove that an assembly implementation is not necessary) and resubmit
them.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[darren.hart at intel dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #39 from Darren Hart <darren.hart at intel dot com> ---
On 4/17/14, 7:53, "gratian.crisan at ni dot com"
<sourceware-bugzilla@sourceware.org> wrote:

>https://sourceware.org/bugzilla/show_bug.cgi?id=11588
>
>--- Comment #38 from Gratian Crisan <gratian.crisan at ni dot com> ---
>(In reply to Darren Hart from comment #37)
>
>> Is the copyright assignment issue closed?
>
>I apologize for the (lawyer induced) delay. I finally got confirmation
>that the
>assignment was signed on the FSF side.
>I will dust off the patches in the next few days, add benchmarks for x86
>(to
>hopefully prove that an assembly implementation is not necessary) and
>resubmit
>them.


Awesome!

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #40 from Gratian Crisan <gratian.crisan at ni dot com> ---
Sorry, this is taking me longer than I originally thought. It turns out that
this commit:
https://sourceware.org/git/?p=glibc.git;a=commit;h=47c5adebd2c864a098c3af66e61e1147dc3cf0b4
is unconditionally disabling PI futexes on ARM because there is a kernel
configuration (CONFIG_CPU_USE_DOMAINS && CONFIG_SMP) that doesn't support them.
So I need to find a solution for that first before I can finish testing on ARM.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[schwab at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Andreas Schwab <schwab at sourceware dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|jackie.rosen at hushmail dot com   |

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
              Flags|                            |security-

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #41 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7684
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7684&action=edit
[PATCH 4/6] benchtests: Add benchmarks for pthread_cond_* functions

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #42 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7685
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7685&action=edit
[PATCH 5/6] x86_64: Remove assembly implementations for pthread_cond_*

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #43 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7686
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7686&action=edit
[PATCH 6/6] arm: Re-enable PI futex support for ARM kernels >= 3.14.3

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #7442|0                           |1
        is obsolete|                            |

--- Comment #44 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7687
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7687&action=edit
[PATCH 1/6 V5] pi-condvars: add protocol support to pthread_condattr_t

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #7239|0                           |1
        is obsolete|                            |

--- Comment #45 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7688
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7688&action=edit
[PATCH 2/6] pi-condvars: add tst-condpi1.c, basic API test

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[gratian.crisan at ni dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Gratian Crisan <gratian.crisan at ni dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #7240|0                           |1
        is obsolete|                            |

--- Comment #46 from Gratian Crisan <gratian.crisan at ni dot com> ---
Created attachment 7689
  --> https://sourceware.org/bugzilla/attachment.cgi?id=7689&action=edit
[PATCH 3/6] pi-condvars: add tst-condpi2.c, priority inversion avoidance test

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[davidh at realtimeint dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #47 from davidh at realtimeint dot com ---
Looks like this again missed the last glibc release (2.20). Any chance this
might finally get into v2.21 (which looks to be scheduled for 01/2015)?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[triegel at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

--- Comment #48 from Torvald Riegel <triegel at redhat dot com> ---
The problem here is that the Austin Group has clarified that POSIX requires
stronger ordering guarantees wrt. waiters vs. signals than what the current
implementation offers.  We can implement them, but we're not aware of an
(efficient) implementation that fulfills both these clarified POSIX
requirements and can do PI under all circumstances:

https://sourceware.org/ml/libc-alpha/2014-08/msg00287.html

If you have further input on these choices, please comment on libc-alpha.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[carlos at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Carlos O'Donell <carlos at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |carlos at redhat dot com
           See Also|                            |https://bugzilla.redhat.com
                   |                            |/show_bug.cgi?id=605506

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[will at eccles dot dev]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Will Eccles <will at eccles dot dev> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |will at eccles dot dev

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[laurettastokes89 at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

runolfsdottir <laurettastokes89 at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |laurettastokes89 at gmail dot com

--- Comment #57 from runolfsdottir <laurettastokes89 at gmail dot com> ---
Also fixed on 2.35 branch.
https://jigsawpuzzle.io/

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[fweimer at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11588

Florian Weimer <fweimer at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|triegel at redhat dot com          |

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[tglx at linutronix dot de]

------- Additional Comments From tglx at linutronix dot de  2010-06-17 22:17 -------
(In reply to comment #15)
> Ulrich, before I spend more time on the test cases and possibly trying to 
> optimize the current implementation, can you comment on if you have any objection 
> to the proposed API changes themselves?

Is there any chance that this issue is going to make any progress ? It affects
real users and we have a working solution for it.

Are there any issues with the approach itself ? If not, what needs to be done to
get this resolved ?

Thanks,

      tglx



-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-06-17 20:59 -------
Ulrich, before I spend more time on the test cases and possibly trying to 
optimize the current implementation, can you comment on if you have any objection 
to the proposed API changes themselves?

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-28 20:34 -------
Created an attachment (id=4821)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4821&action=view)
pi condvar patch impact on normal path testcase

Updated condvar_perf testcase which runs the previous test multiple times and
calculates various statistics to aid in determining the impact of the
pi_condvar patches.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #4782 is|0                           |1
           obsolete|                            |


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-27 23:18 -------
I have posted V3 of the patches to libc-alpha. Rather than bog bugzilla down with 
additional attachments, I'll wait until an agreement has been made on libc-alpha. 
These patches have proper ChangeLogs and have removed the requeue-pi C 
implementation which the previous versions were carrying along.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 18:44 -------
It was brought to my attention that the Signed-off-by's in the patches attached 
and sent to libc-alpha are not appropriate for glibc development. This was just 
an artifact of my git usage, and I will correct in subsequent versions. May we 
discuss these as is, as we our foremost concern is agreeing on the proposed new 
condattr protocol APIs?

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 08:13 -------
Created an attachment (id=4789)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4789&action=view)
[PATCH 4/4 V2] pi-condvars: tst-condpi2.c, detect priority inversion 

Optional patch. This goes beyond relying on the glibc 2 second timeout to abort
the test. Instead, it detects the unbounded priority inversion and exits
gracefully. Depending on the preferred approach, this patch can be ignored
without a problem.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 08:10 -------
Created an attachment (id=4788)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4788&action=view)
[PATCH 3/4 V2] pi-condvars: add tst-condpi2.c, priority inversion avoidance
test 

New "make check" test, tst-condpi2.c, to test priority inversion avoidance.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 08:08 -------
Created an attachment (id=4787)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4787&action=view)
[PATCH 2/4 V2] pi-condvars: add tst-condpi1.c, basic API test 

Updated patch header.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #4781 is|0                           |1
           obsolete|                            |


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 08:07 -------
Created an attachment (id=4786)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4786&action=view)
[PATCH 1/4 V2] pi-condvars: add protocol support to pthread_condattr_t 

Updated patch header. Functionally equivalent to attachment 4780.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #4780 is|0                           |1
           obsolete|                            |


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 08:05 -------
Created an attachment (id=4785)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4785&action=view)
Unbounded priority inversion test case

The first version (attachment 4779) was missing the "_np" in the dl lookup.
This fixes that.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #4779 is|0                           |1
           obsolete|                            |


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 08:02 -------
Turns out the trouble in comment 5 was due to my forgetting to remove the 
nptl/pthread_cond*.S files before building the library this time around. The 
original 0001 patch did indeed function correctly. I have reworked the patches, 
improved the patch headers and naming as well as added a new tst-condpi2.c 
testcase which tests for priority inversion avoidance. A follow-on patch detects 
an unbounded inversion and exits gracefully. Without the last patch, the glibc 
test framework will kill the test after the 2 second timeout - which may be the 
preferred approach. I'll attach the updated patches now.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-13 01:55 -------
While working up tst-condpi2.c (make check version of pthread_cond_hang.c), I 
found something went wrong with my port attached as "0001-pi-condvar-
pthread_condattr_setprotocol_np-2.12-port.patch" and the test case still hangs 
when affined to CPU 0. I'll look into why and get an updated patch posted as soon 
as possible.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[pasky at suse dot cz]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pasky at suse dot cz


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dino at in dot ibm dot com]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dino at in dot ibm dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-11 18:55 -------
Created an attachment (id=4782)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4782&action=view)
pi condvar patch impact on normal path testcase

This test is designed to test the efficiency of the pthread_cond_wait() and
pthread_cond_signal() code paths within glibc. Ulrich has expressed concern
about these patches negatively impacting the non-pi condvar paths. I felt the
changes to the common case (when PI is not involved) were pretty minimal.
Certainly any significant regression there would be unacceptable.

I built three versions of glibc 2.11.1 from git:
1)	  git: unmodified git sources
2)     c_only: pthread_cond*.S files deleted
3) pi_condvar: same as c_only with the pi_condvar patches applied

Comparing #3 against #2 allows us to eliminate any gains #1 would have solely
from the hand written asm. 3 will eventually contain hand written asm, but
until the non-posix API is agreed upon, it doesn't make sense to expend the
effort of writing the asm code in my opinion.

I then ran 10 runs of 10M iterations each at SCHED_FIFO 1 priority on each of
the three glibcs, the results (following) suggest no significant change in the
non PI condvar performance, sitting right at ~270k (avg) cycles/sec for each
glibc.

build-x86_64-2.11.1-git
Cycles/Second: 279831.187500
Cycles/Second: 261911.421875
Cycles/Second: 277664.125000
Cycles/Second: 284847.718750
Cycles/Second: 285067.281250
Cycles/Second: 267918.718750
Cycles/Second: 284785.656250
Cycles/Second: 277402.843750
Cycles/Second: 202379.703125
Cycles/Second: 266421.718750

Min:  202379.703125  us
Max:  285067.28125  us
Avg:  268823.0375  us


build-x86_64-2.11.1-c_only
Cycles/Second: 277931.781250
Cycles/Second: 275614.093750
Cycles/Second: 271194.125000
Cycles/Second: 280155.093750
Cycles/Second: 284708.156250
Cycles/Second: 190936.031250
Cycles/Second: 264253.468750
Cycles/Second: 281354.281250
Cycles/Second: 290366.218750
Cycles/Second: 279962.000000

Min:  190936.03125  us
Max:  290366.21875  us
Avg:  269647.525  us


build-x86_64-2.11.1-pi_condvar
Cycles/Second: 263975.937500
Cycles/Second: 279577.281250
Cycles/Second: 276504.531250
Cycles/Second: 266163.562500
Cycles/Second: 262115.796875
Cycles/Second: 279219.406250
Cycles/Second: 265263.812500
Cycles/Second: 262226.468750
Cycles/Second: 284592.687500
Cycles/Second: 278975.875000

Min:  262115.796875  us
Max:  284592.6875  us
Avg:  271861.535938  us


This is only the cond_signal case, and doesn't account for cond_timedwait or
cond_broadcast, but I wouldn't expect those to experience any additional impact
from this patch. Are there scenarios you can think of that are likely to suffer
greater impact that are not covered by this rather simple test case?

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-11 18:50 -------
Created an attachment (id=4781)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4781&action=view)
pi condvar: add tst-condpi1.c basic API test


-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-11 18:50 -------
Created an attachment (id=4780)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4780&action=view)
pi condvar: pthread_condattr_setprotocol_np 2.12 port


-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug nptl/11588] pthread condvars are not priority inheritance aware

[dvhltc at us dot ibm dot com]

------- Additional Comments From dvhltc at us dot ibm dot com  2010-05-11 18:48 -------
Created an attachment (id=4779)
 --> (http://sourceware.org/bugzilla/attachment.cgi?id=4779&action=view)
Unbounded priority inversion testcase pthread_cond_hang.c

Using vanilla cond vars results in a pthread_cond_wait priority inversion
deadlock. With PI cond vars it works fine. This test uses dlsym to use the new
API if it exists.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11588

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.