Skip to contentSkip to conversation

mozjs/unhandled signal 11/64k page patch/__powerpc64__ ???

DISCUSSION Robert P. J. Day 3 participants 9 messages Nov 23, 2016 – Nov 24, 2016

lore
Share & tools
lore
#0 lore raw

i know nothing about this issue, i'm asking on behalf of a colleague
who has allegedly tracked down a bug in wind river linux, and comparing sources, it *appears* the very same thing would happen in OE.

the symptom running "poweroff" in a qemuppc session;

... snip ...
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001 ... snip ...

further investigation shows the very same thing happening with:

# systemctl start httpd
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001
Error getting authority: Error initializing authority: Error calling
StartServiceByName for org.freedesktop.PolicyKit1: Timeout was reached
(g-io-error-quark, 24)
Failed to start httpd.service: Connection timed out
#

to make a long story short, said colleague followed bug reports and
ended up here:

https://bugzilla.redhat.com/show_bug.cgi?id=1289432

eventually, he got the problem to go away by tweaking the patch "0005-aarch64-64k-page.patch" for mozjs with the following adjustment:

@@ -113,7 +113,7 @@ struct Cell
#if defined(SOLARIS) && (defined(__sparc) || defined(__sparcv9)) const size_t PageShift = 13;
const size_t ArenaShift = PageShift;
-#elif defined(__powerpc__)
+#elif defined(__powerpc__) || defined(__aarch64__)

               ^^^^^^^^^^^ change that to __powerpc64__

const size_t PageShift = 16;
const size_t ArenaShift = 12;
#else

as my colleague reads it, the original patch above is necessary for
PPC64, but *breaks* PPC32. i haven't followed the logic, but does any of this make sense? by making that change, the error went away and things work fine.

thoughts?

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
#1 lore raw

can you reproduce it on version 24 as well ?

Quoted 67 lines · Robert P. J. Day

i know nothing about this issue, i'm asking on behalf of a colleague
who has allegedly tracked down a bug in wind river linux, and comparing sources, it *appears* the very same thing would happen in OE.

the symptom running "poweroff" in a qemuppc session;

... snip ...
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001 ... snip ...

further investigation shows the very same thing happening with:

# systemctl start httpd
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001
Error getting authority: Error initializing authority: Error calling
StartServiceByName for org.freedesktop.PolicyKit1: Timeout was reached
(g-io-error-quark, 24)
Failed to start httpd.service: Connection timed out
#

to make a long story short, said colleague followed bug reports and
ended up here:

https://bugzilla.redhat.com/show_bug.cgi?id=1289432

eventually, he got the problem to go away by tweaking the patch "0005-aarch64-64k-page.patch" for mozjs with the following adjustment:

(quoted diff) +1 -1
@@ -113,7 +113,7 @@ struct Cell
 #if defined(SOLARIS) && (defined(__sparc) || defined(__sparcv9))
 const size_t PageShift = 13;
 const size_t ArenaShift = PageShift;
-#elif defined(__powerpc__)
+#elif defined(__powerpc__) || defined(__aarch64__)
              ^^^^^^^^^^^ change that to __powerpc64__
 const size_t PageShift = 16;
 const size_t ArenaShift = 12;

#else

as my colleague reads it, the original patch above is necessary for
PPC64, but *breaks* PPC32. i haven't followed the logic, but does any of this make sense? by making that change, the error went away and things work fine.

thoughts?

rday

Signature · 13 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                       http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

-- 
_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-devel
#2 lore raw
Quoted 1 line · Khem Raj

can you reproduce it on version 24 as well ?

not sure what you mean ... this is running on a qemuppc image. i
haven't tried running this on fedora 24 itself. what exactly do you want me to try?

Quoted 1 line
Quoted 1 line

On Nov 23, 2016, at 10:34 AM, Robert P. J. Day

Quoted 48 lines · <[email protected]> wrote:
Quoted 48 lines

i know nothing about this issue, i'm asking on behalf of a colleague
who has allegedly tracked down a bug in wind river linux, and comparing sources, it *appears* the very same thing would happen in OE.

the symptom running "poweroff" in a qemuppc session;

... snip ...
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001 ... snip ...

further investigation shows the very same thing happening with:

# systemctl start httpd
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001
Error getting authority: Error initializing authority: Error calling
StartServiceByName for org.freedesktop.PolicyKit1: Timeout was reached
(g-io-error-quark, 24)
Failed to start httpd.service: Connection timed out
#

to make a long story short, said colleague followed bug reports and
ended up here:

https://bugzilla.redhat.com/show_bug.cgi?id=1289432

eventually, he got the problem to go away by tweaking the patch "0005-aarch64-64k-page.patch" for mozjs with the following adjustment:

(quoted diff) +1 -1
@@ -113,7 +113,7 @@ struct Cell
 #if defined(SOLARIS) && (defined(__sparc) || defined(__sparcv9))
 const size_t PageShift = 13;
 const size_t ArenaShift = PageShift;
-#elif defined(__powerpc__)
+#elif defined(__powerpc__) || defined(__aarch64__)
              ^^^^^^^^^^^ change that to __powerpc64__
 const size_t PageShift = 16;
 const size_t ArenaShift = 12;

#else

as my colleague reads it, the original patch above is necessary for
PPC64, but *breaks* PPC32. i haven't followed the logic, but does any of this make sense? by making that change, the error went away and things work fine.

thoughts?

#3 lore raw
Quoted 11 lines · Robert P. J. Day
Quoted 1 line · Khem Raj

can you reproduce it on version 24 as well ?

not sure what you mean ... this is running on a qemuppc image. i
haven't tried running this on fedora 24 itself. what exactly do you want me to try?

I meant version of mozjs

Quoted 56 lines
Quoted 1 line
Quoted 1 line

On Nov 23, 2016, at 10:34 AM, Robert P. J. Day

Quoted 48 lines · <[email protected]> wrote:
Quoted 48 lines

i know nothing about this issue, i'm asking on behalf of a colleague who has allegedly tracked down a bug in wind river linux, and comparing sources, it *appears* the very same thing would happen in OE.

the symptom running "poweroff" in a qemuppc session;

... snip ...
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001 ... snip ...

further investigation shows the very same thing happening with:

# systemctl start httpd
polkitd[680]: unhandled signal 11 at 00000000 nip 0fa69198 lr 0fa69174code 30001
Error getting authority: Error initializing authority: Error calling
StartServiceByName for org.freedesktop.PolicyKit1: Timeout was reached
(g-io-error-quark, 24)
Failed to start httpd.service: Connection timed out
#

to make a long story short, said colleague followed bug reports and ended up here:

https://bugzilla.redhat.com/show_bug.cgi?id=1289432

eventually, he got the problem to go away by tweaking the patch "0005-aarch64-64k-page.patch" for mozjs with the following adjustment:

(quoted diff) +1 -1
@@ -113,7 +113,7 @@ struct Cell
 #if defined(SOLARIS) && (defined(__sparc) || defined(__sparcv9))
 const size_t PageShift = 13;
 const size_t ArenaShift = PageShift;
-#elif defined(__powerpc__)
+#elif defined(__powerpc__) || defined(__aarch64__)
             ^^^^^^^^^^^ change that to __powerpc64__
 const size_t PageShift = 16;
 const size_t ArenaShift = 12;

#else

as my colleague reads it, the original patch above is necessary for PPC64, but *breaks* PPC32. i haven't followed the logic, but does any of this make sense? by making that change, the error went away and things work fine.

thoughts?

Signature · 4 lines
_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-devel
#4 lore raw
Quoted 5 lines · Khem Raj
Quoted 5 lines · Robert P. J. Day

On Wed, 23 Nov 2016, Khem Raj wrote:

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

can you reproduce it on version 24 as well ?

not sure what you mean ... this is running on a qemuppc image. i
haven't tried running this on fedora 24 itself. what exactly do you want me to try?

I meant version of mozjs

this is version 17.

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
#5 lore raw
Quoted 12 lines · Robert P. J. Day
Quoted 7 lines · Khem Raj
Quoted 7 lines · Robert P. J. Day
Quoted 1 line · Khem Raj

can you reproduce it on version 24 as well ?

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

not sure what you mean ... this is running on a qemuppc image. i haven't tried running this on fedora 24 itself. what exactly do you want me to try?

I meant version of mozjs

this is version 17.

OK, I think 24 has such issues taken care of. I am not near code right now. but it will be good to check if something is already fixed there for this

Quoted 18 lines

rday

Signature · 13 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                       http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================

-- 
_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-devel
#6 lore raw
Quoted 17 lines · Khem Raj
Quoted 17 lines · Robert P. J. Day
Quoted 12 lines · Khem Raj
Quoted 11 lines · Robert P. J. Day
Quoted 1 line · Khem Raj

can you reproduce it on version 24 as well ?

not sure what you mean ... this is running on a qemuppc image. i haven't tried running this on fedora 24 itself. what exactly do you want me to try?

Quoted 7 lines
Quoted 3 lines
Quoted 1 line

I meant version of mozjs

this is version 17.

OK, I think 24 has such issues taken care of. I am not near code right now. but it will be good to check if something is already fixed there for this

ok, i'll take a look, and if it *is* fixed in 24, is there a reason
meta-openembedded/meta-oe still contains only mozjs_17.0.0.bb? would it not make sense to bump that up?

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
#7 lore raw

Op 24-11-16 om 09:58 schreef Robert P. J. Day:

Quoted 23 lines · Khem Raj
Quoted 22 lines
Quoted 21 lines · Robert P. J. Day
Quoted 5 lines · Khem Raj
Quoted 5 lines · Robert P. J. Day

On Wed, 23 Nov 2016, Khem Raj wrote:

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

can you reproduce it on version 24 as well ?

not sure what you mean ... this is running on a qemuppc image. i haven't tried running this on fedora 24 itself. what exactly do you want me to try?

I meant version of mozjs

this is version 17.

Quoted 7 lines
Quoted 3 lines

OK, I think 24 has such issues taken care of. I am not near code right now. but it will be good to check if something is already fixed there for this

ok, i'll take a look, and if it *is* fixed in 24, is there a reason
meta-openembedded/meta-oe still contains only mozjs_17.0.0.bb? would it not make sense to bump that up?

mozjs has no stable API, so 17, 24 and up are all different and incompatible :(

#8 lore raw
In reply to Khem Raj
Quoted 17 lines · Khem Raj
Quoted 17 lines · Robert P. J. Day
Quoted 12 lines · Khem Raj
Quoted 11 lines · Robert P. J. Day
Quoted 1 line · Khem Raj

can you reproduce it on version 24 as well ?

not sure what you mean ... this is running on a qemuppc image. i haven't tried running this on fedora 24 itself. what exactly do you want me to try?

Quoted 7 lines
Quoted 3 lines
Quoted 1 line

I meant version of mozjs

this is version 17.

OK, I think 24 has such issues taken care of. I am not near code right now. but it will be good to check if something is already fixed there for this

is it just me, or is the SRC_URI in mozjs_17.0.0.bb a bit weird?

SRC_URI = " \
    http://ftp.mozilla.org/pub/mozilla.org/js/${BPN}${PV}.tar.gz \
                               ^^^^^^^^^^^ ???

there's no such directory structure.

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================