#5278: On modern fedora 64 installs, sage exists frequently when omalloc things
there is no memory left
---------------------+------------------------------------------------------
Reporter: was | Owner: cwitty
Type: defect | Status: new
Priority: blocker | Milestone: sage-3.3
Component: misc | Resolution:
Keywords: |
---------------------+------------------------------------------------------
Old description:
> This is show-stopper bug for using Sage on any modern install of Fedora.
> It was discussed a lot in sage-devel.
>
> To replicate this issue, install Sage on fedora 64-bit and run a doctest
> a few times. Every so often (maybe 20% chance) it will exit with this
> error:
> {{{
> error: no more memory
> System 0k:0k Appl 0k/0k Malloc 0k/0k Valloc 0k/0k Pages 0/0 Regions 0:0
> }}}
>
> Tracing this through with print statements I found that this happens
> around line 370 of singular.spkg's omalloc/omAllocSystem.c (something
> similar was posted to sage-devel):
> {{{
> void* omAllocFromSystem(size_t size)
> {
> void* ptr;
>
> ptr = OM_MALLOC_FROM_SYSTEM(size);
> if (ptr == NULL)
> {
> OM_MEMORY_LOW_HOOK();
> ptr = OM_MALLOC_FROM_SYSTEM(size);
> if (ptr == NULL)
> BOOM (crash handler called!)
> }}}
>
> The first thing that is suspicious is that OM_MALLOC_FROM_SYSTEM is
> called, fails, and then is called again... for now reason. That is very
> hackish code. Tracing that through it seemed to me that
> OM_MALLOC_FROM_SYSTEM should just be the system malloc.
> To test that theory I modified the above code as follows:
> void* omAllocFromSystem(size_t size)
> {
> void* ptr;
>
> ptr = OM_MALLOC_FROM_SYSTEM(size);
> if (ptr == NULL)
> {
> OM_MEMORY_LOW_HOOK();
> ptr = OM_MALLOC_FROM_SYSTEM(size);
> if (ptr == NULL)
> ptr = malloc(size);
> if (ptr == NULL)
> BOOM (crash handler called!)
> }}}
>
> After making that change, the problem vanishes (for me) and doctesting
> works fine again, at least for me. I haven't tested doctesting the whole
> tree.
New description:
This is show-stopper bug for using Sage on any modern install of Fedora.
It was discussed a lot in sage-devel.
To replicate this issue, install Sage on fedora 64-bit and run a doctest a
few times. Every so often (maybe 20% chance) it will exit with this
error:
{{{
error: no more memory
System 0k:0k Appl 0k/0k Malloc 0k/0k Valloc 0k/0k Pages 0/0 Regions 0:0
}}}
Tracing this through with print statements I found that this happens
around line 370 of singular.spkg's omalloc/omAllocSystem.c (something
similar was posted to sage-devel):
{{{
void* omAllocFromSystem(size_t size)
{
void* ptr;
ptr = OM_MALLOC_FROM_SYSTEM(size);
if (ptr == NULL)
{
OM_MEMORY_LOW_HOOK();
ptr = OM_MALLOC_FROM_SYSTEM(size);
if (ptr == NULL)
BOOM (crash handler called!)
}}}
The first thing that is suspicious is that OM_MALLOC_FROM_SYSTEM is
called, fails, and then is called again... for now reason. That is very
hackish code. Tracing that through it seemed to me that
OM_MALLOC_FROM_SYSTEM should just be the system malloc.
To test that theory I modified the above code as follows:
{{{
void* omAllocFromSystem(size_t size)
{
void* ptr;
ptr = OM_MALLOC_FROM_SYSTEM(size);
if (ptr == NULL)
{
OM_MEMORY_LOW_HOOK();
ptr = OM_MALLOC_FROM_SYSTEM(size);
if (ptr == NULL)
ptr = malloc(size);
if (ptr == NULL)
BOOM (crash handler called!)
}}}
After making that change, the problem vanishes (for me) and doctesting
works fine again, at least for me. I haven't tested doctesting the whole
tree.
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/5278#comment:1>
Sage <http://sagemath.org/>
Sage - Open Source Mathematical Software: Building the Car Instead of
Reinventing the Wheel
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en
-~----------~----~----~----~------~----~------~--~---
