Disabled alsa-restore.service and apport.service, not much else left to do.

How much memory do you think you have? Would it help if we say dumped /proc/slabinfo if the check fails?

(In reply to Chris Wilson from comment #2)
> How much memory do you think you have? Would it help if we say dumped
> /proc/slabinfo if the check fails?

Having both the total amount of memory and the output of /proc/slabinfo would be beneficial, yes!

Wanna write a patch for it?

(In reply to Martin Peres from comment #3)
> (In reply to Chris Wilson from comment #2)
> > How much memory do you think you have? Would it help if we say dumped
> > /proc/slabinfo if the check fails?
> 
> Having both the total amount of memory and the output of /proc/slabinfo
> would be beneficial, yes!
> 
> Wanna write a patch for it?

If so, make sure to add it to the lib/, so as any test that allocates loads of memory could use the helper :)

More info in future:

commit e5829165c2ef96140c222429c0f174fb72c4324b (upstream/master)
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Apr 10 13:16:16 2018 +0100

    lib: Dump meminfo and slabinfo if we complain about insufficient memory
    
    All too frequently, we fail our memory checks to a leak in the driver.
    While we give every opportunity for the driver to release the memory
    before we fail, if we do dump the meminfo and slabinfo, if available, so
    we can assign blame^W^W resolve the problem quickly.
    
    References: https://bugs.freedesktop.org/show_bug.cgi?id=105967
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Cc: Martin Peres <martin.peres@linux.intel.com>
    Cc: Tomi Sarvela <tomi.p.sarvela@intel.com>
    Reviewed-by: Antonio Argenziano <antonio.argenziano@intel.com>

intel_os-WARNING: Insufficient free memory; /proc/meminfo:
MemTotal:         965812 kB
MemFree:          520256 kB
MemAvailable:     510932 kB
Buffers:             184 kB
Cached:            97664 kB
SwapCached:            0 kB
Active:           188800 kB
Inactive:         161496 kB
Active(anon):     128000 kB
Inactive(anon):   125308 kB
Active(file):      60800 kB
Inactive(file):    36188 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:       1030140 kB
SwapFree:        1030140 kB
Dirty:               132 kB
Writeback:             0 kB
AnonPages:        252476 kB
Mapped:            96804 kB
Shmem:               856 kB
Slab:              67848 kB
SReclaimable:      26888 kB
SUnreclaim:        40960 kB
KernelStack:        2512 kB
PageTables:         6080 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:     1513044 kB
Committed_AS:     538144 kB
VmallocTotal:   34359738367 kB
VmallocUsed:           0 kB
VmallocChunk:          0 kB
AnonHugePages:     75776 kB
ShmemHugePages:        0 kB
ShmemPmdMapped:        0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
Hugetlb:               0 kB
DirectMap4k:       81356 kB
DirectMap2M:      950272 kB

Fingers crossed the situation improves:

commit 778497e7965dc8662c770a89ebbd741778feb71e (HEAD, upstream/master)
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Fri Jun 15 14:31:19 2018 +0100

    lib: Purge os caches twice for good luck
    
    If at first we don't succeed, try, try again.
    
    References: https://bugs.freedesktop.org/show_bug.cgi?id=105967
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

And my hopes were dashed. Still sporadic borderline failures.

In today's iteration, give in and include "buffers" as free memory. Buffers corresponds to nr_blockdev_pages() so should just be pages backing file data, and easily disposable...

commit 9ab9268fa7eeda0a7ea6eb2ab02bb6c5b9c91ba0 (upstream/master, origin/master, origin/HEAD)
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Jun 19 10:17:40 2018 +0100

    lib: Conservatively include residual buffers in the available ram estimate
    
    Add any buffers reported by sysinfo to the estimate of available memory.
    We do ask the kernel to purge it's caches before reporting sysinfo, but
    a few remain that may be forced out by our test usage, so include them.
    However, be conservative and only allow them to be swapped out.
    
    References: https://bugs.freedesktop.org/show_bug.cgi?id=105967
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

I proclaim the problem swept under the carpet.

(In reply to Chris Wilson from comment #10)
> I proclaim the problem swept under the carpet.

Let's hope so! Thanks!

Sadly premature, but I have a new plan https://patchwork.freedesktop.org/series/45176/