| 2019-04-08 15:46:46 |
dann frazier |
bug |
|
|
added bug |
| 2019-04-08 16:00:06 |
Ubuntu Kernel Bot |
linux (Ubuntu): status |
New |
Incomplete |
|
| 2019-04-08 19:16:12 |
dann frazier |
attachment added |
|
dmesg w/ CMA_DEBUG=y https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1823753/+attachment/5254279/+files/dmesg-cma-debug.txt |
|
| 2019-04-08 21:03:27 |
dann frazier |
linux (Ubuntu): status |
Incomplete |
Confirmed |
|
| 2019-04-11 07:30:34 |
Andrew Cloke |
bug |
|
|
added subscriber Andrew Cloke |
| 2019-04-17 21:16:25 |
dann frazier |
bug |
|
|
added subscriber Fred Kimmy |
| 2019-04-25 18:57:02 |
Seth Forshee |
bug |
|
|
added subscriber Seth Forshee |
| 2019-05-03 14:06:29 |
Terry Rudd |
bug |
|
|
added subscriber Terry Rudd |
| 2019-05-03 14:10:21 |
Terry Rudd |
linux (Ubuntu): importance |
Undecided |
High |
|
| 2019-05-14 02:29:55 |
Guanhua Ye |
bug |
|
|
added subscriber Guanhua Ye |
| 2019-05-30 21:08:24 |
dann frazier |
nominated for series |
|
Ubuntu Eoan |
|
| 2019-05-30 21:08:24 |
dann frazier |
bug task added |
|
linux (Ubuntu Eoan) |
|
| 2019-05-30 21:08:24 |
dann frazier |
nominated for series |
|
Ubuntu Disco |
|
| 2019-05-30 21:08:24 |
dann frazier |
bug task added |
|
linux (Ubuntu Disco) |
|
| 2019-05-30 21:08:32 |
dann frazier |
linux (Ubuntu Disco): status |
New |
In Progress |
|
| 2019-05-30 21:08:35 |
dann frazier |
linux (Ubuntu Eoan): status |
Confirmed |
In Progress |
|
| 2019-05-30 21:08:38 |
dann frazier |
linux (Ubuntu Disco): assignee |
|
dann frazier (dannf) |
|
| 2019-05-30 21:08:40 |
dann frazier |
linux (Ubuntu Disco): importance |
Undecided |
High |
|
| 2019-05-30 21:08:41 |
dann frazier |
linux (Ubuntu Eoan): assignee |
|
dann frazier (dannf) |
|
| 2019-06-04 20:44:12 |
dann frazier |
description |
On some arm64 systems[*] we are seeing a spew of messages on the console:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
This appears to be non-fatal - impacted systems all eventually boot. But, at least in the case of the HP m400, it slows down boot enough that MAAS' default timeout will expire before completing deployment.
[*] Observed on a HiSilicon D06 w/ SMMU disabled in the BIOS, as well as an HP m400 (APM X-Gene) cartridge - although, not on another one that - in theory - should be identical. |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
where other systems began spewing on the order of 10K of these messages on boot:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
|
| 2019-06-04 20:44:31 |
dann frazier |
description |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
where other systems began spewing on the order of 10K of these messages on boot:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing on the order of 10K of these messages on boot:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
|
| 2019-06-04 20:44:53 |
dann frazier |
description |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing on the order of 10K of these messages on boot:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
|
| 2019-06-04 20:45:29 |
dann frazier |
description |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements, but we should track those in other bugs.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
|
| 2019-06-04 21:01:06 |
dann frazier |
description |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements, but we should track those in other bugs.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64) |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements, but we should track those in other bugs.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64)
Test builds are staged at ppa:dannf/cma. |
|
| 2019-06-04 21:01:17 |
dann frazier |
description |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements, but we should track those in other bugs.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64)
Test builds are staged at ppa:dannf/cma. |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements, but we should track those in other bugs.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64)
Test builds are staged in ppa:dannf/cma |
|
| 2019-06-04 21:02:06 |
dann frazier |
description |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements, but we should track those in other bugs.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64)
Test builds are staged in ppa:dannf/cma |
[Impact]
We enabled CONFIG_DMA_CMA to fix bug 1803206, but that led to a regression
on other arm64 systems that began spewing these messages on boot - sometimes > 10K of them:
[ 19.534097] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534109] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534113] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534126] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534130] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534142] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534146] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534157] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534161] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
[ 19.534173] cma: cma_alloc: alloc failed, req-size: 16 pages, ret: -12
[ 19.534177] cma: cma_alloc: alloc failed, req-size: 64 pages, ret: -12
In a previous SRU (bug 1828092), we worked around this by just rate-limiting these messages. These are "err" priority messages though, so even a lower number of them is still disconcerting.
[Fix]
1) Bump up the amount of available CMA on arm64 to 32M (same as upstream defconfig)
2) A patch-set from linux-next that redirects dma-direct contiguous allocations to alloc_pages() for single page allocations (single pages are by definition contiguous), avoiding CMA usage/fragmentation.
[Test Case]
dmesg | grep "cma_alloc: alloc failed"
Some system configs will still have some of these errors even after this fix - but this should reduce them significantly. Per-driver optimizations can be used to make further improvements, but we should track those in other bugs.
[Regression Risk]
Tested on a HiSilicon D06 and HP m400 (Hi1620 & X-Gene arm64).
Regression tested on:
- Raspberry Pi 3B (see Comment #22)
- Power9 system (ppc64el)
- z/VM instance (s390x)
- Intel Centerton system (amd64)
Test builds are staged at http://ppa.launchpad.net/~dannf/ubuntu/cma |
|
| 2019-06-07 20:25:28 |
Seth Forshee |
linux (Ubuntu Eoan): status |
In Progress |
Fix Committed |
|
| 2019-07-01 08:44:31 |
Kleber Sacilotto de Souza |
linux (Ubuntu Disco): status |
In Progress |
Fix Committed |
|
| 2019-07-03 11:02:48 |
Ubuntu Kernel Bot |
tags |
|
verification-needed-disco |
|
| 2019-07-10 22:23:32 |
dann frazier |
tags |
verification-needed-disco |
verification-done-disco |
|
| 2019-07-15 22:40:05 |
Launchpad Janitor |
linux (Ubuntu Eoan): status |
Fix Committed |
Fix Released |
|
| 2019-07-23 05:25:24 |
Launchpad Janitor |
linux (Ubuntu Disco): status |
Fix Committed |
Fix Released |
|
| 2019-07-23 05:25:24 |
Launchpad Janitor |
cve linked |
|
2018-12126 |
|
| 2019-07-23 05:25:24 |
Launchpad Janitor |
cve linked |
|
2018-12127 |
|
| 2019-07-23 05:25:24 |
Launchpad Janitor |
cve linked |
|
2018-12130 |
|
| 2019-07-23 05:25:24 |
Launchpad Janitor |
cve linked |
|
2019-11091 |
|
| 2019-07-23 05:25:24 |
Launchpad Janitor |
cve linked |
|
2019-11833 |
|
| 2019-07-23 05:25:24 |
Launchpad Janitor |
cve linked |
|
2019-11884 |
|
| 2019-07-24 20:59:11 |
Brad Figg |
tags |
verification-done-disco |
cscc verification-done-disco |
|
