Android Oreo, Sabre IMX6Q/QP, PFNs busy

Showing results for 
Show  only  | Search instead for 
Did you mean: 

Android Oreo, Sabre IMX6Q/QP, PFNs busy

Contributor IV

I am using the SABRE kit (I have one with i.MX6Q and another with i.MX6QP) and I am using the last Android release (Oreo 8.0.0) released by NXP.

When using the system, for example going to the settings menu, I get a lot of graphical lag, and these errors in the kernel log:

[ 263.819436] alloc_contig_range: [32e00, 335ea) PFNs busy
[ 263.863451] alloc_contig_range: [37200, 379ea) PFNs busy
[ 263.870269] alloc_contig_range: [37200, 37aea) PFNs busy
[ 263.876422] alloc_contig_range: [37400, 37bea) PFNs busy
[ 263.882981] alloc_contig_range: [37400, 37cea) PFNs busy
[ 263.889460] alloc_contig_range: [37600, 37dea) PFNs busy
[ 263.979731] alloc_contig_range: [32e00, 335ea) PFNs busy
[ 263.997406] alloc_contig_range: [37f00, 386ea) PFNs busy
[ 264.008030] alloc_contig_range: [38000, 387ea) PFNs busy
[ 264.016250] alloc_contig_range: [38100, 388ea) PFNs busy
[ 264.035584] alloc_contig_range: [38200, 389ea) PFNs busy
[ 264.274224] alloc_contig_range: [32e00, 335ea) PFNs busy
[ 264.285375] alloc_contig_range: [37f00, 386ea) PFNs busy
[ 264.304507] alloc_contig_range: [38000, 387ea) PFNs busy
[ 264.311177] alloc_contig_range: [38100, 388ea) PFNs busy
[ 264.317586] alloc_contig_range: [38200, 389ea) PFNs busy

Is this a known problem? Are there any fixes available for this?

0 Kudos
2 Replies

NXP Employee
NXP Employee


This is an expected behavior. This is due to the lack of memory that the system has, an example could be the i.MX8M-EVK where the memory is higher than the i.MX6 en the android oreo image runs well. Unfortunately, the only solution could be through hardware. I apologize for the inconveniences this could give you.

Best Regards,


0 Kudos

Contributor III

I don't think it's due to the lack of memory,  I'm running a very simple example in imx yocto GA release, in such circumstances the total used memory is around 200MB only and my system has 2GB memory but this message still prints. I guess as CMA is a SHARED memory pool, there might be some bugs on memory management.

0 Kudos