-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Simon Kagstrom wrote:
> However, I think the spi_async() method should do quite a bit better > than this. Instead of basically doing 5 queue_work()s per interrupt > (the interrupt itself, plus reading of status,x,y,z), it would do just > one. Well 2000 work queue events a second just for reading accels would be a bit much. 200 interrupts plus 200 workqueue events is better. But, I wonder what it does under load when the scheduler is full and what the latency is for those 200 slots we are asking for then. The chips do not have any FIFO, you have to turn up and read them - both of them, asynchronously - before 10ms or the sample is lost. The chips don't give a toss "how you are meant to do SPI", they demand service in their window or there is data loss. Anyway thanks again for looking into this, we can at least stress it when it's done and see if the issue is real or not. - -Andy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAkjwdNkACgkQOjLpvpq7dMpbxQCgj/b72S29z8Tuk2TwWgojqzap 9yYAoJGcfcvQSyLXOT0SuEHrO1PwsI+l =D0og -----END PGP SIGNATURE-----
