-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Somebody in the thread at some point said: | On Fri, Feb 20, 2009 at 10:04:21AM +0000, Andy Green wrote: |> This patch moves the bulk transfer action outside of |> interrupt context, along with the STOP transmission action |> for multiblock transfers. |> |> It's prompted by |> |> https://docs.openmoko.org/trac/ticket/2180 | | As far as I can tell, the patch fixes the audio breakups on SDcard writes | that I reported on the hardware mailing list (subject: "Re: cpu usage way | too high (?)"). Tested with mikmod, xsidplay and xmms2.
Great, it makes sense. But does it seem to reduce SD throughput significantly? It introduces some delay before it drags back the data from Glamo, but on the other hand there is a new scatter-gather combining config option used recently to increase the size of these multiblock transfers more often, reducing that hit. If we're lucky it doesn't really impact throughput typically, what's your opinion from running it? - -Andy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAkmgkrgACgkQOjLpvpq7dMoLMgCaAxIkftI2XEq9Gt+D0VoqlXP+ dasAn1h49bU/5Ir7ja8W1y6gbf1ATByp =YukI -----END PGP SIGNATURE-----
