Kris\Slyka on Nostr: I know it's a long shot, but is anyone here familiar with the mechanics involved in ...
I know it's a long shot, but is anyone here familiar with the mechanics involved in the scheduling of isochronous USB transfers in the linux kernel? I'm running into some weird problems where the URBs fail to submit because the EHCI scheduler reports that I'm out of bandwidth depending on which order they're submitted in. This is specifically related to USB audio, where sometimes the ISO OUT URBs are submitted before the ISO IN ones when starting a full duplex stream, in which case it works fine, but for some reason it fails with `cannot submit urb 0, error -28: not enough bandwidth` when submitting them the other way around.
#linuxkernel #kerneldev
Published at
2023-08-04 14:26:13Event JSON
{
"id": "2dd092d9b5796ab2b05bfbdea4a3b66c6e8806063d5d1bf5be6653e18dd5469b",
"pubkey": "d1042f75585bc5ecb6fe3185f7bd4c3a75fbfba76011e86ef2bc221eb56f7195",
"created_at": 1691159173,
"kind": 1,
"tags": [
[
"t",
"linuxkernel"
],
[
"t",
"kerneldev"
],
[
"proxy",
"https://pawb.fun/users/slyka/statuses/110831807618756187",
"activitypub"
]
],
"content": "I know it's a long shot, but is anyone here familiar with the mechanics involved in the scheduling of isochronous USB transfers in the linux kernel? I'm running into some weird problems where the URBs fail to submit because the EHCI scheduler reports that I'm out of bandwidth depending on which order they're submitted in. This is specifically related to USB audio, where sometimes the ISO OUT URBs are submitted before the ISO IN ones when starting a full duplex stream, in which case it works fine, but for some reason it fails with `cannot submit urb 0, error -28: not enough bandwidth` when submitting them the other way around.\n\n#linuxkernel #kerneldev",
"sig": "dd70d3002728ed3f968db214083f9285eb3f1cafc8b0fc6b31bca1e080d0787bf0d93645d0828a9463d67706ce82a79d376b05b554a7c2e248a74acf3fc41740"
}
