Hi, Jerome Brunet <jbrunet@xxxxxxxxxxxx> writes: >> I am testing the three Ruslan's async feedback patches as modified by >> Jerome and I hit a regression in windows 10 enumeration. >> >> Ruslan posted an already accepted patch >> https://github.com/torvalds/linux/commit/789ea77310f0200c84002884ffd628e2baf3ad8a#diff-876615ece7fb56ce8d45fc70623bef9caa2548e810426f217fb785ffa10b33d5 >> which allowed win10 enumeration. >> >> Ruslan's async feedback patchset kept the change: >> https://patchwork.kernel.org/project/linux-usb/patch/1614603943-11668-5-git-send-email-ruslan.bilovol@xxxxxxxxx/ >> >> diff --git a/drivers/usb/gadget/function/f_uac2.c >> b/drivers/usb/gadget/function/f_uac2.c >> index 72b42f8..91b22fb 100644 >> --- a/drivers/usb/gadget/function/f_uac2.c >> +++ b/drivers/usb/gadget/function/f_uac2.c >> @@ -506,6 +506,10 @@ static int set_ep_max_packet_size(const struct >> f_uac2_opts *uac2_opts, >> >> max_size_bw = num_channels(chmask) * ssize * >> ((srate / (factor / (1 << (ep_desc->bInterval - 1)))) + 1); >> + >> + if (!is_playback && (uac2_opts->c_sync == USB_ENDPOINT_SYNC_ASYNC)) >> + max_size_bw = max_size_bw * FBACK_FREQ_MAX / 100; >> + >> ep_desc->wMaxPacketSize = cpu_to_le16(min_t(u16, max_size_bw, >> max_size_ep)); >> >> >> Jerome's rebase patch which was accepted recently changed the functionality >> to the original code: >> https://patchwork.kernel.org/project/linux-usb/patch/20210603220104.1216001-4-jbrunet@xxxxxxxxxxxx/ >> >> diff --git a/drivers/usb/gadget/function/f_uac2.c >> b/drivers/usb/gadget/function/f_uac2.c >> index 321e6c05ba93..ae29ff2b2b68 100644 >> --- a/drivers/usb/gadget/function/f_uac2.c >> +++ b/drivers/usb/gadget/function/f_uac2.c >> @@ -584,8 +584,11 @@ static int set_ep_max_packet_size(const struct >> f_uac2_opts *uac2_opts, >> ssize = uac2_opts->c_ssize; >> } >> >> + if (!is_playback && (uac2_opts->c_sync == USB_ENDPOINT_SYNC_ASYNC)) >> + srate = srate * (1000 + uac2_opts->fb_max) / 1000; >> + >> max_size_bw = num_channels(chmask) * ssize * >> - ((srate / (factor / (1 << (ep_desc->bInterval - 1)))) + 1); >> + DIV_ROUND_UP(srate, factor / (1 << (ep_desc->bInterval - 1))); >> ep_desc->wMaxPacketSize = cpu_to_le16(min_t(u16, max_size_bw, >> max_size_ep)); >> >> With this version my Win10 does not enumerate the USB AUDIO device, even if >> it has only EP-IN capability (i.e. is_playback = true). For my setup the >> EP-IN wMaxPacketSize is 192bytes vs. 196bytes in Ruslan's version, causing >> win10 reporting "The specified range could not be found in the range list." >> > > Maybe I am lacking USB expertize, but is there any reason why a 192bytes > maximum packet size should be considered invalid ? Just from your > comment, I can't figure it out. it sounds to me like one part of the descriptor claims 192 while another claims 196, then there is a mismatch and Windows is ignoring the interface. A quick dump of the descriptors would prove this. -- balbi
Attachment:
signature.asc
Description: PGP signature
