USB:UAC2: Incorrect req->length > maxpacket*mc
Hi,
Together with dwc2 maintainer Minas Harutyunyan we have been troubleshooting various issues of dwc2 on RPi4. We hit a problem where the g_audio in capture (EP OUT, playback from USB host) requests req-> length larger than maxpacket*mc.
As a workaround we removed the check in dwc2/gadget.c, however that is
not a proper solution. Minas with his team decided to add a patch where dwc2 will rejecting this type of wrong requests in dwc2_hsotg_ep_queue() function, to not process these request at all. The f_uac2 + g_audio
gadget should restrain from sending such requests.
Steps to reproduce:
* Changing fs_epout_desc.bInterval in f_uac2.c from 4 (1ms) to 1 (125us) - the goal is to maximize available throughput of the audio gadget
* Loading the g_audio module with c_srate=48000, c_ssize=2, c_chmask=2 - i.e. standard 48kHz/16bit/2ch USB playback -> alsa capture
This combination produces mps=24 and mc=1 for EP OUT. Yet the audio function driver sometimes queues request with req->length 192.
Please may I ask for fixing the audio function in this respect? Unfortunately that is way above my knowledge of that code/requirements. However, I can send debug dump for u_audio.c, f_uac2.c, gadget.c if needed.
Thanks a lot in advance.
Best regards,
Pavel.
PATCH 1/1 usb: gadget: f_uac2: fix packet size calculation
Quoted 13 lines · On Fri, 10 Jan 2020 08:29:06 +0100 Pavel Hofman <[email protected]>…
Together with dwc2 maintainer Minas Harutyunyan we have been troubleshooting various issues of dwc2 on RPi4. We hit a problem where the
g_audioin capture (EP OUT, playback from USB host) requests req-> length larger than maxpacket*mc.As a workaround we removed the check in
dwc2/gadget.c, however that is not a proper solution. Minas with his team decided to add a patch where dwc2 will rejecting this type of wrong requests indwc2_hsotg_ep_queue()function, to not process these request at all. Thef_uac2+g_audiogadget should restrain from sending such requests.Steps to reproduce:
Quoted 8 lines
* Changing fs_epout_desc.bInterval in
f_uac2.cfrom 4 (1ms) to 1 (125us)
- the goal is to maximize available throughput of the audio gadget* Loading the
g_audiomodule with c_srate=48000, c_ssize=2, c_chmask=2 - i.e. standard 48kHz/16bit/2ch USB playback -> alsa captureThis combination produces mps=24 and mc=1 for EP OUT. Yet the audio function driver sometimes queues request with
req->length192.
This sounded familiar to me, and I realised that I've been running with a patch for this case that I never submitted upstream.
Does the patch below fix the issue you're seeing?
-- >8 --
The packet size for USB audio must always be a multiple of the frame size, otherwise we are transmitting a partial frame which omits some channels (and these end up at the wrong offset in the next packet). Furthermore, it breaks the residue handling such that we end up trying to send a packet exceeding the maximum packet size for the endpoint.
--- drivers/usb/gadget/function/u_audio.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)
drivers/usb/gadget/function/u_audio.c +6 -4
@@ -407,7 +407,7 @@ int u_audio_start_playback(struct g_audio *audio_dev)struct usb_ep *ep;struct uac_rtd_params *prm;struct uac_params *params = &audio_dev->params; -unsigned int factor, rate; +unsigned int factor;const struct usb_endpoint_descriptor *ep_desc;int req_len, i;
@@ -426,13 +426,15 @@ int u_audio_start_playback(struct g_audio *audio_dev)/* pre-compute some values for iso_complete() */ uac->p_framesize = params->p_ssize *num_channels (params->p_chmask); - rate = params->p_srate * uac->p_framesize; uac->p_interval = factor / (1 << (ep_desc->bInterval -1 )); - uac->p_pktsize = min_t(unsigned int, rate / uac->p_interval, + uac->p_pktsize = min_t(unsigned int, + uac->p_framesize * + (params->p_srate / uac->p_interval), prm->max_psize);if (uac->p_pktsize < prm->max_psize) - uac->p_pktsize_residue = rate % uac->p_interval; + uac->p_pktsize_residue = uac->p_framesize * + (params->p_srate % uac->p_interval);else uac->p_pktsize_residue =0 ;
Signature · 1 line
2.24.1
Re: [PATCH] usb: gadget: f_uac2: fix packet size calculation
Hi,
Dne 10. 01. 20 v 12:28 John Keeping napsal(a):
Review int u_audio_start_playback(struct g_audio *audio_dev)
Quoted 30 lines
Quoted 13 lines · On Fri, 10 Jan 2020 08:29:06 +0100 Pavel Hofman <[email protected]>…
Together with dwc2 maintainer Minas Harutyunyan we have been troubleshooting various issues of dwc2 on RPi4. We hit a problem where the
g_audioin capture (EP OUT, playback from USB host) requests req-> length larger than maxpacket*mc.As a workaround we removed the check in
dwc2/gadget.c, however that is not a proper solution. Minas with his team decided to add a patch where dwc2 will rejecting this type of wrong requests indwc2_hsotg_ep_queue()function, to not process these request at all. Thef_uac2+g_audiogadget should restrain from sending such requests.Steps to reproduce:
Quoted 8 lines
* Changing fs_epout_desc.bInterval in
f_uac2.cfrom 4 (1ms) to 1 (125us)
- the goal is to maximize available throughput of the audio gadget* Loading the
g_audiomodule with c_srate=48000, c_ssize=2, c_chmask=2 - i.e. standard 48kHz/16bit/2ch USB playback -> alsa captureThis combination produces mps=24 and mc=1 for EP OUT. Yet the audio function driver sometimes queues request with
req->length192.This sounded familiar to me, and I realised that I've been running with a patch for this case that I never submitted upstream.
Does the patch below fix the issue you're seeing?
(quoted diff) +0
-0@@ -407,7 +407,7 @@ int u_audio_start_playback(struct g_audio *audio_dev)
Thanks for your hint. Your patch handles u_audio playback which is the other direction than I am trying to fix (usb host playback -> u_audio capture).
With regards,
Pavel.
Re: USB:UAC2: Incorrect req->length > maxpacket*mc - cause likely found
In reply to Pavel HofmanHi,
Dne 10. 01. 20 v 8:29 Pavel Hofman napsal(a):
Quoted 12 lines
Hi,
Together with dwc2 maintainer Minas Harutyunyan we have been troubleshooting various issues of dwc2 on RPi4. We hit a problem where the
g_audioin capture (EP OUT, playback from USB host) requests req-> length larger than maxpacket*mc.As a workaround we removed the check in
dwc2/gadget.c, however that is
not a proper solution. Minas with his team decided to add a patch where dwc2 will rejecting this type of wrong requests indwc2_hsotg_ep_queue()function, to not process these request at all. Thef_uac2+g_audio
gadget should restrain from sending such requests.
Quoted 9 lines
Steps to reproduce:
* Changing fs_epout_desc.bInterval in
f_uac2.cfrom 4 (1ms) to 1 (125us)
- the goal is to maximize available throughput of the audio gadget* Loading the
g_audiomodule with c_srate=48000, c_ssize=2, c_chmask=2 - i.e. standard 48kHz/16bit/2ch USB playback -> alsa captureThis combination produces mps=24 and mc=1 for EP OUT. Yet the audio function driver sometimes queues request with
req->length192.
IMO the problem is here https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/f_uac2.c#L675 :
First wMaxPacketSize is correctly calculated for HS and FS. In my setup it yields (confirmed by an added debug in set_ep_max_packet_size):
FS: ep_desc->wMaxPacketSize 192
HS: ep_desc->wMaxPacketSize 24
However, a few lines later the agdev->out_ep_maxpsize is set as maximum from these two values https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/f_uac2.c#L700 :
agdev->out_ep_maxpsize = max_t(u16, le16_to_cpu(fs_epout_desc.wMaxPacketSize), le16_to_cpu(hs_epout_desc.wMaxPacketSize));
In g_audio_setup this value is used for max_psize https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/u_audio.c#L519 :
prm->max_psize = g_audio->out_ep_maxpsize;
And max_psize is used as req->length in u_audio_start_capture https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/u_audio.c#L379 :
req_len = prm->max_psize;
...req->length = req_len;
192bytes (the value for FS) will be assigned here, instead of the correct 24 bytes for HS. The reason it has worked so far is the original bInterval=4 for HS which yields the same 192bytes as the original for FS. But lowering the fs_epout_desc.bInterval reduces the packets, yet the maximum of the two values is being used.
IMO the very same problem will be on u_audio playback side, the preceeding line:
agdev->in_ep_maxpsize = max_t(u16, le16_to_cpu(fs_epin_desc.wMaxPacketSize), le16_to_cpu(hs_epin_desc.wMaxPacketSize));
Unfortunately I do not know the reason for selection of the maximum value from FS and HS, I cannot create a patch. Very likely there is more hidden know-how which I do not know.
Thanks a lot for looking at the issue.
With regards,
Pavel.
Re: USB:UAC2: Incorrect req->length > maxpacket*mc - cause likely found
Dne 11. 01. 20 v 10:31 Pavel Hofman napsal(a):
Quoted 2 lines
Unfortunately I do not know the reason for selection of the maximum value from FS and HS, I cannot create a patch. Very likely there is more hidden know-how which I do not know.
The change was introduced by https://github.com/torvalds/linux/commit/eb9fecb9e69b0be8c267c55b0bb52a08e8fb6bee#diff-db92e74adcaaf2659ecf1a2135fd5901R572
Re: USB:UAC2: Incorrect req->length > maxpacket*mc - cause likely found
In reply to Pavel HofmanHi,
Dne 11. 01. 20 v 10:31 Pavel Hofman napsal(a):
Quoted 9 lines
Hi,
Dne 10. 01. 20 v 8:29 Pavel Hofman napsal(a):
Quoted 6 lines
Hi,
Together with dwc2 maintainer Minas Harutyunyan we have been troubleshooting various issues of dwc2 on RPi4. We hit a problem where the
g_audioin capture (EP OUT, playback from USB host) requests req-> length larger than maxpacket*mc.
My question relates to the recent patch https://marc.info/?l=linux-usb&m=157901102706577&w=2
Quoted 10 lines
IMO the problem is here https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/f_uac2.c#L675 :
However, a few lines later the
agdev->out_ep_maxpsizeis set as maximum from these two values https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/f_uac2.c#L700 :
agdev->out_ep_maxpsize=max_t(u16, le16_to_cpu(fs_epout_desc.wMaxPacketSize), le16_to_cpu(hs_epout_desc.wMaxPacketSize));Unfortunately I do not know the reason for selection of the maximum value from FS and HS, I cannot create a patch. Very likely there is more hidden know-how which I do not know.
Please can we solve this issue so that the gadget can work for any bInterval value?
Thanks a lot,
Pavel.
Re: USB:UAC2: Incorrect req->length > maxpacket*mc - cause likely found
Hi Pavel,
Quoted 28 lines · On Tue, 14 Jan 2020 19:39:05 +0100 Pavel Hofman <[email protected]>…
Dne 11. 01. 20 v 10:31 Pavel Hofman napsal(a):
Quoted 9 lines
Hi,
Dne 10. 01. 20 v 8:29 Pavel Hofman napsal(a):
Quoted 6 lines
Hi,
Together with dwc2 maintainer Minas Harutyunyan we have been troubleshooting various issues of dwc2 on RPi4. We hit a problem where the
g_audioin capture (EP OUT, playback from USB host) requests req-> length larger than maxpacket*mc.My question relates to the recent patch
https://marc.info/?l=linux-usb&m=157901102706577&w=2Quoted 15 lines
IMO the problem is here
https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/f_uac2.c#L675
:However, a few lines later the
agdev->out_ep_maxpsizeis set as maximum from these two values
https://github.com/torvalds/linux/blob/master/drivers/usb/gadget/function/f_uac2.c#L700
:
agdev->out_ep_maxpsize= max_t(u16,le16_to_cpu(fs_epout_desc.wMaxPacketSize),le16_to_cpu(hs_epout_desc.wMaxPacketSize));
Quoted 7 lines
Quoted 4 lines
Unfortunately I do not know the reason for selection of the maximum value from FS and HS, I cannot create a patch. Very likely there is more hidden know-how which I do not know.
Please can we solve this issue so that the gadget can work for any bInterval value?
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's min_period_bytes which will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.
I'm not sure what the right answer is for that; we could update min_period_bytes if the PCM is opened after the gadget attaches, but then if it is re-attached at a slower speed the PCM configuration will be wrong.
-- >8 --
From 1809582b8acfa4127fb507a97206e598fa47f55a Mon Sep 17 00:00:00 2001
From: John Keeping <[email protected]>
Date: Tue, 14 Jan 2020 19:16:10 +0000
Subject: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
Prior to commit eb9fecb9e69b ("usb: gadget: f_uac2: split out audio core") the maximum packet size was calculated only from the high-speed descriptor but now we use the largest of the full-speed and high-speed descriptors.
This is correct, but the full-speed value is likely to be higher than that for high-speed and this leads to submitting requests for OUT transfers (received by the gadget) which are larger than the endpoint's maximum packet size. These are rightly rejected by the gadget core.
config_ep_by_speed() already sets up the correct maximum packet size for the enumerated speed in the usb_ep structure, so we can simply use this instead of the overall value that has been used to allocate buffers for requests.
Note that the minimum period for ALSA is still set from the largest value, and this is unavoidable because it's possible to open the audio device before the gadget has been enumerated.
--- drivers/usb/gadget/function/u_audio.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/usb/gadget/function/u_audio.c +5 -5
@@ -361,7 +361,7 @@ int u_audio_start_capture(struct g_audio *audio_dev) ep = audio_dev->out_ep; prm = &uac->c_prm;config_ep_by_speed (gadget, &audio_dev->func, ep); - req_len = prm->max_psize; + req_len = ep->maxpacket; prm->ep_enabled = true;usb_ep_enable (ep);
@@ -379,7 +379,7 @@ int u_audio_start_capture(struct g_audio *audio_dev) req->context = &prm->ureq[i]; req->length = req_len; req->complete = u_audio_iso_complete; - req->buf = prm->rbuf + i * prm->max_psize; + req->buf = prm->rbuf + i * ep->maxpacket; }if (usb_ep_queue (ep, prm->ureq[i].req,GFP_ATOMIC ))
@@ -430,9 +430,9 @@ int u_audio_start_playback(struct g_audio *audio_dev) uac->p_pktsize = min_t(unsigned int, uac->p_framesize * (params->p_srate / uac->p_interval), - prm->max_psize); + ep->maxpacket); -if (uac->p_pktsize < prm->max_psize) +if (uac->p_pktsize < ep->maxpacket) uac->p_pktsize_residue = uac->p_framesize * (params->p_srate % uac->p_interval);else
@@ -457,7 +457,7 @@ int u_audio_start_playback(struct g_audio *audio_dev) req->context = &prm->ureq[i]; req->length = req_len; req->complete = u_audio_iso_complete; - req->buf = prm->rbuf + i * prm->max_psize; + req->buf = prm->rbuf + i * ep->maxpacket; }if (usb_ep_queue (ep, prm->ureq[i].req,GFP_ATOMIC ))
Signature · 1 line
2.24.1
Re: USB:UAC2: Incorrect req->length > maxpacket*mc - cause likely found
Hi John,
Dne 14. 01. 20 v 21:04 John Keeping napsal(a):
Quoted 3 lines
I've taken a look at this and the patch below fixes it in my simple testing.
I like your solution, simple and understandable.
Quoted 3 lines
But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.
My motivation for the smaller bInterval is higher attainable throughput. In fact to reach stable operation (avoiding random xruns) I have to use larger period on RPi4 - hence larger latency anyway.
Quoted 5 lines
I'm not sure what the right answer is for that; we could update
min_period_bytesif the PCM is opened after the gadget attaches, but then if it is re-attached at a slower speed the PCM configuration will be wrong.
I would suggest to keep the minimum period setting as is.
Thanks a lot for your help.
Pavel.
Re: USB:UAC2: Incorrect req->length > maxpacket*mc - cause likely found
In reply to John KeepingHi John,
Quoted 5 lines
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.
Please can I ask you to submit your patch? IMO your perhaps slightly suboptimal solution is much better than the current broken version.
Thanks a lot,
Pavel.
PATCH 1/1 usb: gadget: u_audio: Fix high-speed max packet size
Quoted 8 lines · On Thu, 16 Jan 2020 16:39:50 +0100 Pavel Hofman <[email protected]>…
Quoted 5 lines
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.Please can I ask you to submit your patch? IMO your perhaps slightly suboptimal solution is much better than the current broken version.
Yes, the patch is definitely an improvement. I thought it would be picked up from the earlier mail, but I think Patchwork requires the subject to match, so I'm including it again here.
Are you able to provide a Tested-by for this change?
-- >8 --
Prior to commit eb9fecb9e69b ("usb: gadget: f_uac2: split out audio core") the maximum packet size was calculated only from the high-speed descriptor but now we use the largest of the full-speed and high-speed descriptors.
This is correct, but the full-speed value is likely to be higher than that for high-speed and this leads to submitting requests for OUT transfers (received by the gadget) which are larger than the endpoint's maximum packet size. These are rightly rejected by the gadget core.
config_ep_by_speed() already sets up the correct maximum packet size for the enumerated speed in the usb_ep structure, so we can simply use this instead of the overall value that has been used to allocate buffers for requests.
Note that the minimum period for ALSA is still set from the largest value, and this is unavoidable because it's possible to open the audio device before the gadget has been enumerated.
--- drivers/usb/gadget/function/u_audio.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/usb/gadget/function/u_audio.c +5 -5
@@ -361,7 +361,7 @@ int u_audio_start_capture(struct g_audio *audio_dev) ep = audio_dev->out_ep; prm = &uac->c_prm;config_ep_by_speed (gadget, &audio_dev->func, ep); - req_len = prm->max_psize; + req_len = ep->maxpacket; prm->ep_enabled = true;usb_ep_enable (ep);
@@ -379,7 +379,7 @@ int u_audio_start_capture(struct g_audio *audio_dev) req->context = &prm->ureq[i]; req->length = req_len; req->complete = u_audio_iso_complete; - req->buf = prm->rbuf + i * prm->max_psize; + req->buf = prm->rbuf + i * ep->maxpacket; }if (usb_ep_queue (ep, prm->ureq[i].req,GFP_ATOMIC ))
@@ -430,9 +430,9 @@ int u_audio_start_playback(struct g_audio *audio_dev) uac->p_pktsize = min_t(unsigned int, uac->p_framesize * (params->p_srate / uac->p_interval), - prm->max_psize); + ep->maxpacket); -if (uac->p_pktsize < prm->max_psize) +if (uac->p_pktsize < ep->maxpacket) uac->p_pktsize_residue = uac->p_framesize * (params->p_srate % uac->p_interval);else
@@ -457,7 +457,7 @@ int u_audio_start_playback(struct g_audio *audio_dev) req->context = &prm->ureq[i]; req->length = req_len; req->complete = u_audio_iso_complete; - req->buf = prm->rbuf + i * prm->max_psize; + req->buf = prm->rbuf + i * ep->maxpacket; }if (usb_ep_queue (ep, prm->ureq[i].req,GFP_ATOMIC ))
Signature · 1 line
2.24.1
Re: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
Dne 17. 01. 20 v 11:40 John Keeping napsal(a):
Quoted 9 lines
Quoted 6 lines · On Thu, 16 Jan 2020 16:39:50 +0100 Pavel Hofman <[email protected]>…
Quoted 5 lines
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.
Quoted 8 lines
Quoted 2 lines
Please can I ask you to submit your patch? IMO your perhaps slightly suboptimal solution is much better than the current broken version.
Yes, the patch is definitely an improvement. I thought it would be picked up from the earlier mail, but I think Patchwork requires the subject to match, so I'm including it again here.
Are you able to provide a Tested-by for this change?
Testing looks OK, thanks a lot!
Re: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
Dne 19. 01. 20 v 15:53 Pavel Hofman napsal(a):
Quoted 17 lines
Dne 17. 01. 20 v 11:40 John Keeping napsal(a):
Quoted 15 lines
Quoted 7 lines · On Thu, 16 Jan 2020 16:39:50 +0100 Pavel Hofman <[email protected]>…
Quoted 4 lines
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.Please can I ask you to submit your patch? IMO your perhaps slightly suboptimal solution is much better than the current broken version.
Yes, the patch is definitely an improvement. I thought it would be picked up from the earlier mail, but I think Patchwork requires the subject to match, so I'm including it again here.
Quoted 7 lines
Quoted 1 line
Are you able to provide a Tested-by for this change?
Testing looks OK, thanks a lot!
Tested-by Pavel Hofman <[email protected]>
Please may I ask for finishing the patch submittal procedure, when it is already prepared and useful?
Thanks a lot,
Pavel.
Re: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
Dne 24. 01. 20 v 13:16 Pavel Hofman napsal(a):
Quoted 27 lines
Dne 19. 01. 20 v 15:53 Pavel Hofman napsal(a):
Quoted 24 lines
Dne 17. 01. 20 v 11:40 John Keeping napsal(a):
Quoted 8 lines
Quoted 5 lines · On Thu, 16 Jan 2020 16:39:50 +0100 Pavel Hofman <[email protected]>…
Quoted 4 lines
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.Quoted 8 lines
Quoted 2 lines
Please can I ask you to submit your patch? IMO your perhaps slightly suboptimal solution is much better than the current broken version.
Yes, the patch is definitely an improvement. I thought it would be picked up from the earlier mail, but I think Patchwork requires the subject to match, so I'm including it again here.
Are you able to provide a Tested-by for this change?
Testing looks OK, thanks a lot!
Tested-by Pavel Hofman <[email protected]>
I apologize for a basic question - please which official repository to check status of a gadget patch after being accepted? Thanks a lot for the information.
Best regards,
Pavel.
Re: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
Quoted 26 lines · On Fri, 31 Jan 2020 11:30:11 +0100 Pavel Hofman <[email protected]>…
Dne 24. 01. 20 v 13:16 Pavel Hofman napsal(a):
Quoted 25 lines
Dne 19. 01. 20 v 15:53 Pavel Hofman napsal(a):
Quoted 23 lines
Dne 17. 01. 20 v 11:40 John Keeping napsal(a):
Quoted 9 lines
Quoted 6 lines · On Thu, 16 Jan 2020 16:39:50 +0100 Pavel Hofman <[email protected]>…
Quoted 5 lines
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted
high-speed bInterval setting.Quoted 8 lines
Quoted 2 lines
Please can I ask you to submit your patch? IMO your perhaps slightly suboptimal solution is much better than the current broken version.
Yes, the patch is definitely an improvement. I thought it would be picked up from the earlier mail, but I think Patchwork requires the subject to match, so I'm including it again here.
Are you able to provide a Tested-by for this change?
Testing looks OK, thanks a lot!
Quoted 7 lines
Quoted 3 lines
Quoted 2 lines
Tested-by Pavel Hofman <[email protected]>I apologize for a basic question - please which official repository to check status of a gadget patch after being accepted? Thanks a lot for the information.
If you have a kernel tree, you can ask the MAINTAINERS file:
./scripts/get_maintainer.pl --scm -f drivers/usb/gadget/function/u_audio.c
I'd expect this to appear in Felipe's tree first at:
https://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git
but I don't see it yet. I guess it won't be picked up until after the merge window.
Regards,
John
Re: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
Hi John,
Quoted 6 lines
Quoted 3 lines
I apologize for a basic question - please which official repository to check status of a gadget patch after being accepted? Thanks a lot for the information.
If you have a kernel tree, you can ask the
MAINTAINERSfile:
Quoted 9 lines
./scripts/get_maintainer.pl --scm -f drivers/usb/gadget/function/u_audio.cI'd expect this to appear in Felipe's tree first at:
https://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git
but I don't see it yet. I guess it won't be picked up until after the merge window.
Thanks a lot for you info. How does the maintainer pick a patch from the flood of messages? Some extra headers (Tested:by, Acked-by:) are sent separately by different people, does the maintainer have to keep track of all of that manually?
Thanks for your enlightenment.
Pavel.
Re: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
Quoted 14 lines · Pavel Hofman
Hi John,
Quoted 12 lines
Quoted 3 lines
I apologize for a basic question - please which official repository to check status of a gadget patch after being accepted? Thanks a lot for the information.
If you have a kernel tree, you can ask the
MAINTAINERSfile:./scripts/get_maintainer.pl --scm -f drivers/usb/gadget/function/u_audio.cI'd expect this to appear in Felipe's tree first at:
https://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git
Quoted 8 lines
Quoted 3 lines
but I don't see it yet. I guess it won't be picked up until after the merge window.
Thanks a lot for you info. How does the maintainer pick a patch from the flood of messages? Some extra headers (Tested:by, Acked-by:) are sent separately by different people, does the maintainer have to keep track of all of that manually?
I had acked it, so I was under the expectation that Greg would pick it up. Unfortunately, it seems like it has slipped through the cracks. I have now queued it in my testing/fixes branch. It'll be sent forward after -rc1 is tagged.
Signature · 1 line
balbi
Re: [PATCH] usb: gadget: u_audio: Fix high-speed max packet size
In reply to John KeepingHi,
Quoted 34 lines · John Keeping
Quoted 8 lines · On Thu, 16 Jan 2020 16:39:50 +0100 Pavel Hofman <[email protected]>…
Quoted 5 lines
I've taken a look at this and the patch below fixes it in my simple testing. But note that this doesn't adjust the PCM's
min_period_byteswhich will be necessary if you want to minimize latency with an adjusted high-speed bInterval setting.Please can I ask you to submit your patch? IMO your perhaps slightly suboptimal solution is much better than the current broken version.
Yes, the patch is definitely an improvement. I thought it would be picked up from the earlier mail, but I think Patchwork requires the subject to match, so I'm including it again here.
Are you able to provide a Tested-by for this change?
-- >8 --
Prior to commiteb9fecb9e69b("usb: gadget:f_uac2: split out audio core") the maximum packet size was calculated only from the high-speed descriptor but now we use the largest of the full-speed and high-speed descriptors.This is correct, but the full-speed value is likely to be higher than that for high-speed and this leads to submitting requests for OUT transfers (received by the gadget) which are larger than the endpoint's maximum packet size. These are rightly rejected by the gadget core.
config_ep_by_speed()already sets up the correct maximum packet size for the enumerated speed in theusb_epstructure, so we can simply use this instead of the overall value that has been used to allocate buffers for requests.
Quoted 5 lines
Note that the minimum period for ALSA is still set from the largest value, and this is unavoidable because it's possible to open the audio device before the gadget has been enumerated.
Signed-off-by John Keeping <[email protected]>
Signature · 1 line
balbi