V4L/DVB: Add sleep for firmware ready
Subsystems (2)
PATCH 4/4 V4L/DVB: Add sleep for firmware ready
Antti,
After the firmware load changes, is this patch still applicable?
Regards,
Mauro
Forwarded message:
Date: Tue, 05 Aug 2014 13:48:03 +0800
From: Bimow Chen <[email protected]>
To: [email protected]
Subject: [PATCH 4/4] V4L/DVB: Add sleep for firmware ready
Quoted 1 line
From
b19fa868ce937a6ef10f1591a49b2a7ad14964a9Mon Sep 17 00:00:00 2001
From: Bimow Chen <[email protected]>
Date: Tue, 5 Aug 2014 11:20:53 +0800
Subject: [PATCH 4/4] Add sleep for firmware ready.
--- drivers/media/usb/dvb-usb-v2/af9035.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)
drivers/media/usb/dvb-usb-v2/af9035.c +2 -0
@@ -602,6 +602,8 @@ static int af9035_download_firmware(struct dvb_usb_device *d,if (ret <0 ) goto err; +msleep (30 ); +/* firmware loaded, request boot */ req.cmd =CMD_FW_BOOT ; ret =af9035_ctrl_msg (d, &req);
Signature · 1 line
1.7.0.4
I am not sure as I cannot reproduce it. Also 30ms wait here is long as hell, whilst it is not critical.
When I look that firmware downloading from the 1-2 month old Hauppauge driver sniffs, it is not there:
That line is CMD_FW_BOOT, command 0x23 it is 3rd number:
#define CMD_FW_BOOT 0x23
000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dc
Here is whole sequence:
000311: OUT: 000000 ms 001489 ms BULK[00002] >>> 15 00 29 99 03 01 00 01 57 f7 09 02 6d 6c 02 4f 9f 02 4f a2 0b 16 000312: OUT: 000001 ms 001489 ms BULK[00081] <<< 04 99 00 66 ff 000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dc 000314: OUT: 000011 ms 001490 ms BULK[00081] <<< 04 9a 00 65 ff 000315: OUT: 000000 ms 001501 ms BULK[00002] >>> 0b 00 00 9b 01 02 00 00 12 22 40 ec 000316: OUT: 000000 ms 001501 ms BULK[00081] <<< 05 9b 00 02 62 ff
So windows driver waits 10ms after boot, not before.
Due to these reasons, I would like to skip that patch until I see error or get good explanation why it is needed and so.
regards
Antti
Quoted 40 lines · Mauro Carvalho Chehab
Antti,
After the firmware load changes, is this patch still applicable?
Regards,
MauroForwarded message:
Date: Tue, 05 Aug 2014 13:48:03 +0800
From: Bimow Chen <[email protected]>
To: [email protected]
Subject: [PATCH 4/4] V4L/DVB: Add sleep for firmware readyFrom
b19fa868ce937a6ef10f1591a49b2a7ad14964a9Mon Sep 17 00:00:00 2001
From: Bimow Chen <[email protected]>
Date: Tue, 5 Aug 2014 11:20:53 +0800
Subject: [PATCH 4/4] Add sleep for firmware ready.Signed-off-by Bimow Chen <[email protected]>--- drivers/media/usb/dvb-usb-v2/af9035.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)drivers/media/usb/dvb-usb-v2/af9035.c +2
-0@@ -602,6 +602,8 @@ static int af9035_download_firmware(struct dvb_usb_device *d,if (ret <0 ) goto err; +msleep (30 ); +/* firmware loaded, request boot */ req.cmd =CMD_FW_BOOT ; ret =af9035_ctrl_msg (d, &req);
Signature · 1 line
http://palosaari.fi/
Em Tue, 23 Sep 2014 15:00:48 +0300
Antti Palosaari <[email protected]> escreveu:
Quoted 17 lines
I am not sure as I cannot reproduce it. Also 30ms wait here is long as hell, whilst it is not critical.
When I look that firmware downloading from the 1-2 month old Hauppauge driver sniffs, it is not there:
That line is
CMD_FW_BOOT, command0x23it is 3rd number:
#defineCMD_FW_BOOT0x23
000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dcHere is whole sequence:
000311: OUT: 000000 ms 001489 ms BULK[00002] >>> 15 00 29 99 03 01 00 01 57 f7 09 02 6d 6c 02 4f 9f 02 4f a2 0b 16000312: OUT: 000001 ms 001489 ms BULK[00081] <<< 04 99 00 66 ff 000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dc 000314: OUT: 000011 ms 001490 ms BULK[00081] <<< 04 9a 00 65 ff 000315: OUT: 000000 ms 001501 ms BULK[00002] >>> 0b 00 00 9b 01 02 00
Quoted 8 lines
00 12 22 40 ec
000316: OUT: 000000 ms 001501 ms BULK[00081] <<< 05 9b 00 02 62 ffSo windows driver waits 10ms after boot, not before.
Due to these reasons, I would like to skip that patch until I see error or get good explanation why it is needed and so.
Ok. I'll tag it as RFC then.
Quoted 48 lines
regards
AnttiQuoted 40 lines · Mauro Carvalho Chehab
Antti,
After the firmware load changes, is this patch still applicable?
Regards,
MauroForwarded message:
Date: Tue, 05 Aug 2014 13:48:03 +0800
From: Bimow Chen <[email protected]>
To: [email protected]
Subject: [PATCH 4/4] V4L/DVB: Add sleep for firmware readyFrom
b19fa868ce937a6ef10f1591a49b2a7ad14964a9Mon Sep 17 00:00:00 2001
From: Bimow Chen <[email protected]>
Date: Tue, 5 Aug 2014 11:20:53 +0800
Subject: [PATCH 4/4] Add sleep for firmware ready.Signed-off-by Bimow Chen <[email protected]>--- drivers/media/usb/dvb-usb-v2/af9035.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)drivers/media/usb/dvb-usb-v2/af9035.c +2
-0@@ -602,6 +602,8 @@ static int af9035_download_firmware(struct dvb_usb_device *d,if (ret <0 ) goto err; +msleep (30 ); +/* firmware loaded, request boot */ req.cmd =CMD_FW_BOOT ; ret =af9035_ctrl_msg (d, &req);
Hi all,
It's my mistake. Sleep after boot for firmware ready, not before. Please reject this patch.
Thank you.
Best regards,
Bimow
-----Original Message-----
From: Mauro Carvalho Chehab [mailto:[email protected]]
Sent: Tuesday, September 23, 2014 9:11 PM
To: Antti Palosaari
Cc: [email protected]; Bimow Chen (陳心懋)
Subject: Re: [PATCH 4/4] V4L/DVB: Add sleep for firmware ready
Em Tue, 23 Sep 2014 15:00:48 +0300
Antti Palosaari <[email protected]> escreveu:
Quoted 17 lines
I am not sure as I cannot reproduce it. Also 30ms wait here is long as hell, whilst it is not critical.
When I look that firmware downloading from the 1-2 month old Hauppauge driver sniffs, it is not there:
That line is
CMD_FW_BOOT, command0x23it is 3rd number:
#defineCMD_FW_BOOT0x23
000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dcHere is whole sequence:
000311: OUT: 000000 ms 001489 ms BULK[00002] >>> 15 00 29 99 03 01 00 01 57 f7 09 02 6d 6c 02 4f 9f 02 4f a2 0b 16000312: OUT: 000001 ms 001489 ms BULK[00081] <<< 04 99 00 66 ff 000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dc 000314: OUT: 000011 ms 001490 ms BULK[00081] <<< 04 9a 00 65 ff 000315: OUT: 000000 ms 001501 ms BULK[00002] >>> 0b 00 00 9b 01 02 00
Quoted 8 lines
00 12 22 40 ec
000316: OUT: 000000 ms 001501 ms BULK[00081] <<< 05 9b 00 02 62 ffSo windows driver waits 10ms after boot, not before.
Due to these reasons, I would like to skip that patch until I see error or get good explanation why it is needed and so.
Ok. I'll tag it as RFC then.
Quoted 50 lines
regards
AnttiQuoted 42 lines · Mauro Carvalho Chehab
Antti,
After the firmware load changes, is this patch still applicable?
Regards,
MauroForwarded message:
Date: Tue, 05 Aug 2014 13:48:03 +0800
From: Bimow Chen <[email protected]>
To: [email protected]
Subject: [PATCH 4/4] V4L/DVB: Add sleep for firmware readyFrom
b19fa868ce937a6ef10f1591a49b2a7ad14964a9Mon Sep 17 00:00:00
2001
From: Bimow Chen <[email protected]>
Date: Tue, 5 Aug 2014 11:20:53 +0800
Subject: [PATCH 4/4] Add sleep for firmware ready.Signed-off-by Bimow Chen <[email protected]>--- drivers/media/usb/dvb-usb-v2/af9035.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)drivers/media/usb/dvb-usb-v2/af9035.c +2
-0@@ -602,6 +602,8 @@ static int af9035_download_firmware(struct dvb_usb_device *d,if (ret <0 ) goto err; +msleep (30 ); +/* firmware loaded, request boot */ req.cmd =CMD_FW_BOOT ; ret =af9035_ctrl_msg (d, &req);
After looking it again, the 10ms delay seen in sniff is between boot request and reply. Looks like it took 10ms from chip to reply boot command. But there is no any other delays before nor after booth command. If some small extra delay is really needed for Linux, but not Windows, it indicates linux USB stack or driver is more optimal having less delays.
regards
Antti
Quoted 96 lines · On 09/24/2014 06:47 AM, [email protected] wrote:
Hi all,
It's my mistake. Sleep after boot for firmware ready, not before. Please reject this patch.
Thank you.Best regards,
Bimow
-----Original Message-----
From: Mauro Carvalho Chehab [mailto:[email protected]]
Sent: Tuesday, September 23, 2014 9:11 PM
To: Antti Palosaari
Cc: [email protected]; Bimow Chen (陳心懋)
Subject: Re: [PATCH 4/4] V4L/DVB: Add sleep for firmware readyEm Tue, 23 Sep 2014 15:00:48 +0300
Antti Palosaari <[email protected]> escreveu:Quoted 17 lines
I am not sure as I cannot reproduce it. Also 30ms wait here is long as hell, whilst it is not critical.
When I look that firmware downloading from the 1-2 month old Hauppauge driver sniffs, it is not there:
That line is
CMD_FW_BOOT, command0x23it is 3rd number:
#defineCMD_FW_BOOT0x23
000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dcHere is whole sequence:
000311: OUT: 000000 ms 001489 ms BULK[00002] >>> 15 00 29 99 03 01 00 01 57 f7 09 02 6d 6c 02 4f 9f 02 4f a2 0b 16000312: OUT: 000001 ms 001489 ms BULK[00081] <<< 04 99 00 66 ff 000313: OUT: 000000 ms 001490 ms BULK[00002] >>> 05 00 23 9a 65 dc 000314: OUT: 000011 ms 001490 ms BULK[00081] <<< 04 9a 00 65 ff 000315: OUT: 000000 ms 001501 ms BULK[00002] >>> 0b 00 00 9b 01 02 00Quoted 8 lines
00 12 22 40 ec
000316: OUT: 000000 ms 001501 ms BULK[00081] <<< 05 9b 00 02 62 ffSo windows driver waits 10ms after boot, not before.
Due to these reasons, I would like to skip that patch until I see error or get good explanation why it is needed and so.
Ok. I'll tag it as RFC then.
Quoted 50 lines
regards
AnttiQuoted 42 lines · Mauro Carvalho Chehab
Antti,
After the firmware load changes, is this patch still applicable?
Regards,
MauroForwarded message:
Date: Tue, 05 Aug 2014 13:48:03 +0800
From: Bimow Chen <[email protected]>
To: [email protected]
Subject: [PATCH 4/4] V4L/DVB: Add sleep for firmware readyFrom
b19fa868ce937a6ef10f1591a49b2a7ad14964a9Mon Sep 17 00:00:00
2001
From: Bimow Chen <[email protected]>
Date: Tue, 5 Aug 2014 11:20:53 +0800
Subject: [PATCH 4/4] Add sleep for firmware ready.Signed-off-by Bimow Chen <[email protected]>--- drivers/media/usb/dvb-usb-v2/af9035.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-)drivers/media/usb/dvb-usb-v2/af9035.c +2
-0@@ -602,6 +602,8 @@ static int af9035_download_firmware(struct dvb_usb_device *d,if (ret <0 ) goto err; +msleep (30 ); +/* firmware loaded, request boot */ req.cmd =CMD_FW_BOOT ; ret =af9035_ctrl_msg (d, &req);
Signature · 1 line
http://palosaari.fi/