Closed euspectre closed 9 years ago
Linux 4.2 doesn't have the four commits from Jun 23, 2015. I will contemplate pushing those commits into a stable release (4.2.1), they are quite big thought.
I see. Thanks for your quick reply. I will try to use the code from this project on kernel 4.2 as well then.
Well, I patched kernel 4.2 with the changes from these four commits from this repository that you mentioned, namely: c10471f714826f623e65e915d73efb845066cec4 a65998456202ba099f264bfbb3500e4f075cc02e f53f406150fe8d9c2ec1cf783eef3d4d5df0f6c4 d3faadb5da88f2d2874572757b087e4b95942796
The problem is gone, the drivers has worked OK so far.
Thanks for your help!
I'm using Kernel 4.2.3-3 from OSMC and I'm getting the following warnings:
[ 486.961249] ------------[ cut here ]------------
[ 486.961379] WARNING: CPU: 0 PID: 95 at drivers/net/wireless/mediatek/mt7601u/dma.c:291 mt7601u_dma_enqueue_tx+0x29c/0x2b4 [mt7601u]()
[ 486.961397] Modules linked in: evdev uinput sha256_generic sha256_arm hmac drbg ctr ccm arc4 mt7601u mac80211 sg cfg80211 uas rfkill bcm2835_thermal bcm2835_gpiomem lirc_rpi(C) lirc_dev uio_pdrv_genirq uio rc_core fuse ipv6
[ 486.961503] CPU: 0 PID: 95 Comm: systemd-journal Tainted: G WC 4.2.3-3-osmc #1
[ 486.961516] Hardware name: BCM2708
[ 486.961563] [<c001dc88>] (unwind_backtrace) from [<c001ad28>] (show_stack+0x20/0x24)
[ 486.961605] [<c001ad28>] (show_stack) from [<c0729348>] (dump_stack+0x20/0x28)
[ 486.961634] [<c0729348>] (dump_stack) from [<c002d8fc>] (warn_slowpath_null+0x78/0xfc)
[ 486.961708] [<c002d8fc>] (warn_slowpath_null) from [<bf2a86b4>] (mt7601u_dma_enqueue_tx+0x29c/0x2b4 [mt7601u])
[ 486.961822] [<bf2a86b4>] (mt7601u_dma_enqueue_tx [mt7601u]) from [<bf2af1ec>] (mt7601u_tx+0x2dc/0x408 [mt7601u])
[ 486.962349] [<bf2af1ec>] (mt7601u_tx [mt7601u]) from [<bf1f97d8>] (__ieee80211_tx+0x340/0x69c [mac80211])
[ 486.962918] [<bf1f97d8>] (__ieee80211_tx [mac80211]) from [<bf1fa92c>] (ieee80211_xmit+0x314/0x690 [mac80211])
[ 486.963474] [<bf1fa92c>] (ieee80211_xmit [mac80211]) from [<bf1fc768>] (__ieee80211_subif_start_xmit+0x470/0x1650 [mac80211])
[ 486.964032] [<bf1fc768>] (__ieee80211_subif_start_xmit [mac80211]) from [<bf1fd964>] (ieee80211_subif_start_xmit+0x1c/0x24 [mac80211])
[ 486.964346] [<bf1fd964>] (ieee80211_subif_start_xmit [mac80211]) from [<c061e518>] (dev_hard_start_xmit+0x24c/0x458)
[ 486.964392] [<c061e518>] (dev_hard_start_xmit) from [<c06455cc>] (__qdisc_run+0x1b8/0x38c)
[ 486.964420] [<c06455cc>] (__qdisc_run) from [<c0614fe4>] (net_tx_action+0x11c/0x244)
[ 486.964450] [<c0614fe4>] (net_tx_action) from [<c0031a8c>] (__do_softirq+0x170/0x49c)
[ 486.964473] [<c0031a8c>] (__do_softirq) from [<c0032258>] (irq_exit+0xf8/0x140)
[ 486.964494] [<c0032258>] (irq_exit) from [<c0083df8>] (__handle_domain_irq+0x88/0xe8)
[ 486.964522] [<c0083df8>] (__handle_domain_irq) from [<c0017680>] (handle_IRQ+0x2c/0x30)
[ 486.964548] [<c0017680>] (handle_IRQ) from [<c0009714>] (bcm2835_handle_irq+0x1c4/0x1ec)
[ 486.964576] [<c0009714>] (bcm2835_handle_irq) from [<c0730834>] (__irq_usr+0x54/0x80)
[ 486.964588] Exception stack(0xd4e47fb0 to 0xd4e47ff8)
[ 486.964602] 7fa0: 7f6d8df8 00000002 00000002 00000002
[ 486.964622] 7fc0: 7f6db150 7f6d8030 00000010 00000000 00000016 00000001 00000003 00000000
[ 486.964637] 7fe0: 000000ec be951970 7f575fd0 7f564580 20000010 ffffffff
[ 486.964654] ---[ end trace c7fd71fbe4ef81b6 ]---
My Raspberry is freezing sometimes and the connection hangs frequently.
OS: ROSA R6 x86_64 Device: TP-Link WN727Nv4, based on MT7601U. USB IDs: 148f-7601 Kernel: 4.2.0-rc8
With kernel 4.2.0-rc8 and mt7601u that comes with it, the system hangs when I test the network speed with http://speedof.me/ or open YouTube. No reaction even to the magic SysRq.
Unfortunately, I could not obtain the backtrace of the crash itself yet. However, there was the following kernel warning some time before it, not sure if it is related:
Other hardware info: http://hw.rosalinux.ru/index.php?probe=f0a0117ca9
I encountered no such problems with kernel 4.1.6 and mt7601u from this project (rev. ae46ece5). Looks like the source code of the driver itself is the same as in 4.2.0-rc8, which makes it strange.