3.7 kernels and kswapd

[JD]

On 11/27/2012 02:31 PM, Bruno Wolff III wrote:
> On Tue, Nov 27, 2012 at 13:11:11 -0800,
>   JD <jd1008 at gmail.com> wrote:
>>
>> However, because of the current kswapd, building a kernel is taking an
>> inordinately long time, so it will be likely 2 or 3 days on my old 
>> unicore
>> Athlon64 3700+ before I can boot the new 3.7X
>
> Can you try the kernel-3.7.0-0.rc7.git1.1.fc19 from koji or today's 
> rawhide?
> The is also a nodebug build of this in the rawhide nodebug repo.
>
Hi Bruno,
Just booted kernel-3.7.0-0.rc7.git1.1 compiled under fc16.
2 issues:

a. The old bug with the ath9k driver has resurfaced:
dmesg shows numerous
[  393.968566] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  425.968166] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  463.983223] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  503.037956] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  541.946321] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  579.962084] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  617.975928] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  655.992814] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  694.007949] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  732.026419] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  769.976403] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  807.931055] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  846.974291] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  884.984676] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  923.004154] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  961.014261] ath: phy0: Failed to stop TX DMA, queues=0x001!
[  999.031663] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1036.986683] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1150.970389] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1188.986381] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1227.005782] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1265.015690] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1303.031654] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1340.983216] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1378.931162] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1416.947164] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1454.969391] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1492.986550] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1531.046241] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1569.019706] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1646.983235] ath: phy0: Failed to stop TX DMA, queues=0x001!
[ 1686.009143] ath: phy0: Failed to stop TX DMA, queues=0x001!

This was fixed way back when....

b. Disk driver still does not handle sleeping drives correctly
[ 3377.322059] ata4: link is slow to respond, please be patient (ready=-19)
[ 3382.041110] ata4: COMRESET failed (errno=-16)
[ 3382.041131] ata4: hard resetting link
[ 3387.359058] ata4: link is slow to respond, please be patient (ready=-19)
[ 3392.051060] ata4: COMRESET failed (errno=-16)
[ 3392.051080] ata4: hard resetting link
[ 3397.363054] ata4: link is slow to respond, please be patient (ready=-19)
[ 3526.017977] Buffer I/O error on device sdc1, logical block 61014530
[ 3526.017982] lost page write due to I/O error on sdc1
[ 3526.018544] JBD2: Error -5 detected when updating journal superblock 
for sdc1-8.
[ 3529.128778] sd 3:0:0:0: [sdc] Unhandled error code
[ 3529.128792] sd 3:0:0:0: [sdc]
[ 3529.128799] Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
[ 3529.128806] sd 3:0:0:0: [sdc] CDB:
[ 3529.128811] Read(10): 28 00 00 07 90 57 00 00 08 00
[ 3529.128843] end_request: I/O error, dev sdc, sector 495703

Which is terribly bad because it will lead to data loss and FS metadata 
corruption.