Mail Archive Search: <+subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*bk16\s+changes\s+to\s+cbq\s*$/>

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*bk16\s+changes\s+to\s+cbq\s*$/: 18 ]

Total 18 documents matching your query.

1. Re: bk16 changes to cbq (score: 1): Author: zik <jgarzik@xxxxxxxxx>; Date: Sat, 3 Jul 2004 10:16:46 -0700; The test is racy with drivers, on the very next line the driver could take a TX completion interrupt and unplug the queue invalidating the test entirely. If the test proves wrong, that's OK because w; /archives/netdev/2004-07/msg00060.html (8,789 bytes)
2. Re: bk16 changes to cbq (score: 1): Author: rt@xxxxxxxxxxxxxxxxxxx>; Date: 03 Jul 2004 21:21:51 -0400; Ok, I looked at Stephens email; i found it here: http://marc.theaimsgroup.com/?l=linux-netdev&m=108751387802077&w=2 Basically similar to what you said. My suggestion is we back out this change - i th; /archives/netdev/2004-07/msg00064.html (9,740 bytes)
3. Re: bk16 changes to cbq (score: 1): Author: xxxx>; Date: Tue, 6 Jul 2004 00:27:27 +0400; Actually, I do not. It looks like an optimization: if the device is throttled we just do not need to add timer, qdisc will be woken up by unthrottle. I do not see any race condition here, but also I; /archives/netdev/2004-07/msg00098.html (8,548 bytes)
4. Re: bk16 changes to cbq (score: 1): Author: avid S. Miller" <davem@xxxxxxxxxx>; Date: 05 Jul 2004 21:28:09 -0400; Indeed it looks like an optimization; instead of two events (queue watchdog and probably EOL )or device wdog) trying to reschedule device, only one does. I searched the archives couldnt find anything; /archives/netdev/2004-07/msg00116.html (9,210 bytes)
5. bk16 changes to cbq (score: 1): Author: S. Miller" <davem@xxxxxxxxxx>; Date: 03 Jul 2004 09:36:50 -0400; I am noticing this in bk16; it may be one of the changes that came from Stephen recently.. -- a/net/sched/sch_cbq.c 2004-02-20 18:37:25 -08:00 +++ b/net/sched/sch_cbq.c 2004-06-18 13:51:18 -07:00 @@; /archives/netdev/2004-07/msg00120.html (8,574 bytes)
6. Re: bk16 changes to cbq (score: 1): Author: red Spraul <manfred@xxxxxxxxxxxxxxxx>; Date: 03 Jul 2004 20:03:56 -0400; I could be wrong: The only thing the tx complete can do is open up the device for more packets to tx into the device... i.e !netif_queue_stopped(sch->dev) is true in that specific case. Which would m; /archives/netdev/2004-07/msg00139.html (8,923 bytes)
7. Re: bk16 changes to cbq (score: 1): Author: asahide NAKAMURA <nakam@xxxxxxxxxxxxxx>; Date: Tue, 6 Jul 2004 22:13:17 -0700; e1000 would hang the delay scheduler sometimes Let me look for the exact email from Stephen. Here it is: http://mailman.ds9a.nl/pipermail/lartc/2004q2/012736.html Jamal read this, and that is what pr; /archives/netdev/2004-07/msg00191.html (9,503 bytes)
8. Re: bk16 changes to cbq (score: 1): Author: xxxxxxxxxx>; Date: Wed, 7 Jul 2004 19:59:58 +0400; Well, that test was really meaningless. So, Stephen's patch is right. The problem is that Stephen's argument just does not look right and forces to suspect some worse bug somewhere. If netif_queue_s; /archives/netdev/2004-07/msg00208.html (11,011 bytes)
9. Re: bk16 changes to cbq (score: 1): Author: Hemminger <shemminger@xxxxxxxx>; Date: 07 Jul 2004 14:11:32 -0400; Not totaly Alexey. Its like you said before, the code there was an optimization. Heres a scenario especialy with drivers returning 0 on everything in their txmit() (discussion going elsewhere with Je; /archives/netdev/2004-07/msg00214.html (9,700 bytes)
10. Re: bk16 changes to cbq (score: 1): Author: "David S. Miller" <davem@xxxxxxxxxx>; Date: Sat, 3 Jul 2004 10:16:46 -0700; The test is racy with drivers, on the very next line the driver could take a TX completion interrupt and unplug the queue invalidating the test entirely. If the test proves wrong, that's OK because w; /archives/netdev/2004-07/msg00876.html (8,855 bytes)
11. Re: bk16 changes to cbq (score: 1): Author: jamal <hadi@xxxxxxxxxx>; Date: 03 Jul 2004 21:21:51 -0400; Ok, I looked at Stephens email; i found it here: http://marc.theaimsgroup.com/?l=linux-netdev&m=108751387802077&w=2 Basically similar to what you said. My suggestion is we back out this change - i th; /archives/netdev/2004-07/msg00880.html (9,862 bytes)
12. Re: bk16 changes to cbq (score: 1): Author: Alexey Kuznetsov <kuznet@xxxxxxxxxxxxx>; Date: Tue, 6 Jul 2004 00:27:27 +0400; Hello! Actually, I do not. It looks like an optimization: if the device is throttled we just do not need to add timer, qdisc will be woken up by unthrottle. I do not see any race condition here, but; /archives/netdev/2004-07/msg00914.html (8,614 bytes)
13. Re: bk16 changes to cbq (score: 1): Author: jamal <hadi@xxxxxxxxxx>; Date: 05 Jul 2004 21:28:09 -0400; Indeed it looks like an optimization; instead of two events (queue watchdog and probably EOL )or device wdog) trying to reschedule device, only one does. I searched the archives couldnt find anything; /archives/netdev/2004-07/msg00932.html (9,303 bytes)
14. bk16 changes to cbq (score: 1): Author: jamal <hadi@xxxxxxxxxx>; Date: 03 Jul 2004 09:36:50 -0400; I am noticing this in bk16; it may be one of the changes that came from Stephen recently.. -- a/net/sched/sch_cbq.c 2004-02-20 18:37:25 -08:00 +++ b/net/sched/sch_cbq.c 2004-06-18 13:51:18 -07:00 @@; /archives/netdev/2004-07/msg00936.html (8,574 bytes)
15. Re: bk16 changes to cbq (score: 1): Author: jamal <hadi@xxxxxxxxxx>; Date: 03 Jul 2004 20:03:56 -0400; I could be wrong: The only thing the tx complete can do is open up the device for more packets to tx into the device... i.e !netif_queue_stopped(sch->dev) is true in that specific case. Which would m; /archives/netdev/2004-07/msg00955.html (9,030 bytes)
16. Re: bk16 changes to cbq (score: 1): Author: "David S. Miller" <davem@xxxxxxxxxx>; Date: Tue, 6 Jul 2004 22:13:17 -0700; e1000 would hang the delay scheduler sometimes Let me look for the exact email from Stephen. Here it is: http://mailman.ds9a.nl/pipermail/lartc/2004q2/012736.html Jamal read this, and that is what pr; /archives/netdev/2004-07/msg01007.html (9,611 bytes)
17. Re: bk16 changes to cbq (score: 1): Author: Alexey Kuznetsov <kuznet@xxxxxxxxxxxxx>; Date: Wed, 7 Jul 2004 19:59:58 +0400; Hello! Well, that test was really meaningless. So, Stephen's patch is right. The problem is that Stephen's argument just does not look right and forces to suspect some worse bug somewhere. If netif_q; /archives/netdev/2004-07/msg01024.html (11,163 bytes)
18. Re: bk16 changes to cbq (score: 1): Author: jamal <hadi@xxxxxxxxxx>; Date: 07 Jul 2004 14:11:32 -0400; Not totaly Alexey. Its like you said before, the code there was an optimization. Heres a scenario especialy with drivers returning 0 on everything in their txmit() (discussion going elsewhere with Je; /archives/netdev/2004-07/msg01030.html (9,873 bytes)

Current List: 1 - 18
Page: [1]