Search String: Display: Description: Sort:

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*\[PATCH\]\s+Avoid\s+gettimeofday\s+when\s+not\s+needed\s*$/: 20 ]

Total 20 documents matching your query.

1. [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: ipankar Sarma <dipankar@xxxxxxxxxx>
Date: Wed, 31 Mar 2004 14:01:03 +0200
This patch implements an optimization I proposed a long time ago. gettimeofday can be quite costly and it currently is on the fast path of every packet. Normally the packet time stamp is not needed (
/archives/netdev/2004-03/msg00620.html (31,467 bytes)

2. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: @xxxxxxxxxx>
Date: Wed, 31 Mar 2004 13:22:14 +0100
Andi Kleen wrote: This patch implements an optimization I proposed a long time ago. gettimeofday can be quite costly and it currently is on the fast path of every packet. Normally the packet time sta
/archives/netdev/2004-03/msg00621.html (8,791 bytes)

3. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: stopher Chan <cchan@xxxxxxxxxxxx>
Date: Wed, 31 Mar 2004 16:45:31 +0300 (EEST)
Hi Andi, I suspect this breaks proxy arp/ndisc! The arp/ndisc requests are queued up for a random time before being processed by the proxy. Check out pneigh_enqueue() in net/core/neighbour.c for deta
/archives/netdev/2004-03/msg00622.html (9,580 bytes)

4. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: i Kleen <ak@xxxxxx>
Date: Wed, 31 Mar 2004 15:53:25 +0200
The mmap will enable the time stamp, until the socket is closed. -Andi
/archives/netdev/2004-03/msg00623.html (9,128 bytes)

5. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: xxxxxxxxxxxxx
Date: Wed, 31 Mar 2004 15:54:49 +0200
Ok, thanks for clarifying this. I guessed the intention wrong then. A better way may be to just add a new skb flag for this that is managed by the arp code? Comments? -Andi
/archives/netdev/2004-03/msg00624.html (9,116 bytes)

6. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: rvala <vnuorval@xxxxxxxxxx>
Date: Wed, 31 Mar 2004 17:24:47 +0300 (EEST)
I guess that's all right. There just doesn't seem to be any general flag field in sk_buff that we might use at this moment. Should one be added, or should we just hide the flag inside the cb field us
/archives/netdev/2004-03/msg00625.html (9,140 bytes)

7. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: n <ak@xxxxxx>
Date: Wed, 31 Mar 2004 17:10:47 +0200
There are no free bits indeed. I think ->cb is the best choice then. -Andi
/archives/netdev/2004-03/msg00626.html (9,268 bytes)

8. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: <vda@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Wed, 31 Mar 2004 09:30:02 -0800
Andi Kleen wrote: On Wed, Mar 31, 2004 at 05:24:47PM +0300, Ville Nuorvala wrote: There just doesn't seem to be any general flag field in sk_buff that we might use at this moment. Should one be added
/archives/netdev/2004-03/msg00630.html (10,460 bytes)

9. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: xx>
Date: Wed, 31 Mar 2004 19:50:20 +0200
That is what ->cb already is -Andi
/archives/netdev/2004-03/msg00632.html (9,694 bytes)

10. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: xxxxxx>
Date: Wed, 31 Mar 2004 10:53:43 -0800
Andi Kleen wrote: Why not add a 32-bit field to the skb to handle various flag needs going forward? That is what ->cb already is From the comments, it appears each layer can write over these, and sin
/archives/netdev/2004-03/msg00634.html (10,309 bytes)

11. [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Andi Kleen <ak@xxxxxx>
Date: Wed, 31 Mar 2004 14:01:03 +0200
This patch implements an optimization I proposed a long time ago. gettimeofday can be quite costly and it currently is on the fast path of every packet. Normally the packet time stamp is not needed (
/archives/netdev/2004-03/msg01284.html (31,467 bytes)

12. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: P@xxxxxxxxxxxxxx
Date: Wed, 31 Mar 2004 13:22:14 +0100
Andi Kleen wrote: This patch implements an optimization I proposed a long time ago. gettimeofday can be quite costly and it currently is on the fast path of every packet. Normally the packet time sta
/archives/netdev/2004-03/msg01285.html (8,880 bytes)

13. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Ville Nuorvala <vnuorval@xxxxxxxxxx>
Date: Wed, 31 Mar 2004 16:45:31 +0300 (EEST)
Hi Andi, I suspect this breaks proxy arp/ndisc! The arp/ndisc requests are queued up for a random time before being processed by the proxy. Check out pneigh_enqueue() in net/core/neighbour.c for deta
/archives/netdev/2004-03/msg01286.html (9,646 bytes)

14. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Andi Kleen <ak@xxxxxx>
Date: Wed, 31 Mar 2004 15:53:25 +0200
The mmap will enable the time stamp, until the socket is closed. -Andi
/archives/netdev/2004-03/msg01287.html (9,209 bytes)

15. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Andi Kleen <ak@xxxxxx>
Date: Wed, 31 Mar 2004 15:54:49 +0200
Ok, thanks for clarifying this. I guessed the intention wrong then. A better way may be to just add a new skb flag for this that is managed by the arp code? Comments? -Andi
/archives/netdev/2004-03/msg01288.html (9,231 bytes)

16. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Ville Nuorvala <vnuorval@xxxxxxxxxx>
Date: Wed, 31 Mar 2004 17:24:47 +0300 (EEST)
I guess that's all right. There just doesn't seem to be any general flag field in sk_buff that we might use at this moment. Should one be added, or should we just hide the flag inside the cb field us
/archives/netdev/2004-03/msg01289.html (9,274 bytes)

17. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Andi Kleen <ak@xxxxxx>
Date: Wed, 31 Mar 2004 17:10:47 +0200
There are no free bits indeed. I think ->cb is the best choice then. -Andi
/archives/netdev/2004-03/msg01290.html (9,454 bytes)

18. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Ben Greear <greearb@xxxxxxxxxxxxxxx>
Date: Wed, 31 Mar 2004 09:30:02 -0800
There just doesn't seem to be any general flag field in sk_buff that we might use at this moment. Should one be added, or should we just hide the flag inside the cb field using the likes of inet_skb
/archives/netdev/2004-03/msg01294.html (10,572 bytes)

19. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Andi Kleen <ak@xxxxxxx>
Date: Wed, 31 Mar 2004 19:50:20 +0200
That is what ->cb already is -Andi
/archives/netdev/2004-03/msg01296.html (9,916 bytes)

20. Re: [PATCH] Avoid gettimeofday when not needed (score: 1)
Author: Ben Greear <greearb@xxxxxxxxxxxxxxx>
Date: Wed, 31 Mar 2004 10:53:43 -0800
Why not add a 32-bit field to the skb to handle various flag needs going forward? That is what ->cb already is From the comments, it appears each layer can write over these, and since it's just a bl
/archives/netdev/2004-03/msg01298.html (10,460 bytes)


This search system is powered by Namazu