This revision was automatically updated to reflect the committed changes.
Closed by commit rS304202: tcp/lro: Make # of LRO entries tunable (authored by
sephe).
CHANGED PRIOR TO COMMIT
https://reviews.freebsd.org/D7499?vs=19270&id=19319#toc
REPOSITORY
rS FreeBSD src repository
CHANGES SINCE
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
--- Comment #12 from Babak Farrokhi ---
(In reply to Andrey V. Elsukov from comment #10)
No crash with this patch for over an hour (it usually crashed within a few
minutes after boot).
--
You are receiving this mail because:
You are the a
Hi,
ok, can you try 5) but also running with the interrupt threads pinned to CPU 1?
It looks like the interrupt threads are running on CPU 0, and my
/guess/ (looking at the CPU usage distributions) that sometimes the
userland bits run on the same CPU or numa domain as the interrupt
bits, and it l
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211872
Mark Linimon changed:
What|Removed |Added
CC|freebsd-am...@freebsd.org |
Assignee|freebsd-b...@fr
gallatin accepted this revision.
gallatin added a comment.
I tend to think type / range checking something like this is overkill.. I'm
OK with it as-is..
REVISION DETAIL
https://reviews.freebsd.org/D7499
EMAIL PREFERENCES
https://reviews.freebsd.org/settings/panel/emailpreferences/
To
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
--- Comment #11 from Adrian Chadd ---
hiya,
Hm, I thought i captured/fixed this.
I'd prefer the solution involve writing a function that takes the CPU and
returns the pointer in question, and that function enforces the maximum netisr
CPU
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
--- Comment #10 from Andrey V. Elsukov ---
(In reply to Babak Farrokhi from comment #9)
> Created attachment 173703 [details]
> textdump after applying patch from @ae
>
> Panicked again after patch
Ok, this is yet another case :)
Can you
> On 12 Aug 2016, at 00:52, Adrian Chadd wrote:
>
> Which ones of these hit the line rate comfortably?
So Adrian, I ran tests again using FreeBSD 11-RC1.
I put iperf throughput in result files (so that we can classify them), as well
as top -P ALL and pcm-memory.x.
iperf results : columns 3&4 a
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
--- Comment #9 from Babak Farrokhi ---
Created attachment 173703
--> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=173703&action=edit
textdump after applying patch from @ae
Panicked again after patch
--
You are receiving this mai
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
--- Comment #8 from Babak Farrokhi ---
(In reply to Andrey V. Elsukov from comment #7)
Worked for me. Thank you!
--
You are receiving this mail because:
You are the assignee for the bug.
___
fre
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
--- Comment #7 from Andrey V. Elsukov ---
So, it looks like this is enough to fix the panic:
Index: sys/net/netisr.c
===
--- sys/net/netisr.c(revision 304101)
+++ sys/net/
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
Kubilay Kocak changed:
What|Removed |Added
Status|New |Open
Flags|
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
Andrey V. Elsukov changed:
What|Removed |Added
CC||a...@freebsd.org
--- Comment #
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
--- Comment #5 from Babak Farrokhi ---
(In reply to Andrey V. Elsukov from comment #4)
So you suggest setting net.isr.maxthreads=32 would fix this (aside from locking
overhead)? Isn't RSS aware of maximum CPUs used by netisr?
--
You are r
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=211836
Andrey V. Elsukov changed:
What|Removed |Added
CC||adr...@freebsd.org,
hselasky accepted this revision.
REVISION DETAIL
https://reviews.freebsd.org/D7499
EMAIL PREFERENCES
https://reviews.freebsd.org/settings/panel/emailpreferences/
To: sepherosa_gmail.com, rrs, gallatin, np, #transport, hselasky
Cc: freebsd-net-list
hselasky added inline comments.
INLINE COMMENTS
> tcp_lro.c:79
> +
> +static unsigned tcp_lro_entries = TCP_LRO_ENTRIES;
> +SYSCTL_UINT(_net_inet_tcp_lro, OID_AUTO, entries,
Maybe you should use a procedure type and range check the argument.
REVISION DETAIL
https://reviews.freebsd.org/D7
17 matches
Mail list logo
