Bug 78121
Summary: | WARNING at inet_csk_destroy_sock | ||
---|---|---|---|
Product: | Networking | Reporter: | Andrey Rahmatullin (wrar) |
Component: | Other | Assignee: | Stephen Hemminger (stephen) |
Status: | NEW --- | ||
Severity: | normal | CC: | ben, szg00000, xerofoify |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | 3.14.5 | Subsystem: | |
Regression: | No | Bisected commit-id: | |
Attachments: | Patch Test Verison |
Description
Andrey Rahmatullin
2014-06-16 15:34:12 UTC
Created attachment 140091 [details]
Patch Test Verison
This is the patch , I think may fix it. Please tell it and let me known if it fixes it. Cheers Nick (In reply to xerofoify from comment #2) > This is the patch , I think may fix it. Please tell it and let > me known if it fixes it. > Cheers Nick Doesn't look like a "fix". WARN_ON(sk->sk_wmem_queued); WARN_ON(sk->sk_forward_alloc); There are the two lines that you are having issues with based on your trace since both are checking for not Null values this with give you the above error most likely. This is due to warn on printing the above data from your log. Again please try the patch it may work. Thanks Nick Nick, you're really not helping. The assertions are valid; the bug is probably either (1) sk_forward_alloc is not always being decremented when the associated memory is freed, or (2) some associated memory is not being freed. Fair enough , now that I look at my patch you are correct i will see if I can find out where your memory errors are coming from. Thanks Nick Weird allocations seem to be correct in the file. sk_mem_reclaim(sk); Can you do a printk after this line as it seems to maybe reclaiming memory in error as the WarnOns before it are creating any errors based on your trace back. Nick |