Bug 87418 - [em] [patch] The em driver sometimes stop receive data
Summary: [em] [patch] The em driver sometimes stop receive data
Status: Closed FIXED
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: 5.4-STABLE
Hardware: Any Any
: Normal Affects Only Me
Assignee: Gleb Smirnoff
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-10-14 01:40 UTC by Dmitrij Tejblum
Modified: 2006-01-11 10:32 UTC (History)
0 users

See Also:

Attachments
file.diff (3.45 KB, patch)
2005-10-14 01:40 UTC, Dmitrij Tejblum 		no flags Details | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Dmitrij Tejblum 2005-10-14 01:40:15 UTC 
The receive function em_process_receive_interrupts() unlock the adapter
while ether_input() process the packet, and then lock it back. In the 
meantime, em_init() may be called (either from em_watchdog() from softclock
interrupt or from the ifconfig program). em_init() reset the card, in
particular it set adapter->next_rx_desc_to_check to 0 and reset hardware 
RX Head and Tail descriptor pointers. The loop in
em_process_receive_interrupts() does not expect these things to change, and
a mess may result.

Fix: The following patch tries to make em_process_receive_interrupts() fully
reenterant (1.73 was incomplete) and fix the described problem.

(I removed a code for FreeBSD 4: I don't want to deal with it and BTW Intel
themselves ships different drivers for RELENG_4, RELENG_5 and RELENG_6.)
Comment 1 Gleb Smirnoff freebsd_committer freebsd_triage 2005-10-14 11:53:17 UTC 
State Changed
From-To: open->patched

Committed, thanks!
Comment 2 Gleb Smirnoff freebsd_committer freebsd_triage 2005-10-14 11:53:17 UTC 
Responsible Changed
From-To: freebsd-bugs->glebius

I am handling this.
Comment 3 Gleb Smirnoff freebsd_committer freebsd_triage 2006-01-11 10:32:08 UTC 
State Changed
From-To: patched->closed

Merged to RELENG_5.