netdev - Re: Re: [PATCH] net: less interrupt masking in NAPI

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date:	Sat, 10 Jan 2015 22:27:58 +0200
From:	Oded Gabbay <oded.gabbay@....com>
To:	<davem@...emloft.net>, <eric.dumazet@...il.com>
CC:	<netdev@...r.kernel.org>, <willemb@...gle.com>
Subject: Re: Re: [PATCH] net: less interrupt masking in NAPI

On 2014/11/4 1:25, David Miller wrote:
> From: Eric Dumazet <eric.dumazet@...xxxxxx>
> Date: Sun, 02 Nov 2014 06:19:33 -0800
> 
>> From: Eric Dumazet <edumazet@...xxxxxxx>
>>
>> net_rx_action() can mask irqs a single time to transfert sd->poll_list
>> into a private list, for a very short duration.
>>
>> Then, napi_complete() can avoid masking irqs again,
>> and net_rx_action() only needs to mask irq again in slow path.
>>
>> This patch removes 2 couples of irq mask/unmask per typical NAPI run,
>> more if multiple napi were triggered.
>>
>> Note this also allows to give control back to caller (do_softirq())
>> more often, so that other softirq handlers can be called a bit earlier,
>> or ksoftirqd can be wakeup earlier under pressure.
>>
>> This was developed while testing an alternative to RX interrupt
>> mitigation to reduce latencies while keeping or improving GRO
>> aggregation on fast NIC.
>>
>> Idea is to test napi->gro_list at the end of a napi->poll() and
>> reschedule one NAPI poll, but after servicing a full round of
>> softirqs (timers, TX, rcu, ...). This will be allowed only if softirq
>> is currently serviced by idle task or ksoftirqd, and resched not needed.
>>
>> Signed-off-by: Eric Dumazet <edumazet@...xxxxxxx>
> 
> Also applied, thanks Eric.

Hi,
Unfortunately, this patch breaks my "Qualcomm Atheros AR8161 Gigabit 
Ethernet (rev 10)" Ethernet controller, which is handled by the alx 
network driver.

ogabbay@...dg-ubuntu:~$ lspci -s 01:00.0 -k
01:00.0 Ethernet controller: 
	Qualcomm Atheros AR8161 Gigabit Ethernet (rev 10)
	Subsystem: Qualcomm Atheros Device 1071
	Kernel driver in use: alx

I have this controller on a mobile platform which I use to test amdkfd 
on Kaveri AMD APU and from 3.19-rc1, the network stopped working when 
trying to transfer files through scp or nfs.

I bisected the kernel (from 3.18.0 to 3.19-rc1) and reached this patch.

Here is the log of the bisect:

git bisect start
# bad: [97bf6af1f928216fd6c5a66e8a57bfa95a659672] Linux 3.19-rc1
git bisect bad 97bf6af1f928216fd6c5a66e8a57bfa95a659672

# good: [b2776bf7149bddd1f4161f14f79520f17fc1d71d] Linux 3.18
git bisect good b2776bf7149bddd1f4161f14f79520f17fc1d71d

# bad: [70e71ca0af244f48a5dcf56dc435243792e3a495] Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next
git bisect bad 70e71ca0af244f48a5dcf56dc435243792e3a495

# good: [e28870f9b3e92cd3570925089c6bb789c2603bc4] Merge tag 'backlight-for-linus-3.19' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/backlight
git bisect good e28870f9b3e92cd3570925089c6bb789c2603bc4

# bad: [450fa21942fe2c37f0c9f52d1a33bbc081eee288] sh_eth: Remove redundant alignment adjustment
git bisect bad 450fa21942fe2c37f0c9f52d1a33bbc081eee288

# bad: [5c8d19da950861d0482abc0ac3481acca34b008f] e100e: use netdev_rss_key_fill() helper
git bisect bad 5c8d19da950861d0482abc0ac3481acca34b008f

# good: [bf515fb11ab539c76d04f0e3c5216ed41f41d81f] Merge tag 'mac80211-next-for-john-2014-11-04' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next
git bisect good bf515fb11ab539c76d04f0e3c5216ed41f41d81f

# bad: [2c99cd914d4fed9160d98849c9dd38034616768e] Merge branch 'amd-xgbe-next'
git bisect bad 2c99cd914d4fed9160d98849c9dd38034616768e

# good: [3d762a0f0ab9cb4a6b5993db3ce56c92f9f90ab2] net: dsa: Add support for reading switch registers with ethtool
git bisect good 3d762a0f0ab9cb4a6b5993db3ce56c92f9f90ab2

# bad: [8ce0c8254f15229aa99fc6c04141f28c446e5f8c] Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next
git bisect bad 8ce0c8254f15229aa99fc6c04141f28c446e5f8c

# good: [f0c65567b3c1b23f79e8a49139580a3872a68d1f] Merge branch 'sunvnet-multi-tx-queue'
git bisect good f0c65567b3c1b23f79e8a49139580a3872a68d1f

# bad: [547f2735c20023d7b50a791b1b17cacb652e9237] Merge branch 'mlx4-next'
git bisect bad 547f2735c20023d7b50a791b1b17cacb652e9237

# good: [4cdb1e2e3d3495423db558d3bb7ed11d66aabce7] net: shrink struct softnet_data
git bisect good 4cdb1e2e3d3495423db558d3bb7ed11d66aabce7

# bad: [0a98455666ec87378148a1dde97f1ce5baf75a64] net/mlx4_core: Protect port type setting by mutex
git bisect bad 0a98455666ec87378148a1dde97f1ce5baf75a64

# bad: [6e8066999800d90d52af5c84ac49ebf683d14cdc] net/mlx4_core: Prevent VF from changing port configuration
git bisect bad 6e8066999800d90d52af5c84ac49ebf683d14cdc

# bad: [d75b1ade567ffab085e8adbbdacf0092d10cd09c] net: less interrupt masking in NAPI
git bisect bad d75b1ade567ffab085e8adbbdacf0092d10cd09c

# first bad commit: [d75b1ade567ffab085e8adbbdacf0092d10cd09c] 
net: less interrupt masking in NAPI

Could you please solve this issue as it renders my board quite useless.

If you need more info, please ask.

Thanks,
	Oded
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html