linux-kernel - Re: [PATCH 0/4]nbd: fixes for nbd

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>] [day] [month] [year] [list]

Date:	Mon, 6 Jun 2016 16:37:49 +0530
From:	Pranay Srivastava <pranjas@...il.com>
To:	"mpa@...gutronix.de" <mpa@...gutronix.de>,
	"nbd-general@...ts.sourceforge.net" 
	<nbd-general@...ts.sourceforge.net>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Cc:	"Pranay Kr. Srivastava" <pranjas@...il.com>
Subject: Re: [PATCH 0/4]nbd: fixes for nbd

This got rejected by mailing-list[due to HTML content] so resending it again.

On Sat, Jun 4, 2016 at 10:25 AM, Pranay Srivastava <pranjas@...il.com> wrote:
>
>
> On Thursday, June 2, 2016, Pranay Kr. Srivastava <pranjas@...il.com> wrote:
>> This patch series fixes the following
>>
>> 1) fix might_sleep warning on socket shutdown:
>>    Fix sock_shutdown to avoid calling kernel_sock_shutdown
>>    while holding spin_lock.
>>
>> 2) cleanup nbd_set_socket
>>    Cleanup nbd_set_socket to use spin_lock instead of
>>    irq version and remove the goto statement in favour
>>    of a simple if-else statement.
>>
>> 3) fix various coding standard warnings
>>    Make shutdown get called in a process context instead, using
>>    system_wq.
>>
>> 4) make nbd device wait for its users.
>>    When a timeout or error occurs then nbd driver simply kills
>>    the block device. Many filesystem(s) example ext2/ext3 don't
>>    expect their buffer heads to disappear like that. Fix this
>>    by making nbd device wait for its users.
>>
>>    Introduced a new field to check if the device is currently
>>    in use or not. This helps to check if the kref_put should
>>    be done on device release or not.
>>
>>    This field needs to be atomic as the release function may
>>    be called from NBD_DO_IT as well as from device's release
>>    function.
>>
>> 5) use device_attr macros for sysfs attribute
>>    use DEVICE_ATTR_RO for sysfs pid attribute.
>>
>> Changelog for v2:
>> 1) fix might_sleep warning on socket shutdown
>>    use bool timedout instead of atomic
>>
>> 2) cleanup nbd_set_socket
>>    Added this new patch to this series.
>>
>> 3) fix various coding standard warnings
>>    No Change.
>>
>> 4) make nbd device wait for its users
>>    Earlier version used to do a final kref put when
>>    the kref->counter == 2. This required a check of
>>    the internal atomic counter of kref which was ugly.
>>
>>    v2 of this patch make this more readable and doesn't
>>    do manual check of the internal counter used by kref.
>>
>> 5) use device_attr macros for sysfs attribute
>>    No Change.
>>
>> Pranay Kr. Srivastava (5):
>>   fix might_sleep warning on socket shutdown.
>>   cleanup nbd_set_socket
>>   fix various coding standard warnings
>>   make nbd device wait for its users.
>>   use device_attr macros for sysfs attribute
>>
>>  drivers/block/nbd.c | 173
>> +++++++++++++++++++++++++++++++++++++---------------
>>  1 file changed, 124 insertions(+), 49 deletions(-)
>>
>> --
>> 2.6.2
>>
>>
> Markus can you please review this series.
>
> --
>         ---P.K.S
>

Can anyone kindly review this series?


-- 
        ---P.K.S