From mboxrd@z Thu Jan 1 00:00:00 1970 From: Michal Hocko Subject: Re: Recent removal of bsg read/write support Date: Mon, 3 Sep 2018 14:28:04 +0200 Message-ID: <20180903122804.GA15074@dhcp22.suse.cz> References: <6b907fed-d83f-de75-bde4-11270a0b1b0b@interlog.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <6b907fed-d83f-de75-bde4-11270a0b1b0b@interlog.com> Sender: linux-kernel-owner@vger.kernel.org To: Douglas Gilbert Cc: Richard Weinberger , drorl@infinidat.com, LKML , linux-block@vger.kernel.org, linux-scsi@vger.kernel.org, Linus Torvalds , Christoph Hellwig , Jens Axboe List-Id: linux-scsi@vger.kernel.org On Sun 02-09-18 21:16:10, Douglas Gilbert wrote: > On 2018-09-02 01:44 PM, Richard Weinberger wrote: > > CC'ing relevant people. Otherwise your mail might get lost. > > > > On Sun, Sep 2, 2018 at 1:37 PM Dror Levin wrote: > > > > > > Note: I am not subscribed to LKML so please CC replies to this email. > > > > > > Hi, > > > > > > We have an internal tool that uses the bsg read/write interface to > > > issue SCSI commands as part of a test suite for a storage device. > > > > > > After recently reading on LWN that this interface is to be removed we > > > tried porting our code to use sg instead. However, that raises new > > > issues - mainly getting ENOMEM over iSCSI for unknown reasons. > > > > > > Because of this we would like to continue using the bsg interface, > > > even if some changes are required to meet security concerns. > > > > > > Is there any chance for this removal to be reverted? I saw it was > > > already included in 4.19-rc1. > > Hi, > Both bsg and sg are relatively thin shims over the same block layer > pass-through calls. And neither driver will themselves generate ENOMEM > unless the CPU is running low of memory. > > In my experience, the main reason for unexpected ENOMEMs *** is from > blk_rq_map_user_iov() in block/blk_map.c called from both drivers. > That is a particular resource shortage rather than memory in general. > I do notice the blk_rq_map_user_iov() is/was called with GFP_KERNEL > in bsg and GFP_ATOMIC by sg. That suggests when you call write() on > a sg device and get ENOMEM, then wait a little (depends on your app) > and try again. Well, what is the reason to use GFP_ATOMIC in the first place? I am not familiar with the code so I might be easily wrong but sg_start_req which calls blk_rq_map_user_iov resp. blk_rq_map_user with GFP_ATOMIC uses mutex. It is a conditional usage so the sleeping context might depend on the caller. But I guess it would be better to double check. It looks suspicious to me. -- Michal Hocko SUSE Labs