[dm-devel] DM Regression in 4.16-rc1 - read() returns data when it shouldn't

2018-02-14 Thread Milan Broz
Hi, the commit (found by bisect) commit 18a25da84354c6bb655320de6072c00eda6eb602 Author: NeilBrown Date: Wed Sep 6 09:43:28 2017 +1000 dm: ensure bio submission follows a depth-first tree walk cause serious regression while reading from DM device. The reproducer is below, basicall

Re: [dm-devel] DM Regression in 4.16-rc1 - read() returns data when it shouldn't

2018-02-14 Thread NeilBrown
On Wed, Feb 14 2018, Milan Broz wrote: > Hi, > > the commit (found by bisect) > > commit 18a25da84354c6bb655320de6072c00eda6eb602 > Author: NeilBrown > Date: Wed Sep 6 09:43:28 2017 +1000 > > dm: ensure bio submission follows a depth-first tree walk > > cause serious regression while

Re: [dm-devel] DM Regression in 4.16-rc1 - read() returns data when it shouldn't

2018-02-14 Thread Mike Snitzer
On Wed, Feb 14 2018 at 3:39pm -0500, NeilBrown wrote: > On Wed, Feb 14 2018, Milan Broz wrote: > > > Hi, > > > > the commit (found by bisect) > > > > commit 18a25da84354c6bb655320de6072c00eda6eb602 > > Author: NeilBrown > > Date: Wed Sep 6 09:43:28 2017 +1000 > > > > dm: ensure bio

Re: [dm-devel] DM Regression in 4.16-rc1 - read() returns data when it shouldn't

2018-02-14 Thread NeilBrown
On Wed, Feb 14 2018, Mike Snitzer wrote: > On Wed, Feb 14 2018 at 3:39pm -0500, > NeilBrown wrote: > >> On Wed, Feb 14 2018, Milan Broz wrote: >> >> > Hi, >> > >> > the commit (found by bisect) >> > >> > commit 18a25da84354c6bb655320de6072c00eda6eb602 >> > Author: NeilBrown >> > Date:

Re: [dm-devel] DM Regression in 4.16-rc1 - read() returns data when it shouldn't

2018-02-14 Thread Milan Broz
On 02/15/2018 01:07 AM, NeilBrown wrote: > > And looking again at the code, it doesn't seem so bad. I was worried > about reassigning ci.io->orig_bio after calling > __split_and_process_non_flush(), but the important thing is to set it > before the last dec_pending() call, and the code gets that