From: Jeff Moyer Subject: Re: [PATCH v2 5/5] dax: handle media errors in dax_do_io Date: Fri, 15 Apr 2016 13:11:17 -0400 Message-ID: References: <1459303190-20072-1-git-send-email-vishal.l.verma@intel.com> <1459303190-20072-6-git-send-email-vishal.l.verma@intel.com> <1460739288.3012.3.camel@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Cc: "Wilcox\, Matthew R" , "linux-kernel\@vger.kernel.org" , "linux-block\@vger.kernel.org" , "hch\@infradead.org" , "linux-nvdimm\@ml01.01.org" , "xfs\@oss.sgi.com" , "linux-mm\@kvack.org" , "viro\@zeniv.linux.org.uk" , "akpm\@linux-foundation.org" , "axboe\@fb.com" , "linux-fsdevel\@vger.kernel.org" , "linux-ext4\@vger.kernel.org" , "david\@fromorbit.com" , "jack\@suse.cz" To: "Verma\, Vishal L" Return-path: In-Reply-To: <1460739288.3012.3.camel@intel.com> (Vishal L. Verma's message of "Fri, 15 Apr 2016 16:54:48 +0000") Sender: owner-linux-mm@kvack.org List-Id: linux-ext4.vger.kernel.org "Verma, Vishal L" writes: > On Fri, 2016-04-15 at 12:11 -0400, Jeff Moyer wrote: >> Vishal Verma writes: >> > + if (IS_DAX(inode)) { >> > + ret =3D dax_do_io(iocb, inode, iter, offset, >> > blkdev_get_block, >> > =C2=A0 NULL, DIO_SKIP_DIO_COUNT); >> > - return __blockdev_direct_IO(iocb, inode, I_BDEV(inode), >> > iter, offset, >> > + if (ret =3D=3D -EIO && (iov_iter_rw(iter) =3D=3D WRITE)) >> > + ret_saved =3D ret; >> > + else >> > + return ret; >> > + } >> > + >> > + ret =3D __blockdev_direct_IO(iocb, inode, I_BDEV(inode), >> > iter, offset, >> > =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0blkdev_get_block, NULL, NULL, >> > =C2=A0 =C2=A0=C2=A0=C2=A0=C2=A0DIO_SKIP_DIO_COUNT); >> > + if (ret < 0 && ret_saved) >> > + return ret_saved; >> > + >> Hmm, did you just break async DIO?=C2=A0=C2=A0I think you did!=C2=A0=C2= =A0:) >> __blockdev_direct_IO can return -EIOCBQUEUED, and you've now turned >> that >> into -EIO.=C2=A0=C2=A0Really, I don't see a reason to save that first >> -EIO.=C2=A0=C2=A0The >> same applies to all instances in this patch. > > The reason I saved it was if __blockdev_direct_IO fails for some > reason, we should return the original cause o the error, which was an > EIO.. i.e. we shouldn't be hiding the EIO if the direct_IO fails with > something else.. OK. > But, how does _EIOCBQUEUED work? Maybe we need an exception for it? For async direct I/O, only the setup phase of the I/O is performed and then we return to the caller. -EIOCBQUEUED signifies this. You're heading towards code that looks like this: if (IS_DAX(inode)) { ret =3D dax_do_io(iocb, inode, iter, offset, blkdev_get_blo= ck, NULL, DIO_SKIP_DIO_COUNT); if (ret =3D=3D -EIO && (iov_iter_rw(iter) =3D=3D WRITE)) ret_saved =3D ret; else return ret; } ret =3D __blockdev_direct_IO(iocb, inode, I_BDEV(inode), iter, offs= et, blkdev_get_block, NULL, NULL, DIO_SKIP_DIO_COUNT); if (ret < 0 && ret !=3D -EIOCBQUEUED && ret_saved) return ret_saved; There's a lot of special casing here, so you might consider adding comments. Cheers, Jeff -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org