From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S965551AbeCACuM (ORCPT <rfc822;w@1wt.eu>);
        Wed, 28 Feb 2018 21:50:12 -0500
Received: from mail.kernel.org ([198.145.29.99]:47028 "EHLO mail.kernel.org"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S965488AbeCACuL (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Wed, 28 Feb 2018 21:50:11 -0500
DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 90C8D214EE
Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=kernel.org
Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=jaegeuk@kernel.org
Date: Wed, 28 Feb 2018 18:50:09 -0800
From: Jaegeuk Kim <jaegeuk@kernel.org>
To: Chao Yu <yuchao0@huawei.com>
Cc: Yunlong Song <yunlong.song@huawei.com>, chao@kernel.org,
        yunlong.song@icloud.com, miaoxie@huawei.com, bintian.wang@huawei.com,
        shengyong1@huawei.com, heyunlei@huawei.com,
        linux-f2fs-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v4] Revert "f2fs crypto: avoid unneeded memory allocation
 in ->readdir"
Message-ID: <20180301025009.GC18665@jaegeuk-macbookpro.roam.corp.google.com>
References: <1519463698-60555-1-git-send-email-yunlong.song@huawei.com>
 <1519787857-107910-1-git-send-email-yunlong.song@huawei.com>
 <20180228054810.GB86647@jaegeuk-macbookpro.roam.corp.google.com>
 <c36a6024-209c-b660-ba61-e681a98e6f3b@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <c36a6024-209c-b660-ba61-e681a98e6f3b@huawei.com>
User-Agent: Mutt/1.8.2 (2017-04-18)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 02/28, Chao Yu wrote:
> Hi Jaegeuk,
> 
> On 2018/2/28 13:48, Jaegeuk Kim wrote:
> > Hi Yunlong,
> > 
> > As Eric pointed out, how do you think using nohighmem for directory likewise
> 
> I'd like to ask, at the beginning, why we choose to use highmem for dentry page?
> any history reason there?

There was no huge preference on it based on performance. I just wanted not to
abuse lowmem.

Thanks,

> 
> > ext4, which looks like more efficient? Actually, we don't need to do this in
> > most of recent kernels, right?
> 
> It's OK to me to keep a line with ext4.
> 
> Thanks,
> 
> > 
> > Thanks,
> > 
> > On 02/28, Yunlong Song wrote:
> >> This reverts commit e06f86e61d7a67fe6e826010f57aa39c674f4b1b.
> >>
> >> Conflicts:
> >> 	fs/f2fs/dir.c
> >>
> >> In some platforms (such as arm), high memory is used, then the
> >> decrypting filename will cause panic, the reason see commit
> >> 569cf1876a32e574ba8a7fb825cd91bafd003882 ("f2fs crypto: allocate buffer
> >> for decrypting filename"):
> >>
> >>  We got dentry pages from high_mem, and its address space directly goes into the
> >>  decryption path via f2fs_fname_disk_to_usr.
> >>  But, sg_init_one assumes the address is not from high_mem, so we can get this
> >>  panic since it doesn't call kmap_high but kunmap_high is triggered at the end.
> >>
> >>  kernel BUG at ../../../../../../kernel/mm/highmem.c:290!
> >>  Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
> >>  ...
> >>   (kunmap_high+0xb0/0xb8) from [<c0114534>] (__kunmap_atomic+0xa0/0xa4)
> >>   (__kunmap_atomic+0xa0/0xa4) from [<c035f028>] (blkcipher_walk_done+0x128/0x1ec)
> >>   (blkcipher_walk_done+0x128/0x1ec) from [<c0366c24>] (crypto_cbc_decrypt+0xc0/0x170)
> >>   (crypto_cbc_decrypt+0xc0/0x170) from [<c0367148>] (crypto_cts_decrypt+0xc0/0x114)
> >>   (crypto_cts_decrypt+0xc0/0x114) from [<c035ea98>] (async_decrypt+0x40/0x48)
> >>   (async_decrypt+0x40/0x48) from [<c032ca34>] (f2fs_fname_disk_to_usr+0x124/0x304)
> >>   (f2fs_fname_disk_to_usr+0x124/0x304) from [<c03056fc>] (f2fs_fill_dentries+0xac/0x188)
> >>   (f2fs_fill_dentries+0xac/0x188) from [<c03059c8>] (f2fs_readdir+0x1f0/0x300)
> >>   (f2fs_readdir+0x1f0/0x300) from [<c0218054>] (vfs_readdir+0x90/0xb4)
> >>   (vfs_readdir+0x90/0xb4) from [<c0218418>] (SyS_getdents64+0x64/0xcc)
> >>   (SyS_getdents64+0x64/0xcc) from [<c0105ba0>] (ret_fast_syscall+0x0/0x30)
> >>
> >> Howerver, later patch:
> >> commit e06f86e61d7a ("f2fs crypto: avoid unneeded memory allocation in ->readdir")
> >> reverts the codes, which causes panic again in arm, so fix it back to the old version.
> >>
> >> Signed-off-by: Yunlong Song <yunlong.song@huawei.com>
> >> Reviewed-by: Chao Yu <yuchao0@huawei.com>
> >> ---
> >>  fs/f2fs/dir.c | 7 +++++++
> >>  1 file changed, 7 insertions(+)
> >>
> >> diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c
> >> index f00b5ed..de2e295 100644
> >> --- a/fs/f2fs/dir.c
> >> +++ b/fs/f2fs/dir.c
> >> @@ -825,9 +825,16 @@ int f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d,
> >>  			int save_len = fstr->len;
> >>  			int err;
> >>  
> >> +			de_name.name = f2fs_kmalloc(sbi, de_name.len, GFP_NOFS);
> >> +			if (!de_name.name)
> >> +				return -ENOMEM;
> >> +
> >> +			memcpy(de_name.name, d->filename[bit_pos], de_name.len);
> >> +
> >>  			err = fscrypt_fname_disk_to_usr(d->inode,
> >>  						(u32)de->hash_code, 0,
> >>  						&de_name, fstr);
> >> +			kfree(de_name.name);
> >>  			if (err)
> >>  				return err;
> >>  
> >> -- 
> >> 1.8.5.2
> > 
> > .
> >