From: Yongqiang Yang Subject: Re: [PATCH v2] ext4: avoid eh_entries overflow before insert extent_idx Date: Fri, 24 Jun 2011 14:49:34 +0800 Message-ID: References: <1308818837-5243-1-git-send-email-sanbai@taobao.com> <4E035453.8080808@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Robin Dong , linux-ext4@vger.kernel.org, Robin Dong To: Eric Sandeen Return-path: Received: from mail-pz0-f46.google.com ([209.85.210.46]:32841 "EHLO mail-pz0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751132Ab1FXGtf convert rfc822-to-8bit (ORCPT ); Fri, 24 Jun 2011 02:49:35 -0400 Received: by pzk9 with SMTP id 9so1609216pzk.19 for ; Thu, 23 Jun 2011 23:49:34 -0700 (PDT) In-Reply-To: <4E035453.8080808@redhat.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Thu, Jun 23, 2011 at 10:57 PM, Eric Sandeen wro= te: > On 6/23/11 3:47 AM, Robin Dong wrote: >> If eh_entries is equal to (or greater than) eh_max, the operation of >> inserting new extent_idx will make number of entries overflow. >> So check eh_entries before inserting the new extent_idx. > > Do you have any testcase you can share which shows this bug? I am not sure if Robin has any test case. According to code, I think there is no bug case. Because this function is called by ext4_ext_split() and ext4_ext_split() is called only if the index block has free space. I think the right logic should be as this patch shows, that is, we should lookup the capacity before insertion. Yongqiang. > > Thanks, > -Eric > >> Signed-off-by: Robin Dong >> --- >> =A0fs/ext4/extents.c | =A0 18 ++++++++++-------- >> =A01 files changed, 10 insertions(+), 8 deletions(-) >> >> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c >> index eb63c7b..792e77e 100644 >> --- a/fs/ext4/extents.c >> +++ b/fs/ext4/extents.c >> @@ -776,6 +776,16 @@ static int ext4_ext_insert_index(handle_t *hand= le, struct inode *inode, >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0logic= al, le32_to_cpu(curp->p_idx->ei_block)); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -EIO; >> =A0 =A0 =A0 } >> + >> + =A0 =A0 if (unlikely(le16_to_cpu(curp->p_hdr->eh_entries) >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0>=3D le16_to_cp= u(curp->p_hdr->eh_max))) { >> + =A0 =A0 =A0 =A0 =A0 =A0 EXT4_ERROR_INODE(inode, >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0"eh_ent= ries %d >=3D eh_max %d!", >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0le16_to= _cpu(curp->p_hdr->eh_entries), >> + =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0le16_to= _cpu(curp->p_hdr->eh_max)); >> + =A0 =A0 =A0 =A0 =A0 =A0 return -EIO; >> + =A0 =A0 } >> + >> =A0 =A0 =A0 len =3D EXT_MAX_INDEX(curp->p_hdr) - curp->p_idx; >> =A0 =A0 =A0 if (logical > le32_to_cpu(curp->p_idx->ei_block)) { >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 /* insert after */ >> @@ -805,14 +815,6 @@ static int ext4_ext_insert_index(handle_t *hand= le, struct inode *inode, >> =A0 =A0 =A0 ext4_idx_store_pblock(ix, ptr); >> =A0 =A0 =A0 le16_add_cpu(&curp->p_hdr->eh_entries, 1); >> >> - =A0 =A0 if (unlikely(le16_to_cpu(curp->p_hdr->eh_entries) >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0> le16_to_cpu(c= urp->p_hdr->eh_max))) { >> - =A0 =A0 =A0 =A0 =A0 =A0 EXT4_ERROR_INODE(inode, >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0"eh_ent= ries %d > eh_max %d!", >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0le16_to= _cpu(curp->p_hdr->eh_entries), >> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0le16_to= _cpu(curp->p_hdr->eh_max)); >> - =A0 =A0 =A0 =A0 =A0 =A0 return -EIO; >> - =A0 =A0 } >> =A0 =A0 =A0 if (unlikely(ix > EXT_LAST_INDEX(curp->p_hdr))) { >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 EXT4_ERROR_INODE(inode, "ix > EXT_LAST_I= NDEX!"); >> =A0 =A0 =A0 =A0 =A0 =A0 =A0 return -EIO; > > -- > To unsubscribe from this list: send the line "unsubscribe linux-ext4"= in > the body of a message to majordomo@vger.kernel.org > More majordomo info at =A0http://vger.kernel.org/majordomo-info.html > --=20 Best Wishes Yongqiang Yang -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html