| Message ID | 1545096879-123632-1-git-send-email-yuehaibing@huawei.com |
|---|---|
| State | New |
| Headers | show |
| Series | [-next] cifs: Use GFP_ATOMIC when a lock is held in cifs_mount() | expand |
On December 17, 2018 11:34:39 PM GMT-02:00, YueHaibing <yuehaibing@huawei.com> wrote: >A spin lock is held before kstrndup, it may sleep with holding >the spinlock, so we should use GFP_ATOMIC instead. > >Fixes: e58c31d5e387 ("cifs: Add support for failover in >cifs_reconnect()") >Signed-off-by: YueHaibing <yuehaibing@huawei.com> >--- > fs/cifs/connect.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > >diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c >index a83aa61..3d5e308 100644 >--- a/fs/cifs/connect.c >+++ b/fs/cifs/connect.c >@@ -4574,7 +4574,8 @@ int cifs_mount(struct cifs_sb_info *cifs_sb, >struct smb_vol *vol) > tcon->remap = cifs_remap(cifs_sb); > } > cifs_sb->origin_fullpath = kstrndup(tcon->dfs_path, >- strlen(tcon->dfs_path), GFP_KERNEL); >+ strlen(tcon->dfs_path), >+ GFP_ATOMIC); > if (!cifs_sb->origin_fullpath) { > spin_unlock(&cifs_tcp_ses_lock); > rc = -ENOMEM; Reviewed-by: Paulo Alcantara <palcantara@suse.de> Thanks Paulo
merged into cifs-2.6.git for-next On Mon, Dec 17, 2018 at 7:58 PM Paulo Alcantara <paulo@paulo.ac> wrote: > > > > On December 17, 2018 11:34:39 PM GMT-02:00, YueHaibing <yuehaibing@huawei.com> wrote: > >A spin lock is held before kstrndup, it may sleep with holding > >the spinlock, so we should use GFP_ATOMIC instead. > > > >Fixes: e58c31d5e387 ("cifs: Add support for failover in > >cifs_reconnect()") > >Signed-off-by: YueHaibing <yuehaibing@huawei.com> > >--- > > fs/cifs/connect.c | 3 ++- > > 1 file changed, 2 insertions(+), 1 deletion(-) > > > >diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c > >index a83aa61..3d5e308 100644 > >--- a/fs/cifs/connect.c > >+++ b/fs/cifs/connect.c > >@@ -4574,7 +4574,8 @@ int cifs_mount(struct cifs_sb_info *cifs_sb, > >struct smb_vol *vol) > > tcon->remap = cifs_remap(cifs_sb); > > } > > cifs_sb->origin_fullpath = kstrndup(tcon->dfs_path, > >- strlen(tcon->dfs_path), GFP_KERNEL); > >+ strlen(tcon->dfs_path), > >+ GFP_ATOMIC); > > if (!cifs_sb->origin_fullpath) { > > spin_unlock(&cifs_tcp_ses_lock); > > rc = -ENOMEM; > > Reviewed-by: Paulo Alcantara <palcantara@suse.de> > > Thanks > Paulo > -- > Sent from my Android device with K-9 Mail. Please excuse my brevity.
> -----Original Message----- > From: linux-cifs-owner@vger.kernel.org <linux-cifs-owner@vger.kernel.org> On > Behalf Of Steve French > Sent: Monday, December 17, 2018 9:06 PM > To: Paulo Alcantara <paulo@paulo.ac> > Cc: YueHaibing <yuehaibing@huawei.com>; Steve French > <sfrench@samba.org>; CIFS <linux-cifs@vger.kernel.org>; samba-technical > <samba-technical@lists.samba.org>; kernel-janitors <kernel- > janitors@vger.kernel.org> > Subject: Re: [PATCH -next] cifs: Use GFP_ATOMIC when a lock is held in > cifs_mount() > > merged into cifs-2.6.git for-next > > On Mon, Dec 17, 2018 at 7:58 PM Paulo Alcantara <paulo@paulo.ac> wrote: > > > > > > > > On December 17, 2018 11:34:39 PM GMT-02:00, YueHaibing > <yuehaibing@huawei.com> wrote: > > >A spin lock is held before kstrndup, it may sleep with holding > > >the spinlock, so we should use GFP_ATOMIC instead. Wouldn't it be better to allocate the memory before entering the spinlock-protected section altogether? Tom. > > >Fixes: e58c31d5e387 ("cifs: Add support for failover in > > >cifs_reconnect()") > > >Signed-off-by: YueHaibing <yuehaibing@huawei.com> > > >--- > > > fs/cifs/connect.c | 3 ++- > > > 1 file changed, 2 insertions(+), 1 deletion(-) > > > > > >diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c > > >index a83aa61..3d5e308 100644 > > >--- a/fs/cifs/connect.c > > >+++ b/fs/cifs/connect.c > > >@@ -4574,7 +4574,8 @@ int cifs_mount(struct cifs_sb_info *cifs_sb, > > >struct smb_vol *vol) > > > tcon->remap = cifs_remap(cifs_sb); > > > } > > > cifs_sb->origin_fullpath = kstrndup(tcon->dfs_path, > > >- strlen(tcon->dfs_path), GFP_KERNEL); > > >+ strlen(tcon->dfs_path), > > >+ GFP_ATOMIC); > > > if (!cifs_sb->origin_fullpath) { > > > spin_unlock(&cifs_tcp_ses_lock); > > > rc = -ENOMEM; > > > > Reviewed-by: Paulo Alcantara <palcantara@suse.de> > > > > Thanks > > Paulo > > -- > > Sent from my Android device with K-9 Mail. Please excuse my brevity. > > > > -- > Thanks, > > Steve
diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index a83aa61..3d5e308 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -4574,7 +4574,8 @@ int cifs_mount(struct cifs_sb_info *cifs_sb, struct smb_vol *vol) tcon->remap = cifs_remap(cifs_sb); } cifs_sb->origin_fullpath = kstrndup(tcon->dfs_path, - strlen(tcon->dfs_path), GFP_KERNEL); + strlen(tcon->dfs_path), + GFP_ATOMIC); if (!cifs_sb->origin_fullpath) { spin_unlock(&cifs_tcp_ses_lock); rc = -ENOMEM;
A spin lock is held before kstrndup, it may sleep with holding the spinlock, so we should use GFP_ATOMIC instead. Fixes: e58c31d5e387 ("cifs: Add support for failover in cifs_reconnect()") Signed-off-by: YueHaibing <yuehaibing@huawei.com> --- fs/cifs/connect.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)