Using pr_<loglevel>() is more concise than
printk(KERN_<LOGLEVEL>).
Replace printks having a log level with the appropriate
pr_*() macros.
Signed-off-by: Arushi Singhal <[email protected]>
---
changes in v2
*in v1 printk() were replaced with netdev_*()
net/netfilter/nf_conntrack_acct.c | 2 +-
net/netfilter/nf_conntrack_ecache.c | 2 +-
net/netfilter/nf_conntrack_timestamp.c | 2 +-
net/netfilter/nf_nat_core.c | 2 +-
net/netfilter/nfnetlink_queue.c | 4 ++--
5 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/net/netfilter/nf_conntrack_acct.c b/net/netfilter/nf_conntrack_acct.c
index 8669167..b1c3286 100644
--- a/net/netfilter/nf_conntrack_acct.c
+++ b/net/netfilter/nf_conntrack_acct.c
@@ -80,7 +80,7 @@ static int nf_conntrack_acct_init_sysctl(struct net *net)
net->ct.acct_sysctl_header = register_net_sysctl(net, "net/netfilter",
table);
if (!net->ct.acct_sysctl_header) {
- printk(KERN_ERR "nf_conntrack_acct: can't register to sysctl.\n");
+ pr_err("nf_conntrack_acct: can't register to sysctl.\n");
goto out_register;
}
return 0;
diff --git a/net/netfilter/nf_conntrack_ecache.c b/net/netfilter/nf_conntrack_ecache.c
index caac41a..21a3048 100644
--- a/net/netfilter/nf_conntrack_ecache.c
+++ b/net/netfilter/nf_conntrack_ecache.c
@@ -372,7 +372,7 @@ static int nf_conntrack_event_init_sysctl(struct net *net)
net->ct.event_sysctl_header =
register_net_sysctl(net, "net/netfilter", table);
if (!net->ct.event_sysctl_header) {
- printk(KERN_ERR "nf_ct_event: can't register to sysctl.\n");
+ pr_err("nf_ct_event: can't register to sysctl.\n");
goto out_register;
}
return 0;
diff --git a/net/netfilter/nf_conntrack_timestamp.c b/net/netfilter/nf_conntrack_timestamp.c
index 4c4734b..f32cc86 100644
--- a/net/netfilter/nf_conntrack_timestamp.c
+++ b/net/netfilter/nf_conntrack_timestamp.c
@@ -58,7 +58,7 @@ static int nf_conntrack_tstamp_init_sysctl(struct net *net)
net->ct.tstamp_sysctl_header = register_net_sysctl(net, "net/netfilter",
table);
if (!net->ct.tstamp_sysctl_header) {
- printk(KERN_ERR "nf_ct_tstamp: can't register to sysctl.\n");
+ pr_err("nf_ct_tstamp: can't register to sysctl.\n");
goto out_register;
}
return 0;
diff --git a/net/netfilter/nf_nat_core.c b/net/netfilter/nf_nat_core.c
index 6c38421..dcda5ac 100644
--- a/net/netfilter/nf_nat_core.c
+++ b/net/netfilter/nf_nat_core.c
@@ -814,7 +814,7 @@ static int __init nf_nat_init(void)
ret = nf_ct_extend_register(&nat_extend);
if (ret < 0) {
nf_ct_free_hashtable(nf_nat_bysource, nf_nat_htable_size);
- printk(KERN_ERR "nf_nat_core: Unable to register extension\n");
+ pr_err("nf_nat_core: Unable to register extension\n");
return ret;
}
diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index 8bba231..f5ddab1 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -834,8 +834,8 @@ nfqnl_mangle(void *data, int data_len, struct nf_queue_entry *e, int diff)
nskb = skb_copy_expand(e->skb, skb_headroom(e->skb),
diff, GFP_ATOMIC);
if (!nskb) {
- printk(KERN_WARNING "nf_queue: OOM "
- "in mangle, dropping packet\n");
+ pr_warn("nf_queue: OOM "
+ "in mangle, dropping packet\n");
return -ENOMEM;
}
kfree_skb(e->skb);
--
2.7.4
On Mon, 2018-03-12 at 01:11 +0530, Arushi Singhal wrote:
> Using pr_<loglevel>() is more concise than
> printk(KERN_<LOGLEVEL>).
> Replace printks having a log level with the appropriate
> pr_*() macros.
>
> Signed-off-by: Arushi Singhal <[email protected]>
> ---
> changes in v2
> *in v1 printk() were replaced with netdev_*()
> net/netfilter/nf_conntrack_acct.c | 2 +-
> net/netfilter/nf_conntrack_ecache.c | 2 +-
> net/netfilter/nf_conntrack_timestamp.c | 2 +-
> net/netfilter/nf_nat_core.c | 2 +-
> net/netfilter/nfnetlink_queue.c | 4 ++--
> 5 files changed, 6 insertions(+), 6 deletions(-)
None of these files have a #define for pr_fmt so this
should be OK.
Perhaps coalesce the formats and remove the unnecessary periods too.
> diff --git a/net/netfilter/nf_conntrack_acct.c b/net/netfilter/nf_conntrack_acct.c
[]
> @@ -80,7 +80,7 @@ static int nf_conntrack_acct_init_sysctl(struct net *net)
> net->ct.acct_sysctl_header = register_net_sysctl(net, "net/netfilter",
> table);
> if (!net->ct.acct_sysctl_header) {
> - printk(KERN_ERR "nf_conntrack_acct: can't register to sysctl.\n");
> + pr_err("nf_conntrack_acct: can't register to sysctl.\n");
> goto out_register;
> }
> return 0;
etc...
> diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
[]
> @@ -834,8 +834,8 @@ nfqnl_mangle(void *data, int data_len, struct nf_queue_entry *e, int diff)
> nskb = skb_copy_expand(e->skb, skb_headroom(e->skb),
> diff, GFP_ATOMIC);
> if (!nskb) {
> - printk(KERN_WARNING "nf_queue: OOM "
> - "in mangle, dropping packet\n");
> + pr_warn("nf_queue: OOM "
> + "in mangle, dropping packet\n");
> return -ENOMEM;
> }
> kfree_skb(e->skb);
Hi Joe,
On Sun, Mar 11, 2018 at 12:52:41PM -0700, Joe Perches wrote:
> On Mon, 2018-03-12 at 01:11 +0530, Arushi Singhal wrote:
> > Using pr_<loglevel>() is more concise than
> > printk(KERN_<LOGLEVEL>).
> > Replace printks having a log level with the appropriate
> > pr_*() macros.
> >
> > Signed-off-by: Arushi Singhal <[email protected]>
> > ---
> > changes in v2
> > *in v1 printk() were replaced with netdev_*()
>
> > net/netfilter/nf_conntrack_acct.c | 2 +-
> > net/netfilter/nf_conntrack_ecache.c | 2 +-
> > net/netfilter/nf_conntrack_timestamp.c | 2 +-
> > net/netfilter/nf_nat_core.c | 2 +-
> > net/netfilter/nfnetlink_queue.c | 4 ++--
> > 5 files changed, 6 insertions(+), 6 deletions(-)
>
> None of these files have a #define for pr_fmt so this
> should be OK.
I think Arushi could add pr_fmt in the same go, so we skip another
follow up patch for this. @Arushi: I suggested this in my previous
email, please have a look.
This would also probably allows us to save the line break in the error
message, which IIRC is not a good practise either, eg.
pr_warn("nf_queue: OOM "
"in mangle, dropping packet\n");
> Perhaps coalesce the formats and remove the unnecessary periods too.
Yes, removing periods seems fine.
Thanks.
On Mon, Mar 12, 2018 at 03:56:15AM +0530, Arushi Singhal wrote:
> On Mon, Mar 12, 2018 at 2:17 AM, Pablo Neira Ayuso <[email protected]>
> wrote:
>
> > Hi Joe,
> >
> > On Sun, Mar 11, 2018 at 12:52:41PM -0700, Joe Perches wrote:
> > > On Mon, 2018-03-12 at 01:11 +0530, Arushi Singhal wrote:
> > > > Using pr_<loglevel>() is more concise than
> > > > printk(KERN_<LOGLEVEL>).
> > > > Replace printks having a log level with the appropriate
> > > > pr_*() macros.
> > > >
> > > > Signed-off-by: Arushi Singhal <[email protected]>
> > > > ---
> > > > changes in v2
> > > > *in v1 printk() were replaced with netdev_*()
> > >
> > > > net/netfilter/nf_conntrack_acct.c | 2 +-
> > > > net/netfilter/nf_conntrack_ecache.c | 2 +-
> > > > net/netfilter/nf_conntrack_timestamp.c | 2 +-
> > > > net/netfilter/nf_nat_core.c | 2 +-
> > > > net/netfilter/nfnetlink_queue.c | 4 ++--
> > > > 5 files changed, 6 insertions(+), 6 deletions(-)
> > >
> > > None of these files have a #define for pr_fmt so this
> > > should be OK.
> >
> > I think Arushi could add pr_fmt in the same go, so we skip another
> > follow up patch for this. @Arushi: I suggested this in my previous
> > email, please have a look.
> >
> > Hello Pablo
>
> Should I send two patches, one with the conversion of printk() to pr_() and
> another for defining pr_fmt().
>
> Or
>
> only one patch with all the changes?
I think adding pr_fmt and use pr_() belongs to the same logical
change, so one patch for this is fine.
Thanks Arushi.
P.S: Please, just send your patch [email protected]
next time, no need to Cc every list. Thanks!
Arushi Singhal <[email protected]> wrote:
> On Mon, Mar 12, 2018 at 2:17 AM, Pablo Neira Ayuso <[email protected]>
> wrote:
>
> > Hi Joe,
> >
> > On Sun, Mar 11, 2018 at 12:52:41PM -0700, Joe Perches wrote:
> > > On Mon, 2018-03-12 at 01:11 +0530, Arushi Singhal wrote:
> > > > Using pr_<loglevel>() is more concise than
> > > > printk(KERN_<LOGLEVEL>).
> > > > Replace printks having a log level with the appropriate
> > > > pr_*() macros.
> > > >
> > > > Signed-off-by: Arushi Singhal <[email protected]>
> > > > ---
> > > > changes in v2
> > > > *in v1 printk() were replaced with netdev_*()
> > >
> > > > net/netfilter/nf_conntrack_acct.c | 2 +-
> > > > net/netfilter/nf_conntrack_ecache.c | 2 +-
> > > > net/netfilter/nf_conntrack_timestamp.c | 2 +-
> > > > net/netfilter/nf_nat_core.c | 2 +-
> > > > net/netfilter/nfnetlink_queue.c | 4 ++--
> > > > 5 files changed, 6 insertions(+), 6 deletions(-)
> > >
> > > None of these files have a #define for pr_fmt so this
> > > should be OK.
> >
> > I think Arushi could add pr_fmt in the same go, so we skip another
> > follow up patch for this. @Arushi: I suggested this in my previous
> > email, please have a look.
> >
> > Hello Pablo
>
> Should I send two patches, one with the conversion of printk() to pr_() and
> another for defining pr_fmt().
>
> Or
>
> only one patch with all the changes?
Both in one, it reduces code churn.
With pr_* + pr_fmt, the module name will be prefixed automatically, see
e.g. commit e016c5e43db51875c2b541b59bd217494d213174 as an example.
> > This would also probably allows us to save the line break in the error
> > message, which IIRC is not a good practise either, eg.
> >
> > pr_warn("nf_queue: OOM "
> > "in mangle, dropping packet\n");
> >
> > > Perhaps coalesce the formats and remove the unnecessary periods too.
The above message should be removed, its useless (on oom allocator
already warns).