[dpdk-dev] power: fix argument cannot be negative
Commit Message
Fix issue reported by Coverity.
Coverity ID 13269 & 13266:
Function strerror(errno) has built strings only for non-negative errno values.
for negative values of errno it describe error as "Unknown error -errno"
to be more descriptive i put string "channel not found" taken from header.
The negative argument will be interpreted as a very large unsigned value.
In send_msg: Negative value used as argument to a function expecting
a positive value (for example, size of buffer or allocation)
Fixes: 445c6528b55f ("power: common interface for guest and host")
Signed-off-by: Daniel Mrzyglod <danielx.t.mrzyglod@intel.com>
---
lib/librte_power/guest_channel.c | 3 ++-
lib/librte_power/rte_power_kvm_vm.c | 3 ++-
2 files changed, 4 insertions(+), 2 deletions(-)
Comments
The title do not convey the real issue.
We should be more concerned by an issue of "wrong error message"
rather than an "argument" which "cannot be negative".
2016-04-20 16:39, Daniel Mrzyglod:
> Fix issue reported by Coverity.
> Coverity ID 13269 & 13266:
It is better to put these references below and start with the
explanation of the issue.
> Function strerror(errno) has built strings only for non-negative errno values.
> for negative values of errno it describe error as "Unknown error -errno"
> to be more descriptive i put string "channel not found" taken from header.
>
> The negative argument will be interpreted as a very large unsigned value.
OK.
The next statement is probably a useless copy paste of the coverity report.
> In send_msg: Negative value used as argument to a function expecting
> a positive value (for example, size of buffer or allocation)
Coverity issue: 13266
Coverity issue: 13269
> Fixes: 445c6528b55f ("power: common interface for guest and host")
>
> Signed-off-by: Daniel Mrzyglod <danielx.t.mrzyglod@intel.com>
[...]
> RTE_LOG(ERR, GUEST_CHANNEL, "Error on channel '%s' communications "
> - "test: %s\n", fd_path, strerror(ret));
> + "test: %s\n", fd_path, ret > 0 ? strerror(ret) :
> + "channel not connected");
The indent is messy. I sugest this:
+ RTE_LOG(ERR, GUEST_CHANNEL,
+ "Error on channel '%s' communications test: %s\n",
+ fd_path, ret > 0 ? strerror(ret) :
+ "channel not connected");
> - RTE_LOG(DEBUG, POWER, "Error sending message: %s\n", strerror(ret));
> + RTE_LOG(DEBUG, POWER, "Error sending message: %s\n", ret > 0 ? strerror(ret)
> + : "channel not connected");
+ RTE_LOG(DEBUG, POWER, "Error sending message: %s\n",
+ ret > 0 ? strerror(ret) : "channel not connected");
Applied with above changes, thanks.
> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Thomas Monjalon
> Sent: Monday, May 16, 2016 1:17 PM
> To: Mrzyglod, DanielX T <danielx.t.mrzyglod@intel.com>
> Cc: dev@dpdk.org; Carew, Alan <alan.carew@intel.com>
> Subject: Re: [dpdk-dev] [PATCH] power: fix argument cannot be negative
>
> The next statement is probably a useless copy paste of the coverity
> report.
>
> > In send_msg: Negative value used as argument to a function expecting a
> > positive value (for example, size of buffer or allocation)
A question on this point. Is it just that the Coverity message is useless
in this case or in general? For other error/warning fixes we include the
message in the commit.
John
2016-05-16 14:39 GMT+02:00 Mcnamara, John <john.mcnamara@intel.com>:
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Thomas Monjalon
>> The next statement is probably a useless copy paste of the coverity
>> report.
>>
>> > In send_msg: Negative value used as argument to a function expecting a
>> > positive value (for example, size of buffer or allocation)
>
> A question on this point. Is it just that the Coverity message is useless
> in this case or in general? For other error/warning fixes we include the
> message in the commit.
Sometimes, the coverity message is accurate, sometimes it s better to reword it.
Anyway, having 2 sentences saying the same thing is disturbing.
> -----Original Message-----
> From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com]
> Sent: Monday, May 16, 2016 2:00 PM
> ...
> > A question on this point. Is it just that the Coverity message is
> > useless in this case or in general? For other error/warning fixes we
> > include the message in the commit.
>
> Sometimes, the coverity message is accurate, sometimes it s better to
> reword it.
> Anyway, having 2 sentences saying the same thing is disturbing.
Ok. Noted.
@@ -104,7 +104,8 @@ guest_channel_host_connect(const char *path, unsigned lcore_id)
ret = guest_channel_send_msg(&pkt, lcore_id);
if (ret != 0) {
RTE_LOG(ERR, GUEST_CHANNEL, "Error on channel '%s' communications "
- "test: %s\n", fd_path, strerror(ret));
+ "test: %s\n", fd_path, ret > 0 ? strerror(ret) :
+ "channel not connected");
goto error;
}
RTE_LOG(INFO, GUEST_CHANNEL, "Channel '%s' is now connected\n", fd_path);
@@ -106,7 +106,8 @@ send_msg(unsigned lcore_id, uint32_t scale_direction)
ret = guest_channel_send_msg(&pkt[lcore_id], lcore_id);
if (ret == 0)
return 1;
- RTE_LOG(DEBUG, POWER, "Error sending message: %s\n", strerror(ret));
+ RTE_LOG(DEBUG, POWER, "Error sending message: %s\n", ret > 0 ? strerror(ret)
+ : "channel not connected");
return -1;
}