linux-kernel - Re: [PATCH bpf-next v2] libbpf: Optimized return value in libbpf

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20221210073341.233271-1-liuxin350@huawei.com>
Date:   Sat, 10 Dec 2022 15:33:41 +0800
From:   Xin Liu <liuxin350@...wei.com>
To:     <daniel@...earbox.net>
CC:     <andrii@...nel.org>, <ast@...nel.org>, <bpf@...r.kernel.org>,
        <haoluo@...gle.com>, <john.fastabend@...il.com>,
        <jolsa@...nel.org>, <kongweibin2@...wei.com>, <kpsingh@...nel.org>,
        <linux-kernel@...r.kernel.org>, <liuxin350@...wei.com>,
        <martin.lau@...ux.dev>, <sdf@...gle.com>, <song@...nel.org>,
        <wuchangye@...wei.com>, <xiesongyang@...wei.com>,
        <yanan@...wei.com>, <yhs@...com>, <zhangmingyi5@...wei.com>
Subject: Re: [PATCH bpf-next v2] libbpf: Optimized return value in libbpf_strerror when errno is libbpf errno

On Sat, 10 Dec 2022 00:12:58 AM Daniel Borkmann <daniel@...earbox.net> wrote:
> On 12/9/22 12:05 PM, Xin Liu wrote:
> > This is a small improvement in libbpf_strerror. When libbpf_strerror
> > is used to obtain the system error description, if the length of the
> > buf is insufficient, libbpf_sterror returns ERANGE and sets errno to
> > ERANGE.
> > 
> > However, this processing is not performed when the error code
> > customized by libbpf is obtained. Make some minor improvements here,
> > return -ERANGE and set errno to ERANGE when buf is not enough for
> > custom description.
> 
> nit: $subject line got corrupted?
> 
> > Signed-off-by: Xin Liu <liuxin350@...wei.com>
> > ---
> > 
> > v2:
> > Check the return value of snprintf to determine whether the buffer is
> > too small.
> > 
> > v1:
> > https://lore.kernel.org/bpf/20221209084047.229525-1-liuxin350@huawei.com/T/#t
> > 
> >   tools/lib/bpf/libbpf_errno.c | 15 +++++++++++----
> >   1 file changed, 11 insertions(+), 4 deletions(-)
> > 
> > diff --git a/tools/lib/bpf/libbpf_errno.c b/tools/lib/bpf/libbpf_errno.c
> > index 96f67a772a1b..6240c7cb7472 100644
> > --- a/tools/lib/bpf/libbpf_errno.c
> > +++ b/tools/lib/bpf/libbpf_errno.c
> > @@ -39,14 +39,13 @@ static const char *libbpf_strerror_table[NR_ERRNO] = {
> >   
> >   int libbpf_strerror(int err, char *buf, size_t size)
> >   {
> > +	int ret;
> 
> nit: newline after declaration
> 
> >   	if (!buf || !size)
> >   		return libbpf_err(-EINVAL);
> >   
> >   	err = err > 0 ? err : -err;
> >   
> >   	if (err < __LIBBPF_ERRNO__START) {
> > -		int ret;
> > -
> >   		ret = strerror_r(err, buf, size);
> >   		buf[size - 1] = '\0';
> >   		return libbpf_err_errno(ret);
> > @@ -56,12 +55,20 @@ int libbpf_strerror(int err, char *buf, size_t size)
> >   		const char *msg;
> >   
> >   		msg = libbpf_strerror_table[ERRNO_OFFSET(err)];
> > -		snprintf(buf, size, "%s", msg);
> > +		ret = snprintf(buf, size, "%s", msg);
> >   		buf[size - 1] = '\0';
> > +		if (ret < 0)
> > +			return libbpf_err_errno(ret);
> 
> This would pass in ret == -1 and then eventually return 1 which
> is misleading, no?
> 
> We have buf and msg non-NULL and a positive size, afaik, the only
> case where you could get a negative error now is when you pass in
> a buf with size exceeding INT_MAX..
> 
> > +		if (ret >= size)
> > +			return libbpf_err(-ERANGE);
> >   		return 0;
> >   	}
> >   
> > -	snprintf(buf, size, "Unknown libbpf error %d", err);
> > +	ret = snprintf(buf, size, "Unknown libbpf error %d", err);
> >   	buf[size - 1] = '\0';
> > +	if (ret < 0)
> > +		return libbpf_err_errno(ret);
> > +	if (ret >= size)
> > +		return libbpf_err(-ERANGE);
> >   	return libbpf_err(-ENOENT);
> >   }
> > 

The logic of returning negative numbers is really unlikely here,
I'll add some comments and delete this logic, thanks to Daniel
for pointing out.