[LUDOC-503] "Deciding on NID Ranges" incorreclty states network ID can be hex

• Type: Bug
• Resolution: Unresolved
• Priority: Minor
• Severity: 3
• Assignee: Lustre Manual Triage
• Reporter: Nathaniel Clark

Description

Current Doc:

<nidlist>       :== <nidrange> [ ' ' <nidrange> ]
<nidrange>      :== <addrrange> '@' <net>
<addrrange>     :== '*' |
                        <ipaddr_range> |
                        <numaddr_range>
<ipaddr_range>  :==
        <numaddr_range>.<numaddr_range>.<numaddr_range>.<numaddr_range>
<numaddr_range> :== <number> |
                        <expr_list>
<expr_list>     :== '[' <range_expr> [ ',' <range_expr>] ']'
<range_expr>    :== <number> |
                        <number> '-' <number> |
                        <number> '-' <number> '/' <number>
<net>           :== <netname> | <netname><number>
<netname>       :== "lo" | "tcp" | "o2ib" | "gni"
<number>        :== <nonnegative decimal> | <hexadecimal>

 

I do not believe that <number> can be hex when specifying a NID.  libcfs_str2nid() only processes the network number as an unsigned decimal.

 

The inline documentation in (both) nidstrings.c specifies struct nidrange this way:

 /**
 * Nid range list syntax.
 * \verbatim
 *
 * <nidlist>         :== <nidrange> [ ' ' <nidrange> ]
 * <nidrange>         :== <addrrange> '@' <net>
 * <addrrange>         :== '*' |
 *             <ipaddr_range> |
 *             <cfs_expr_list>
 * <ipaddr_range>    :== <cfs_expr_list>.<cfs_expr_list>.<cfs_expr_list>.
 *             <cfs_expr_list>
 * <cfs_expr_list>   :== <number> |
 *             <expr_list>
 * <expr_list>         :== '[' <range_expr> [ ',' <range_expr>] ']'
 * <range_expr>      :== <number> |
 *             <number> '-' <number> |
 *             <number> '-' <number> '/' <number>
 * <net>         :== <netname> | <netname><number>
 * <netname>         :== "lo" | "tcp" | "o2ib" | "cib" | "openib" | "iib" |
 *             "vib" | "ra" | "elan" | "mx" | "ptl"
 * \endverbatim
 */