Messages in this thread | | | Date | Wed, 31 Mar 2004 19:24:07 -0700 | From | Tom Rini <> | Subject | Re: [patch 1/22] Add __early_param for all arches |
| |
On Thu, Apr 01, 2004 at 11:51:41AM +1000, Rusty Russell wrote: > On Thu, 2004-04-01 at 02:13, Tom Rini wrote: > > My first concern is can parse_args & co really be run so very early on ? > > If you can run normal C code, yes. It doesn't require any > initialization.
No malloc'ing, etc? OK. Then yes, please create some sort of merged patchset (I don't know how Andrew wants to handle it, maybe just replace the core patch?)
> > Also: > > > +/* Arch code calls this early on. */ > > > +void __init parse_early_options(const char *saved_command_line) > > > +{ > > > + static char __initdata command_line[COMMAND_LINE_SIZE]; > > > + strcpy(command_line, saved_command_line); > > > > Really should be: > > /* i386 goes right to saved_command_line */ > > if (*cmdline_p != saved_command_line) > > memcpy(saved_command_line, *cmdline_p, COMMAND_LINE_SIZE); > > /* ensure NUL terminated. */ > > saved_command_line[COMMAND_LINE_SIZE - 1] = '\0'; > > Why? Other than the nul term (which I agree with), I didn't understand > that code. > > Get the archs to pass some command line in. eg. i386 can do: > > - parse_cmdline_early(cmdline_p); > + parse_early_options(*cmdline_p); > > I feel that anyone destroying command lines should do their own > saving, and we should head that way...
memcpy(src, src, size) isn't good, is it? On i386 we start out with the commandline in saved_command_line. Everyone else I believe has a local var we point cmdline_p at, which gets copied into saved_command_line.
-- Tom Rini http://gate.crashing.org/~trini/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |