Hi Robert, On Sat, 26 Jan 2013 07:11:18 -0500 (EST), "Robert P. J. Day" <rpj...@crashcourse.ca> wrote:
> On Sat, 26 Jan 2013, Albert ARIBAUD wrote: > > > Hi Robert, > > > > On Sat, 26 Jan 2013 06:38:51 -0500 (EST), "Robert P. J. Day" > > <rpj...@crashcourse.ca> wrote: > > ... snip ... > > > > and second, i'm not sure how to read this out of cmd_pci.c: > > > > > > ===== start > > > > > > #ifdef CONFIG_SYS_LONGHELP > > > static char pci_help_text[] = > > > "[bus] [long]\n" > > > " - short or long list of PCI devices on bus 'bus'\n" > > > #ifdef CONFIG_CMD_PCI_ENUM > > > "pci enum\n" > > > " - re-enumerate PCI buses\n" > > > #endif > > > "pci header b.d.f\n" > > > " - show header of PCI device 'bus.device.function'\n" > > > "pci display[.b, .w, .l] b.d.f [address] [# of objects]\n" > > > " - display PCI configuration space (CFG)\n" > > > "pci next[.b, .w, .l] b.d.f address\n" > > > " - modify, read and keep CFG address\n" > > > "pci modify[.b, .w, .l] b.d.f address\n" > > > " - modify, auto increment CFG address\n" > > > "pci write[.b, .w, .l] b.d.f address value\n" > > > " - write to CFG address"; > > > #endif > > > > > > U_BOOT_CMD( > > > pci, 5, 1, do_pci, > > > "list and access PCI Configuration Space", pci_help_text > > > ); > > > > > > ===== end > > > > > > note how, if CONFIG_SYS_LONGHELP is defined, the symbol > > > "pci_help_text" is created as the text, but its *usage* just below in > > > the U_BOOT_CMD macro is *outside* of that preprocessor check. how > > > would that work if CONFIG_SYS_LONGHELP is undefined? not at my dev > > > host right this minute so i can't test, but it just looks ... weird. > > > > Probably would not work. Submit a fix. :) > > there appears to be a number of common/cmd_*.c files that have that > structure: > > $ grep "ifdef.*CONFIG_SYS_LONGHELP" cmd*.c > cmd_bootm.c:#ifdef CONFIG_SYS_LONGHELP > cmd_bootm.c:#ifdef CONFIG_SYS_LONGHELP > cmd_fdt.c:#ifdef CONFIG_SYS_LONGHELP > cmd_help.c:#ifdef CONFIG_SYS_LONGHELP > cmd_i2c.c:#ifdef CONFIG_SYS_LONGHELP > cmd_mp.c:#ifdef CONFIG_SYS_LONGHELP > cmd_mtdparts.c:#ifdef CONFIG_SYS_LONGHELP > cmd_nand.c:#ifdef CONFIG_SYS_LONGHELP > cmd_nvedit.c:#ifdef CONFIG_SYS_LONGHELP > cmd_pci.c:#ifdef CONFIG_SYS_LONGHELP > cmd_source.c:#ifdef CONFIG_SYS_LONGHELP > cmd_ximg.c:#ifdef CONFIG_SYS_LONGHELP > $ > > so it's not just one file. The grep above just shows that there are files have conditionally compiled code for CONFIG_SYS_LONGHELP; however it does not show per se that these files wound not compile without it. Even assuming some of these files would not build without CONFIG_SYS_LONGHELP (cmd_pci.c is already proven not to), at least for ARM (an architecture for which I frequently build all targets) no target fails with an undefined variable, which implies that no target uses any of those files without defining CONFIG_SYS_LONGHELP. My conclusion is: if you need to build one or more targets without CONFIG_SYS_LONGHELP and it fails in some cmd_*.c file, then submit a patch that fixes the cmd_*.c files and undefines CONFIG_SYS_LONGHELP in the include/configs/*.h header files for those targets. > rday Amicalement, -- Albert. _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
