On Sat, 2 Jul 2011, Peter Rasmussen wrote: > It does seem that NLS_ASCII isn't important (and then I suppose it should be > deprecated?), The fact that _you_ don't need it doesn't mean it should be deprecated. > but the codepages have issues. > > If you look at the pages I have attached, it seems that only codepage 437 is > accepted as valid when mounting a volume that is of vfat type, in my case FAT32 > *). > > Even if I put codepage 465 as default type and have it set, I can't mount the > volume. Of course not, since 465 isn't the codepage you want to use. Look carefully at what you wrote: > CONFIG_FAT_DEFAULT_CODEPAGE=465 > CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-15" > CONFIG_NLS_DEFAULT="utf8" > # CONFIG_NLS_CODEPAGE_437 is not set > CONFIG_NLS_CODEPAGE_850=y > CONFIG_NLS_CODEPAGE_865=y Since the codepage you're really using is 865, you should have set FAT_DEFAULT_CODEPAGE to 865, not 465. > I see several problems with this, and it may or it may not be related to USB > Mass Storage, but as a user I think it is. It isn't. You may notice, for instance, that none of these kernel config settings have anything to do with USB -- their names don't even begin with CONFIG_USB_. > 1. If I have a FAT (12, 16, 32, 64) volume and wants to mount it in a USB > environment, being hindered by an arbitrary language setting doesn't compute. This has nothing to do with USB. If you had a SATA disk with a FAT volume, you'd run into exactly the same problem. Or an SD memory card, or a Firewire drive, etc. > 2. I remember not being able to do stuff because the config.sys and autoexec.bat > files in MSDOS3.2 (and 3.3) didn't match exactly, but as this is more than a few > years ago I don't think something similar should block me now with a current > Linux kernel. What do you expect the kernel to do when you tell it that FAT volumes use codepage 437 but you don't include kernel support for that codepage? > 3. I find it to be problematic that settings in different pages when configuring > a kernel build are so related that a 'bad' setting will affect the usage, but > the error message I can get from the utility (mount) is very broad (read: > vague). The problem is that too many different things can go wrong. The kernel can't tell the mount program about every single possible error. > 4. When building kernels I now always use 'make menuconfig' and if there is an > important relation, it would be great to be able to see that mentioned. This relation isn't built in stone. You can always override it with the "codepage" mount option, as explained in the man page for mount(8). This _is_ mentioned in the help text for CONFIG_FAT_DEFAULT_CODEPAGE, which you can see in menuconfig if you hit the Help button. > I realize that the kernel-userspace channel is thin wrt. good information for > error messages, but is that then not a somewhat larger issue to bring to the > table with the powers that be? You're certainly free to bring any issue you want to the attention of whomever you like. Try posting a message on the linux-kernel mailing list. > Anyway, I hope that I can eventually mount a volume on USB, regardless of what > the language setting may be, because that is a second line priority if you ask > me. > > Best regards, > Peter > > *) When will support for exFAT (aka FAT64) be included in the kernel? Too many > non-technical issues? This is not the right channel to ask about that? No, you should ask the FAT filesystem maintainer about that. (Search for FAT in the MAINTAINERS file.) I have no idea what work is being done in that area. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
