Hi, On Wednesday, February 17, 2016 08:29:14 AM Sascha Hauer wrote: > On Tue, Feb 16, 2016 at 07:54:47PM +0100, Markus Pargmann wrote: > > Currently all fastboot flash commands with UBI images are handled by a > > final call to 'ubiformat'. This only makes sense for flash commands > > where the target file is a mtd device. If we just want to transfer a UBI > > image we would expect a simple copy to the correct location. > > > > This patch checks if the destination file is a MTD device by opening it > > and calling an ioctl MEMGETINFO. Only for MTD devices, ubiformat is > > called. > > > > Signed-off-by: Markus Pargmann > > --- > > drivers/usb/gadget/f_fastboot.c | 17 ++++++++++++++++- > > 1 file changed, 16 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/usb/gadget/f_fastboot.c b/drivers/usb/gadget/f_fastboot.c > > index bf28f7c22aaa..192c0d6eb7f4 100644 > > --- a/drivers/usb/gadget/f_fastboot.c > > +++ b/drivers/usb/gadget/f_fastboot.c > > @@ -686,7 +686,21 @@ static void cb_flash(struct usb_ep *ep, struct usb_request *req, const char *cmd > > } > > > > if (filetype == filetype_ubi) { > > - char *cmd = asprintf("ubiformat -y -f %s %s", FASTBOOT_TMPFILE, filename); > > + char *cmd; > > + int fd; > > + struct mtd_info_user meminfo; > > + > > + fd = open(filename, O_RDONLY); > > + if (!fd) > > + goto copy; > > 0 is a valid file descriptor, though one that will never be returned by > open(). You want to check for fd < 0 here. Right, thanks, will fix that. > > > + > > + ret = ioctl(fd, MEMGETINFO, &meminfo); > > + close(fd); > > + /* Not a MTD device, ubiformat is not a valid operation */ > > + if (ret) > > + goto copy; > > + > > + cmd = asprintf("ubiformat -y -f %s %s", FASTBOOT_TMPFILE, filename); > > Hm, ubiformat should get a C api. Yes, probably. Best Regards, Markus -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |