Hi, (I *finally* got to this topic again...)
Am Sonntag, 8. August 2021, 16:20:12 CET schrieb Patrick Steinhardt: > On Fri, Aug 06, 2021 at 12:51:10PM +0800, Michael Chang via Grub-devel wrote: > [snip] > > diff --git a/grub-core/disk/cryptodisk.c b/grub-core/disk/cryptodisk.c > > index 90f82b2d3..c2bb2b6eb 100644 > > --- a/grub-core/disk/cryptodisk.c > > +++ b/grub-core/disk/cryptodisk.c > > @@ -1040,6 +1040,7 @@ grub_cryptodisk_cheat_mount (const char *sourcedev, > > const char *cheat) > > grub_cryptodisk_t dev; > > grub_cryptodisk_dev_t cr; > > grub_disk_t source; > > + unsigned int cheat_sector_size; > > > > /* Try to open disk. */ > > source = grub_disk_open (sourcedev); > > @@ -1062,6 +1063,25 @@ grub_cryptodisk_cheat_mount (const char *sourcedev, > > const char *cheat) > > if (!dev) > > continue; > > > > + /* Use the sector size and count of the cheat device */ > > + dev->cheat_fd = grub_util_fd_open (cheat, GRUB_UTIL_FD_O_RDONLY); > > + if (!GRUB_UTIL_FD_IS_VALID (dev->cheat_fd)) > > + { > > + grub_free (dev); > > + return grub_errno; > > + } > > + dev->total_sectors = grub_util_get_fd_size (dev->cheat_fd, cheat, > > &cheat_sector_size); > > + if (dev->total_sectors == -1) > > + { > > + grub_util_fd_close (dev->cheat_fd); > > + grub_free (dev); > > + return grub_errno; > > + } > > We may want to print error messages for both error cases. Otherwise it's > hard to tell why cheat mounting might've failed. Agreed. > > + dev->log_sector_size = cheat_sector_size; > > + dev->total_sectors >>= dev->log_sector_size; > > + grub_util_fd_close (dev->cheat_fd); > > + dev->cheat_fd = GRUB_UTIL_FD_INVALID; > > Wouldn't it make more sense to just use a separate variable for the FD? > Feels kind of weird to put it into `dev->cheat_fd`, only to clone and > unset the member variable immediately after we're done again. A bit, yes, but the alternative is having a variable "int cheat_fd" which makes it more confusing IMO. I'm fine with either option. > In any case, thanks for the patch. It does look a lot more sensible > compared to what I'd been doing. Do you want to keep owning the patch > and try to upstream it, or shall I pick it up as part of this patch > series? I'd of course retain author information. Feel free to incorporate it, or I can refine it and resend as v2. (I'll most likely get to the latter only in the next year though) It looks like we have a third patch (series) for this feature meanwhile: [PATCH 0/2] Have LUKS2 cryptomounts be useable with grub-probe I CC'd the author, let's try to coordinate. Thanks, Fabian > Patrick _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel