Date: Fri, 02 May 2014 03:15:51 +0300 From: "Sulev-Madis Silber (ketas)" <madis555@hot.ee> To: Nathan Whitehorn <nwhitehorn@freebsd.org> Cc: FreeBSD ARM <freebsd-arm@freebsd.org> Subject: Re: crochet - why does it (try to) change files in /usr/src? Message-ID: <5362E3B7.6000105@hot.ee> In-Reply-To: <5361BBB9.9050208@freebsd.org> References: <20140501005611.3401d271adf4db31cf8e9246@getmail.no> <CADH-AwGKD1jbxON9VxF2LFwXFOU5NbgyC=Vhktsuqdh5tJvBgg@mail.gmail.com> <CAB3ij4DJZ201qDk9ZQ5fy3NRGkDMu12VCwd3FaRPW_JAzGJD_w@mail.gmail.com> <5361AC27.8020001@freebsd.org> <5361B72D.2000506@hot.ee> <5361BBB9.9050208@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2014-05-01 06:12, Nathan Whitehorn wrote: > On 04/30/14 19:53, Sulev-Madis Silber (ketas) wrote: >> On 2014-05-01 05:06, Nathan Whitehorn wrote: >>> With both mkimg and makefs in the tree now, there should be almost no >>> need for md devices for image generation anymore. >>> -Nathan >> >> That would only work for devices that don't need FAT filesystems. >> >> >> Also, if you want to build image root-less, you need to figure out how >> to get correct file permissions into your image file. Surely you can get >> those into file, just you can't do it directly on filesystem. >> > > makefs knows how to do permissions fine, either from the file itself or > a manifest, and should be acquiring FAT support from NetBSD shortly. It > also supports building opposite-endian UFS file systems, which is very > useful for embedded applications. > -Nathan Not related much, but lately I tried to create partition table where first partition starts from 512k sector #8192... After multiple tries with gpart and fdisk I found that this is totally impossible. Felt like trying to get square peg into round hole, no matter how I tried, it always realigned the partition into wrong sector. I guess this is because of those stupid C/H/S values (does anything use them these days?)? I even looked the code that does this, but didn't know how to exactly fix this. On 11-CURRENT, gpart even complained about partition not aligned to 4m... but it was impossible to have it aligned properly, no matter how I tried. I wish I could force alignment of partitions. Although... now I checked it again and I'm confused what exactly happens there. Can anyone confirm that indeed is wrong behavior. Because it looks like it is.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5362E3B7.6000105>