size-in-bytes
From [email protected] Sat Feb 16 22:09:26 2002 [email protected] wrote: > /* >+ * blk_size_in_bytes contains the size of all block-devices in bytes >+ * (blk_size has too low a resolution, since we really need the size >+ * in 512 byte sectors, and fails on devices > 2 TB) >+ * >+ * blk_size_in_bytes[MAJOR][MINOR] >+ * >+ * if (!blk_size_in_bytes[MAJOR]) then no minor size checking is done. >+ */ >+loff_t * blk_size_in_bytes[MAX_BLKDEV]; >+ >+/* Please pin it up the block device structure not to just another arbitrary global array.
You miss the point of the patch, perhaps forgot to read the introduction :-)
The point of the patch is that all applied occurrences of blk_size[][] have disappeared. All places in the code where a size is needed, that size comes from blkdev_size_in_bytes(), and the size comes in bytes, and the size is a loff_t.
You want to remove these arrays - fine, so do I.
(And of course I removed them a few times, but that is another story.) But that is an independent action. Whatever one planned to do with blk_size[][] now applies to blk_size_in_bytes[][].
So, this patch is a step ahead but does not solve all the worlds problems at once.
Andries
Quoted 27 lines · [email protected] wrote:
From [email protected] Sat Feb 16 22:09:26 2002 [email protected] wrote: > /* >+ * blk_size_in_bytes contains the size of all block-devices in bytes >+ * (blk_size has too low a resolution, since we really need the size >+ * in 512 byte sectors, and fails on devices > 2 TB) >+ * >+ * blk_size_in_bytes[MAJOR][MINOR] >+ * >+ * if (!blk_size_in_bytes[MAJOR]) then no minor size checking is done. >+ */ >+loff_t * blk_size_in_bytes[MAX_BLKDEV]; >+ >+/* Please pin it up the block device structure not to just another arbitrary global array.You miss the point of the patch, perhaps forgot to read the introduction :-)
The point of the patch is that all applied occurrences of blk_size[][] have disappeared. All places in the code where a size is needed, that size comes from
blkdev_size_in_bytes(), and the size comes in bytes, and the size is aloff_t.
Quoted 7 lines
You want to remove these arrays - fine, so do I.
(And of course I removed them a few times, but that is another story.) But that is an independent action. Whatever one planned to do with blk_size[][] now applies to blk_size_in_bytes[][].
So, this patch is a step ahead but does not solve all the worlds problems at once.
Well I see that you have killed all places where it get's used - that's fine! But why don't just kill (fix) the initializations as well. You have already:
g->part[minor(dev)].nr_sects;
and friends right there at your hands ;-).