On 3/20/25 9:06 AM, Matthew Wilcox wrote:
On Thu, Mar 20, 2025 at 09:01:46AM -0700, Bart Van Assche wrote:
On 3/20/25 4:13 AM, Luis Chamberlain wrote:
-/*
- * We should strive for 1 << (PAGE_SHIFT + MAX_PAGECACHE_ORDER)
- * however we constrain this to what we can validate and test.
- */
-#define BLK_MAX_BLOCK_SIZE SZ_64K
+#define BLK_MAX_BLOCK_SIZE 1 << (PAGE_SHIFT + MAX_PAGECACHE_ORDER)
/* blk_validate_limits() validates bsize, so drivers don't usually need to */
static inline int blk_validate_block_size(unsigned long bsize)
All logical block sizes above 4 KiB trigger write amplification if there
are applications that write 4 KiB at a time, isn't it? Isn't that issue
even worse for logical block sizes above 64 KiB?
I think everybody knows this Bart. You've raised it before, we've
talked about it, and you're not bringing anything new to the discussion
this time. Why bring it up again?
The patch description mentions what has been changed but does not
mention why. Shouldn't the description of this patch explain why this
change has been made? Shouldn't the description of this patch explain
for which applications this change is useful?
Thanks,
Bart.
