bcachefs: Remove BCACHEFS_SB_MAX_SIZE & check#3001
bcachefs: Remove BCACHEFS_SB_MAX_SIZE & check#3001karelzak merged 1 commit intoutil-linux:masterfrom
Conversation
|
In this case |
|
These superblocks will also run into the global limit of 8MiB in |
I'm wondering if people would take issue with increasing this to 32MiB? if (len > 8388608 /* 8 Mib */ ) {
DBG(BUFFER, ul_debug("\t too large read request (ignore)"));
return NULL;
} |
|
Or we provide a second function that is able to bypass that limit. |
This constant had an incorrect value. However, the code already does a max. size check which is correct. Note: bcachefs can theoretically have a superblock of 32MiB, but this is very unlikely to happen. Signed-off-by: Tony Asleson <[email protected]>
tasleson
force-pushed
the
bcachefs_correct_max_size_constant
branch
from
5a3099f to
63ffa1c
Compare
|
@t-8ch I updated the PR to remove the max size constant and check. I spoke with Kent about the max size and realistically it should never get that big with most being significantly < 1 MiB. The user who was failing to find bcachefs super block with blkid had 1 that was ~70KB. It should be fine to leave the current max of 8MiB. |
tasleson
changed the title
t-8ch
left a comment
t-8ch
left a comment
There was a problem hiding this comment.
Looks good and should also go to the maintenance branches.
|
Cherry-picked to stable/v2.40 too. |
3 participants
The size of the super block as stated in the kernel:So the existing constant is off by a factor of 512. This was discovered as some existing bcachefs FS have super blocks which exceed (1 << 0x010) or 65536.This constant had an incorrect value. However, the code already
does a max. size check which is correct.
Note: bcachefs can theoretically have a superblock of 32MiB, but
this is very unlikely to happen.