Re: [PATCH] hfsplus: Convert kmap() to kmap_local_page() in bnode.c

[Fabio M. De Francesco]

On mercoledì 20 luglio 2022 17:43:24 CEST Fabio M. De Francesco wrote:
> kmap() is being deprecated in favor of kmap_local_page().
> 
> Two main problems with kmap(): (1) It comes with an overhead as mapping
> space is restricted and protected by a global lock for synchronization 
and
> (2) it also requires global TLB invalidation when the kmap’s pool wraps
> and it might block when the mapping space is fully utilized until a slot
> becomes available.
> 
> With kmap_local_page() the mappings are per thread, CPU local, can take
> page faults, and can be called from any context (including interrupts).
> It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore,
> the tasks can be preempted and, when they are scheduled to run again, the
> kernel virtual addresses are restored and still valid.
> 
> Since its use in bnode.c is safe everywhere, it should be preferred.
> 
> Therefore, replace kmap() with kmap_local_page() in bnode.c. Where
> possible, use the suited standard helpers (memzero_page(), memcpy_page())
> instead of open coding kmap_local_page() plus memset() or memcpy().
> 
> Suggested-by: Ira Weiny <ira.weiny@xxxxxxxxx>
> Signed-off-by: Fabio M. De Francesco <fmdefrancesco@xxxxxxxxx>
> ---
>  fs/hfsplus/bnode.c | 105 +++++++++++++++++++++------------------------
>  1 file changed, 48 insertions(+), 57 deletions(-)

Please drop this patch because I'll send it again in a series of three, one 
per file (bnode.c, bitmap.c, btree.c).

Unfortunately, while working at the last patch (btree.c), I just noticed 
that I had missed one of the call sites of kmap() in bitmap.c.

Therefore, I'd prefer to send that again in a patchset and find a way to 
test each of the three of them on a QEMU/KVM x86_32 VM, booting a kernel 
with HIGHMEM64GB enabled.

Thanks,

Fabio