From: Ira Weiny <ira.weiny@intel.com>
To: Davidlohr Bueso <dave@stgolabs.net>, <vishal.l.verma@intel.com>,
<dan.j.williams@intel.com>, <dave.jiang@intel.com>,
<ira.weiny@intel.com>
Cc: <nvdimm@lists.linux.dev>, <linux-kernel@vger.kernel.org>,
Davidlohr Bueso <dave@stgolabs.net>
Subject: Re: [PATCH] drivers/nvdimm: Use local kmaps
Date: Mon, 12 Jan 2026 17:10:35 -0600 [thread overview]
Message-ID: <69657f6b74508_f1ef61007a@iweiny-mobl.notmuch> (raw)
In-Reply-To: <20251128212303.2170933-1-dave@stgolabs.net>
Davidlohr Bueso wrote:
> Replace the now deprecated kmap_atomic() with kmap_local_page().
>
> Optimizing nvdimm/pmem for highmem makes no sense as this is always
> 64bit, and the mapped regions for both btt and pmem do not require
> disabling preemption and pagefaults. Specifically, kmap does not care
> about the caller's atomic context (such as reads holding the btt arena
> spinlock) or NVDIMM_IO_ATOMIC semantics to avoid error handling when
> accessing the btt arena in general. Same for the memcpy cases. kmap
> local temporary mappings will hold valid across any context switches.
>
> Signed-off-by: Davidlohr Bueso <dave@stgolabs.net>
Queue'd for 7.0
Thanks!
Ira Weiny
> ---
> drivers/nvdimm/btt.c | 12 ++++++------
> drivers/nvdimm/pmem.c | 8 ++++----
> 2 files changed, 10 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/nvdimm/btt.c b/drivers/nvdimm/btt.c
> index a933db961ed7..237edfa1c624 100644
> --- a/drivers/nvdimm/btt.c
> +++ b/drivers/nvdimm/btt.c
> @@ -1104,10 +1104,10 @@ static int btt_data_read(struct arena_info *arena, struct page *page,
> {
> int ret;
> u64 nsoff = to_namespace_offset(arena, lba);
> - void *mem = kmap_atomic(page);
> + void *mem = kmap_local_page(page);
>
> ret = arena_read_bytes(arena, nsoff, mem + off, len, NVDIMM_IO_ATOMIC);
> - kunmap_atomic(mem);
> + kunmap_local(mem);
>
> return ret;
> }
> @@ -1117,20 +1117,20 @@ static int btt_data_write(struct arena_info *arena, u32 lba,
> {
> int ret;
> u64 nsoff = to_namespace_offset(arena, lba);
> - void *mem = kmap_atomic(page);
> + void *mem = kmap_local_page(page);
>
> ret = arena_write_bytes(arena, nsoff, mem + off, len, NVDIMM_IO_ATOMIC);
> - kunmap_atomic(mem);
> + kunmap_local(mem);
>
> return ret;
> }
>
> static void zero_fill_data(struct page *page, unsigned int off, u32 len)
> {
> - void *mem = kmap_atomic(page);
> + void *mem = kmap_local_page(page);
>
> memset(mem + off, 0, len);
> - kunmap_atomic(mem);
> + kunmap_local(mem);
> }
>
> #ifdef CONFIG_BLK_DEV_INTEGRITY
> diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c
> index 05785ff21a8b..92c67fbbc1c8 100644
> --- a/drivers/nvdimm/pmem.c
> +++ b/drivers/nvdimm/pmem.c
> @@ -128,10 +128,10 @@ static void write_pmem(void *pmem_addr, struct page *page,
> void *mem;
>
> while (len) {
> - mem = kmap_atomic(page);
> + mem = kmap_local_page(page);
> chunk = min_t(unsigned int, len, PAGE_SIZE - off);
> memcpy_flushcache(pmem_addr, mem + off, chunk);
> - kunmap_atomic(mem);
> + kunmap_local(mem);
> len -= chunk;
> off = 0;
> page++;
> @@ -147,10 +147,10 @@ static blk_status_t read_pmem(struct page *page, unsigned int off,
> void *mem;
>
> while (len) {
> - mem = kmap_atomic(page);
> + mem = kmap_local_page(page);
> chunk = min_t(unsigned int, len, PAGE_SIZE - off);
> rem = copy_mc_to_kernel(mem + off, pmem_addr, chunk);
> - kunmap_atomic(mem);
> + kunmap_local(mem);
> if (rem)
> return BLK_STS_IOERR;
> len -= chunk;
> --
> 2.39.5
>
prev parent reply other threads:[~2026-01-12 23:07 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-28 21:23 Davidlohr Bueso
2025-12-04 15:32 ` Dave Jiang
2026-01-12 23:10 ` Ira Weiny [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=69657f6b74508_f1ef61007a@iweiny-mobl.notmuch \
--to=ira.weiny@intel.com \
--cc=dan.j.williams@intel.com \
--cc=dave.jiang@intel.com \
--cc=dave@stgolabs.net \
--cc=linux-kernel@vger.kernel.org \
--cc=nvdimm@lists.linux.dev \
--cc=vishal.l.verma@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®