All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Gonda <pgonda@google.com>
To: stable@vger.kernel.org
Cc: Peter Gonda <pgonda@google.com>,
	David Rientjes <rientjes@google.com>,
	Christoph Hellwig <hch@lst.de>
Subject: [PATCH 12/30 for 5.4] dma-pool: scale the default DMA coherent pool size with memory capacity
Date: Fri, 25 Sep 2020 09:18:58 -0700	[thread overview]
Message-ID: <20200925161916.204667-13-pgonda@google.com> (raw)
In-Reply-To: <20200925161916.204667-1-pgonda@google.com>

From: David Rientjes <rientjes@google.com>

upstream 1d659236fb43c4d2b37af7a4309681e834e9ec9a commit.

When AMD memory encryption is enabled, some devices may use more than
256KB/sec from the atomic pools.  It would be more appropriate to scale
the default size based on memory capacity unless the coherent_pool
option is used on the kernel command line.

This provides a slight optimization on initial expansion and is deemed
appropriate due to the increased reliance on the atomic pools.  Note that
the default size of 128KB per pool will normally be larger than the
single coherent pool implementation since there are now up to three
coherent pools (DMA, DMA32, and kernel).

Note that even prior to this patch, coherent_pool= for sizes larger than
1 << (PAGE_SHIFT + MAX_ORDER-1) can fail.  With new dynamic expansion
support, this would be trivially extensible to allow even larger initial
sizes.

Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Peter Gonda <pgonda@google.com>
---
 kernel/dma/pool.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/kernel/dma/pool.c b/kernel/dma/pool.c
index dde6de7f8e83..35bb51c31fff 100644
--- a/kernel/dma/pool.c
+++ b/kernel/dma/pool.c
@@ -20,8 +20,8 @@ static unsigned long pool_size_dma32;
 static struct gen_pool *atomic_pool_kernel __ro_after_init;
 static unsigned long pool_size_kernel;
 
-#define DEFAULT_DMA_COHERENT_POOL_SIZE  SZ_256K
-static size_t atomic_pool_size = DEFAULT_DMA_COHERENT_POOL_SIZE;
+/* Size can be defined by the coherent_pool command line */
+static size_t atomic_pool_size;
 
 /* Dynamic background expansion when the atomic pool is near capacity */
 static struct work_struct atomic_pool_work;
@@ -170,6 +170,16 @@ static int __init dma_atomic_pool_init(void)
 {
 	int ret = 0;
 
+	/*
+	 * If coherent_pool was not used on the command line, default the pool
+	 * sizes to 128KB per 1GB of memory, min 128KB, max MAX_ORDER-1.
+	 */
+	if (!atomic_pool_size) {
+		atomic_pool_size = max(totalram_pages() >> PAGE_SHIFT, 1UL) *
+					SZ_128K;
+		atomic_pool_size = min_t(size_t, atomic_pool_size,
+					 1 << (PAGE_SHIFT + MAX_ORDER-1));
+	}
 	INIT_WORK(&atomic_pool_work, atomic_pool_work_fn);
 
 	atomic_pool_kernel = __dma_atomic_pool_init(atomic_pool_size,
-- 
2.28.0.618.gf4bc123cb7-goog


  parent reply	other threads:[~2020-09-25 16:20 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-25 16:18 [PATCH 00/30 for 5.4] Backport unencrypted non-blocking DMA allocations Peter Gonda
2020-09-25 16:18 ` [PATCH 01/30 for 5.4] dma-direct: remove __dma_direct_free_pages Peter Gonda
2020-09-25 16:18 ` [PATCH 02/30 for 5.4] dma-direct: remove the dma_handle argument to __dma_direct_alloc_pages Peter Gonda
2020-09-25 16:18 ` [PATCH 03/30 for 5.4] dma-direct: provide mmap and get_sgtable method overrides Peter Gonda
2020-09-25 16:18 ` [PATCH 04/30 for 5.4] dma-mapping: merge the generic remapping helpers into dma-direct Peter Gonda
2020-09-25 16:18 ` [PATCH 05/30 for 5.4] lib/genalloc.c: rename addr_in_gen_pool to gen_pool_has_addr Peter Gonda
2020-09-25 16:18 ` [PATCH 06/30 for 5.4] dma-remap: separate DMA atomic pools from direct remap code Peter Gonda
2020-09-25 16:18 ` [PATCH 07/30 for 5.4] dma-pool: add additional coherent pools to map to gfp mask Peter Gonda
2020-09-25 16:18 ` [PATCH 08/30 for 5.4] dma-pool: dynamically expanding atomic pools Peter Gonda
2020-09-25 16:18 ` [PATCH 09/30 for 5.4] dma-direct: atomic allocations must come from atomic coherent pools Peter Gonda
2020-09-25 16:18 ` [PATCH 10/30 for 5.4] dma-pool: add pool sizes to debugfs Peter Gonda
2020-09-25 16:18 ` [PATCH 11/30 for 5.4] x86/mm: unencrypted non-blocking DMA allocations use coherent pools Peter Gonda
2020-09-25 16:18 ` Peter Gonda [this message]
2020-09-25 16:18 ` [PATCH 13/30 for 5.4] dma-pool: fix too large DMA pools on medium memory size systems Peter Gonda
2020-09-25 16:19 ` [PATCH 14/30 for 5.4] dma-pool: decouple DMA_REMAP from DMA_COHERENT_POOL Peter Gonda
2020-09-25 16:19 ` [PATCH 15/30 for 5.4] dma-direct: consolidate the error handling in dma_direct_alloc_pages Peter Gonda
2020-09-25 16:19 ` [PATCH 16/30 for 5.4] xtensa: use the generic uncached segment support Peter Gonda
2020-09-25 16:19 ` [PATCH 17/30 for 5.4] dma-direct: make uncached_kernel_address more general Peter Gonda
2020-09-25 16:19 ` [PATCH 18/30 for 5.4] dma-direct: always align allocation size in dma_direct_alloc_pages() Peter Gonda
2020-09-25 16:19 ` [PATCH 19/30 for 5.4] dma-direct: re-encrypt memory if dma_direct_alloc_pages() fails Peter Gonda
2020-09-25 16:19 ` [PATCH 20/30 for 5.4] dma-direct: check return value when encrypting or decrypting memory Peter Gonda
2020-09-25 16:19 ` [PATCH 21/30 for 5.4] dma-direct: add missing set_memory_decrypted() for coherent mapping Peter Gonda
2020-09-25 16:19 ` [PATCH 22/30 for 5.4] dma-mapping: DMA_COHERENT_POOL should select GENERIC_ALLOCATOR Peter Gonda
2020-09-25 16:19 ` [PATCH 23/30 for 5.4] dma-mapping: warn when coherent pool is depleted Peter Gonda
2020-09-25 16:19 ` [PATCH 24/30 for 5.4] dma-direct: provide function to check physical memory area validity Peter Gonda
2020-09-25 16:19 ` [PATCH 25/30 for 5.4] dma-pool: get rid of dma_in_atomic_pool() Peter Gonda
2020-09-25 16:19 ` [PATCH 26/30 for 5.4] dma-pool: introduce dma_guess_pool() Peter Gonda
2020-09-25 16:19 ` [PATCH 27/30 for 5.4] dma-pool: make sure atomic pool suits device Peter Gonda
2020-09-25 16:19 ` [PATCH 28/30 for 5.4] dma-pool: do not allocate pool memory from CMA Peter Gonda
2020-09-25 16:19 ` [PATCH 29/30 for 5.4] dma-pool: fix coherent pool allocations for IOMMU mappings Peter Gonda
2020-09-25 16:19 ` [PATCH 30/30 for 5.4] dma/direct: turn ARCH_ZONE_DMA_BITS into a variable Peter Gonda
2020-10-05 13:07 ` [PATCH 00/30 for 5.4] Backport unencrypted non-blocking DMA allocations Greg KH
2020-10-06 15:26   ` Peter Gonda
2020-10-06 18:10     ` David Rientjes
2021-01-04 13:00       ` Greg KH
2021-01-04 22:37         ` David Rientjes
2021-01-05  5:58           ` Greg KH
2021-01-05  6:38             ` David Rientjes
2021-05-12 19:06               ` Peter Gonda

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=20200925161916.204667-13-pgonda@google.com \
    --to=pgonda@google.com \
    --cc=hch@lst.de \
    --cc=rientjes@google.com \
    --cc=stable@vger.kernel.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.