From: Yinghai Lu <yinghai@kernel.org>
To: Bjorn Helgaas <bhelgaas@google.com>,
David Miller <davem@davemloft.net>,
David Ahern <david.ahern@oracle.com>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Wei Yang <weiyang@linux.vnet.ibm.com>, TJ <linux@iam.tj>,
Yijing Wang <wangyijing@huawei.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org,
Yinghai Lu <yinghai@kernel.org>
Subject: [PATCH 18/36] PCI: Better support for two alt_size
Date: Mon, 6 Jul 2015 16:39:08 -0700 [thread overview]
Message-ID: <1436225966-27247-19-git-send-email-yinghai@kernel.org> (raw)
In-Reply-To: <1436225966-27247-1-git-send-email-yinghai@kernel.org>
Need to put aligned with max align before not aligned.
For example:
alt align/size: 8M/9M, 4M/8M
before this patch we have 8M/20M.
After this patch we will have 8M/17M.
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
---
drivers/pci/setup-bus.c | 80 +++++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 78 insertions(+), 2 deletions(-)
diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
index 9427baa..90dc096 100644
--- a/drivers/pci/setup-bus.c
+++ b/drivers/pci/setup-bus.c
@@ -58,6 +58,20 @@ static inline bool is_before(resource_size_t align1, resource_size_t size1,
return false;
}
+static inline bool is_before_alt(resource_size_t align, resource_size_t size1,
+ resource_size_t size2)
+{
+ resource_size_t size1_left, size2_left;
+
+ /* aligned is before not aligned */
+ size1_left = size1 & (align - 1);
+ size2_left = size2 & (align - 1);
+ if (!size1_left && size2_left)
+ return true;
+
+ return false;
+}
+
struct pci_dev_resource {
struct list_head list;
struct resource *res;
@@ -314,6 +328,42 @@ static void __sort_resources(struct list_head *head)
}
}
+static void __sort_resources_alt(struct list_head *head)
+{
+ struct pci_dev_resource *res1, *tmp_res, *res2;
+ resource_size_t align = 0;
+
+
+ __sort_resources(head);
+
+ /* get max align at first */
+ list_for_each_entry(res1, head, list) {
+ resource_size_t align1;
+
+ align1 = pci_resource_alignment(res1->dev, res1->res);
+ if (align1 > align)
+ align = align1;
+ }
+
+ list_for_each_entry_safe(res1, tmp_res, head, list) {
+ resource_size_t size1, size2;
+
+ size1 = resource_size(res1->res);
+
+ /* reorder it */
+ list_for_each_entry(res2, head, list) {
+ if (res2 == res1)
+ break;
+
+ size2 = resource_size(res2->res);
+ if (is_before_alt(align, size1, size2)) {
+ list_move_tail(&res1->list, &res2->list);
+ break;
+ }
+ }
+ }
+}
+
/**
* reassign_resources_sorted() - satisfy any additional resource requests
*
@@ -641,7 +691,7 @@ static void __assign_resources_alt_sorted(struct list_head *head,
res->end = res->start + alt_res->alt_size - 1;
}
- __sort_resources(head);
+ __sort_resources_alt(head);
/* Satisfy the alt resource requests */
assign_requested_resources_sorted(head, &local_alt_fail_head);
@@ -1236,6 +1286,32 @@ static void __sort_align_test(struct list_head *head)
}
}
+static void __sort_align_test_alt(struct list_head *head)
+{
+ struct align_test_res *res1, *tmp_res, *res2;
+ resource_size_t align = 0;
+
+ __sort_align_test(head);
+
+ /* get max align at first */
+ list_for_each_entry(res1, head, list)
+ if (res1->align > align)
+ align = res1->align;
+
+ list_for_each_entry_safe(res1, tmp_res, head, list) {
+ /* reorder it */
+ list_for_each_entry(res2, head, list) {
+ if (res2 == res1)
+ break;
+
+ if (is_before_alt(align, res1->size, res2->size)) {
+ list_move_tail(&res1->list, &res2->list);
+ break;
+ }
+ }
+ }
+}
+
static bool is_align_size_good(struct list_head *head,
resource_size_t min_align, resource_size_t size,
resource_size_t start)
@@ -1321,7 +1397,7 @@ static resource_size_t calculate_mem_alt_size(struct list_head *head,
if (count <= 1)
goto out;
- __sort_align_test(head);
+ __sort_align_test_alt(head);
tmp = max(size, max_align);
order = __fls(count);
--
1.8.4.5
next prev parent reply other threads:[~2015-07-06 23:41 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-06 23:38 [PATCH 00/36] PCI: Resource allocation cleanup for v4.3 Yinghai Lu
2015-07-06 23:38 ` [PATCH 01/36] PCI: Cleanup res_to_dev_res() printout for addon resources Yinghai Lu
2015-07-06 23:38 ` [PATCH 02/36] PCI: Reuse res_to_dev_res in reassign_resources_sorted Yinghai Lu
2015-07-06 23:38 ` [PATCH 03/36] PCI: Use correct align for optional only resources during sorting for resource allocation Yinghai Lu
2015-07-06 23:38 ` [PATCH 04/36] PCI: Optimize bus align/size calculation during sizing Yinghai Lu
2015-07-06 23:38 ` [PATCH 05/36] PCI: Optimize bus align/size calculation for optional " Yinghai Lu
2015-07-06 23:38 ` [PATCH 06/36] PCI: Reorder resources list for must/optional resources Yinghai Lu
2015-07-06 23:38 ` [PATCH 07/36] PCI: Remove duplicated code for resource sorting Yinghai Lu
2015-07-06 23:38 ` [PATCH 08/36] PCI: Rename pdev_sort_resources to pdev_check_resources Yinghai Lu
2015-07-06 23:38 ` [PATCH 09/36] PCI: Treat ROM resource as optional during realloc Yinghai Lu
2015-07-06 23:39 ` [PATCH 10/36] PCI: Add debug printout during releasing partial must/optinal assigned resources Yinghai Lu
2015-07-06 23:39 ` [PATCH 11/36] PCI: Simplify res reference using in __assign_resourcs_sorted Yinghai Lu
2015-07-06 23:39 ` [PATCH 12/36] PCI: Separate realloc list checking after allocation Yinghai Lu
2015-07-06 23:39 ` [PATCH 13/36] PCI: Add __add_to_list() Yinghai Lu
2015-07-06 23:39 ` [PATCH 14/36] PCI: Separate must_add assigning to another function Yinghai Lu
2015-07-06 23:39 ` [PATCH 15/36] PCI: Bail out early if there is no addon Yinghai Lu
2015-07-06 23:39 ` [PATCH 16/36] PCI: Add alt_size allocation support Yinghai Lu
2015-07-06 23:39 ` [PATCH 17/36] PCI: Add support for more than two alt_size under same bridge Yinghai Lu
2015-07-15 3:07 ` Yijing Wang
2015-07-15 5:08 ` Yinghai Lu
2015-07-15 5:16 ` Yijing Wang
2015-07-06 23:39 ` Yinghai Lu [this message]
2015-07-06 23:39 ` [PATCH 19/36] resources: Split out __allocate_resource() Yinghai Lu
2015-07-06 23:39 ` [PATCH 20/36] resources: Make allocate_resource return just fit resource Yinghai Lu
2015-07-06 23:39 ` [PATCH 21/36] PCI: Check pref compatible bit for mem64 resource of pcie device Yinghai Lu
2015-07-06 23:39 ` [PATCH 22/36] PCI: Only treat non-pef mmio64 as pref if all bridges has MEM_64 Yinghai Lu
2015-07-06 23:39 ` [PATCH 23/36] PCI: Add has_mem64 for host_bridge Yinghai Lu
2015-07-06 23:39 ` [PATCH 24/36] PCI: Only treat non-pef mmio64 as pref if host-bridge has_mem64 Yinghai Lu
2015-07-06 23:39 ` [PATCH 25/36] PCI: Restore pref mmio allocation logic for hostbridge without mmio64 Yinghai Lu
2015-07-06 23:39 ` [PATCH 26/36] sparc/PCI: Add mem64 resource parsing for root bus Yinghai Lu
2015-07-06 23:39 ` [PATCH 27/36] sparc/PCI: Add IORESOURCE_MEM_64 for 64-bit resource in of parsing Yinghai Lu
2015-07-06 23:39 ` [PATCH 28/36] powerpc/PCI: " Yinghai Lu
2015-07-06 23:39 ` [PATCH 29/36] of/PCI: Add IORESOURCE_MEM_64 for 64-bit resource Yinghai Lu
2015-07-10 20:21 ` Rob Herring
2015-07-06 23:39 ` [PATCH 30/36] PCI: Treat optional as must in first try for bridge rescan Yinghai Lu
2015-07-06 23:39 ` [PATCH 31/36] PCI: Get new realloc size for bridge for last try Yinghai Lu
2015-07-06 23:39 ` [PATCH 32/36] PCI: Don't release sibiling bridge resources during hotplug Yinghai Lu
2015-07-06 23:39 ` [PATCH 33/36] PCI: Don't release fixed resource for realloc Yinghai Lu
2015-07-06 23:39 ` [PATCH 34/36] PCI: Set resource to FIXED for lsi devices Yinghai Lu
2015-07-06 23:39 ` [PATCH 35/36] PCI, x86: Add pci=assign_pref_bars to re-allocate pref bars Yinghai Lu
2015-07-06 23:39 ` [PATCH 36/36] PCI: Don't set flags to 0 when assign resource fail Yinghai Lu
2015-07-09 3:30 ` Wei Yang
2015-07-09 5:01 ` Yinghai Lu
2015-07-09 6:04 ` Wei Yang
2015-07-09 16:20 ` Yinghai Lu
2015-07-10 2:30 ` Wei Yang
2015-07-10 2:48 ` Yinghai Lu
2015-07-10 5:49 ` Yinghai Lu
2015-07-11 0:03 ` Wei Yang
2015-07-11 0:42 ` Yinghai Lu
2015-07-11 1:37 ` Wei Yang
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=1436225966-27247-19-git-send-email-yinghai@kernel.org \
--to=yinghai@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=benh@kernel.crashing.org \
--cc=bhelgaas@google.com \
--cc=davem@davemloft.net \
--cc=david.ahern@oracle.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux@iam.tj \
--cc=wangyijing@huawei.com \
--cc=weiyang@linux.vnet.ibm.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;
as well as URLs for NNTP newsgroup(s).