From: lizf@kernel.org
To: stable@vger.kernel.org
Cc: linux-kernel@vger.kernel.org,
Mike Marciniszyn <mike.marciniszyn@intel.com>,
Doug Ledford <dledford@redhat.com>,
Ben Hutchings <ben@decadent.org.uk>,
Zefan Li <lizefan@huawei.com>
Subject: [PATCH 3.4 63/65] IB/qib: Change lkey table allocation to support more MRs
Date: Tue, 20 Oct 2015 08:48:13 +0800 [thread overview]
Message-ID: <1445302095-4695-63-git-send-email-lizf@kernel.org> (raw)
In-Reply-To: <1445302030-4607-1-git-send-email-lizf@kernel.org>
From: Mike Marciniszyn <mike.marciniszyn@intel.com>
3.4.110-rc1 review patch. If anyone has any objections, please let me know.
------------------
commit d6f1c17e162b2a11e708f28fa93f2f79c164b442 upstream.
The lkey table is allocated with with a get_user_pages() with an
order based on a number of index bits from a module parameter.
The underlying kernel code cannot allocate that many contiguous pages.
There is no reason the underlying memory needs to be physically
contiguous.
This patch:
- switches the allocation/deallocation to vmalloc/vfree
- caps the number of bits to 23 to insure at least 1 generation bit
o this matches the module parameter description
Reviewed-by: Vinit Agnihotri <vinit.abhay.agnihotri@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
[bwh: Backported to 3.2:
- Adjust context
- Add definition of qib_dev_warn(), added upstream by commit ddb887658970
("IB/qib: Convert opcode counters to per-context")]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Zefan Li <lizefan@huawei.com>
---
drivers/infiniband/hw/qib/qib.h | 4 ++++
drivers/infiniband/hw/qib/qib_keys.c | 4 ++++
drivers/infiniband/hw/qib/qib_verbs.c | 14 ++++++++++----
drivers/infiniband/hw/qib/qib_verbs.h | 2 ++
4 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/drivers/infiniband/hw/qib/qib.h b/drivers/infiniband/hw/qib/qib.h
index c7d4ef1..dcff64f 100644
--- a/drivers/infiniband/hw/qib/qib.h
+++ b/drivers/infiniband/hw/qib/qib.h
@@ -1429,6 +1429,10 @@ extern struct mutex qib_mutex;
qib_get_unit_name((dd)->unit), ##__VA_ARGS__); \
} while (0)
+#define qib_dev_warn(dd, fmt, ...) \
+ dev_warn(&(dd)->pcidev->dev, "%s: " fmt, \
+ qib_get_unit_name((dd)->unit), ##__VA_ARGS__)
+
#define qib_dev_porterr(dd, port, fmt, ...) \
do { \
dev_err(&(dd)->pcidev->dev, "%s: IB%u:%u " fmt, \
diff --git a/drivers/infiniband/hw/qib/qib_keys.c b/drivers/infiniband/hw/qib/qib_keys.c
index 8fd19a4..ca6e6cf 100644
--- a/drivers/infiniband/hw/qib/qib_keys.c
+++ b/drivers/infiniband/hw/qib/qib_keys.c
@@ -69,6 +69,10 @@ int qib_alloc_lkey(struct qib_lkey_table *rkt, struct qib_mregion *mr)
* unrestricted LKEY.
*/
rkt->gen++;
+ /*
+ * bits are capped in qib_verbs.c to insure enough bits
+ * for generation number
+ */
mr->lkey = (r << (32 - ib_qib_lkey_table_size)) |
((((1 << (24 - ib_qib_lkey_table_size)) - 1) & rkt->gen)
<< 8);
diff --git a/drivers/infiniband/hw/qib/qib_verbs.c b/drivers/infiniband/hw/qib/qib_verbs.c
index 7b6c3bf..395d9d6 100644
--- a/drivers/infiniband/hw/qib/qib_verbs.c
+++ b/drivers/infiniband/hw/qib/qib_verbs.c
@@ -40,6 +40,7 @@
#include <linux/rculist.h>
#include <linux/mm.h>
#include <linux/random.h>
+#include <linux/vmalloc.h>
#include "qib.h"
#include "qib_common.h"
@@ -2058,10 +2059,16 @@ int qib_register_ib_device(struct qib_devdata *dd)
* the LKEY). The remaining bits act as a generation number or tag.
*/
spin_lock_init(&dev->lk_table.lock);
+ /* insure generation is at least 4 bits see keys.c */
+ if (ib_qib_lkey_table_size > MAX_LKEY_TABLE_BITS) {
+ qib_dev_warn(dd, "lkey bits %u too large, reduced to %u\n",
+ ib_qib_lkey_table_size, MAX_LKEY_TABLE_BITS);
+ ib_qib_lkey_table_size = MAX_LKEY_TABLE_BITS;
+ }
dev->lk_table.max = 1 << ib_qib_lkey_table_size;
lk_tab_size = dev->lk_table.max * sizeof(*dev->lk_table.table);
dev->lk_table.table = (struct qib_mregion **)
- __get_free_pages(GFP_KERNEL, get_order(lk_tab_size));
+ vmalloc(lk_tab_size);
if (dev->lk_table.table == NULL) {
ret = -ENOMEM;
goto err_lk;
@@ -2231,7 +2238,7 @@ err_tx:
sizeof(struct qib_pio_header),
dev->pio_hdrs, dev->pio_hdrs_phys);
err_hdrs:
- free_pages((unsigned long) dev->lk_table.table, get_order(lk_tab_size));
+ vfree(dev->lk_table.table);
err_lk:
kfree(dev->qp_table);
err_qpt:
@@ -2285,7 +2292,6 @@ void qib_unregister_ib_device(struct qib_devdata *dd)
sizeof(struct qib_pio_header),
dev->pio_hdrs, dev->pio_hdrs_phys);
lk_tab_size = dev->lk_table.max * sizeof(*dev->lk_table.table);
- free_pages((unsigned long) dev->lk_table.table,
- get_order(lk_tab_size));
+ vfree(dev->lk_table.table);
kfree(dev->qp_table);
}
diff --git a/drivers/infiniband/hw/qib/qib_verbs.h b/drivers/infiniband/hw/qib/qib_verbs.h
index 0c19ef0..66f7f62 100644
--- a/drivers/infiniband/hw/qib/qib_verbs.h
+++ b/drivers/infiniband/hw/qib/qib_verbs.h
@@ -622,6 +622,8 @@ struct qib_qpn_table {
struct qpn_map map[QPNMAP_ENTRIES];
};
+#define MAX_LKEY_TABLE_BITS 23
+
struct qib_lkey_table {
spinlock_t lock; /* protect changes in this struct */
u32 next; /* next unused index (speeds search) */
--
1.9.1
next prev parent reply other threads:[~2015-10-20 0:52 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-20 0:47 [PATCH 3.4 00/65] 3.4.110-rc1 review lizf
2015-10-20 0:47 ` [PATCH 3.4 01/65] hrtimer: Allow concurrent hrtimer_start() for self restarting timers lizf
2015-10-20 0:47 ` [PATCH 3.4 02/65] mtd: fix: avoid race condition when accessing mtd->usecount lizf
2015-10-20 0:47 ` [PATCH 3.4 03/65] crypto: talitos - avoid memleak in talitos_alg_alloc() lizf
2015-10-20 0:47 ` [PATCH 3.4 04/65] ASoC: wm8737: Fixup setting VMID Impedance control register lizf
2015-10-20 0:47 ` [PATCH 3.4 05/65] ASoC: wm8903: Fix define for WM8903_VMID_RES_250K lizf
2015-10-20 0:47 ` [PATCH 3.4 06/65] ASoC: wm8955: Fix setting wrong register for WM8955_K_8_0_MASK bits lizf
2015-10-20 0:47 ` [PATCH 3.4 07/65] pktgen: adjust spacing in proc file interface output lizf
2015-10-20 0:47 ` [PATCH 3.4 08/65] pktgen: document ability to add same device to several threads lizf
2015-10-20 0:47 ` [PATCH 3.4 09/65] tty/serial: at91: RS485 mode: 0 is valid for delay_rts_after_send lizf
2015-10-20 0:47 ` [PATCH 3.4 10/65] rndis_wlan: harmless issue calling set_bit() lizf
2015-10-20 0:47 ` [PATCH 3.4 11/65] drm/radeon: take the mode_config mutex when dealing with hpds (v2) lizf
2015-10-20 0:47 ` [PATCH 3.4 12/65] usb: dwc3: gadget: return error if command sent to DEPCMD register fails lizf
2015-10-20 0:47 ` [PATCH 3.4 13/65] rcu: Correctly handle non-empty Tiny RCU callback list with none ready lizf
2015-10-20 0:47 ` [PATCH 3.4 14/65] mtd: dc21285: use raw spinlock functions for nw_gpio_lock lizf
2015-10-20 0:47 ` [PATCH 3.4 15/65] staging: rtl8712: prevent buffer overrun in recvbuf2recvframe lizf
2015-10-20 0:47 ` [PATCH 3.4 16/65] usb: core: Fix USB 3.0 devices lost in NOTATTACHED state after a hub port reset lizf
2015-10-20 0:47 ` [PATCH 3.4 17/65] fixing infinite OPEN loop in 4.0 stateid recovery lizf
[not found] ` <3B7DC48D-0D7F-4F9A-9CE0-FAC640F60199@netapp.com>
2015-10-21 8:04 ` Zefan Li
2015-10-20 0:47 ` [PATCH 3.4 18/65] NFS: Fix size of NFSACL SETACL operations lizf
2015-10-20 0:47 ` [PATCH 3.4 19/65] SUNRPC: Fix a memory leak in the backchannel code lizf
2015-10-20 0:47 ` [PATCH 3.4 20/65] ipr: Increase default adapter init stage change timeout lizf
2015-10-20 0:47 ` [PATCH 3.4 21/65] ath3k: add support of 13d3:3474 AR3012 device lizf
2015-10-20 0:47 ` [PATCH 3.4 22/65] ath9k: fix DMA stop sequence for AR9003+ lizf
2015-10-20 0:47 ` [PATCH 3.4 23/65] regulator: core: fix constraints output buffer lizf
2015-10-20 0:47 ` [PATCH 3.4 24/65] x86/PCI: Use host bridge _CRS info on Foxconn K8M890-8237A lizf
2015-10-20 0:47 ` [PATCH 3.4 25/65] dmaengine: mv_xor: bug fix for racing condition in descriptors cleanup lizf
2015-10-20 0:47 ` [PATCH 3.4 26/65] ASoC: wm8960: the enum of "DAC Polarity" should be wm8960_enum[1] lizf
2015-10-20 0:47 ` [PATCH 3.4 27/65] ext4: fix race between truncate and __ext4_journalled_writepage() lizf
2015-10-20 0:47 ` [PATCH 3.4 28/65] Disable write buffering on Toshiba ToPIC95 lizf
2015-10-20 0:47 ` [PATCH 3.4 29/65] sctp: fix ASCONF list handling lizf
2015-10-20 0:47 ` [PATCH 3.4 30/65] jbd2: use GFP_NOFS in jbd2_cleanup_journal_tail() lizf
2015-10-20 0:47 ` [PATCH 3.4 31/65] regmap: Fix regmap_bulk_read in BE mode lizf
2015-10-20 0:47 ` [PATCH 3.4 32/65] jbd2: fix ocfs2 corrupt when updating journal superblock fails lizf
2015-10-20 0:47 ` [PATCH 3.4 33/65] ideapad: fix software rfkill setting lizf
2015-10-20 0:47 ` [PATCH 3.4 34/65] mmc: card: Fixup request missing in mmc_blk_issue_rw_rq lizf
2015-10-20 0:47 ` [PATCH 3.4 35/65] nfs: increase size of EXCHANGE_ID name string buffer lizf
2015-10-20 0:47 ` [PATCH 3.4 36/65] bridge: fix br_stp_set_bridge_priority race conditions lizf
2015-10-20 0:47 ` [PATCH 3.4 37/65] ext4: call sync_blockdev() before invalidate_bdev() in put_super() lizf
2015-10-20 0:47 ` [PATCH 3.4 38/65] packet: read num_members once in packet_rcv_fanout() lizf
2015-10-20 0:47 ` [PATCH 3.4 39/65] packet: avoid out of bounds read in round robin fanout lizf
2015-10-20 0:47 ` [PATCH 3.4 40/65] ext4: don't retry file block mapping on bigalloc fs with non-extent file lizf
2015-10-20 0:47 ` [PATCH 3.4 41/65] watchdog: omap: assert the counter being stopped before reprogramming lizf
2015-10-20 0:47 ` [PATCH 3.4 42/65] bridge: multicast: restore router configuration on port link down/up lizf
2015-10-20 0:47 ` [PATCH 3.4 43/65] stmmac: troubleshoot unexpected bits in des0 & des1 lizf
2015-10-20 0:47 ` [PATCH 3.4 44/65] mm: kmemleak: allow safe memory scanning during kmemleak disabling lizf
2015-10-20 0:47 ` [PATCH 3.4 45/65] dell-laptop: Fix allocating & freeing SMI buffer page lizf
2015-10-20 0:47 ` [PATCH 3.4 46/65] tracing/filter: Do not WARN on operand count going below zero lizf
2015-10-20 0:47 ` [PATCH 3.4 47/65] tracing/filter: Do not allow infix to exceed end of string lizf
2015-10-20 0:47 ` [PATCH 3.4 48/65] __bitmap_parselist: fix bug in empty string handling lizf
2015-10-20 0:47 ` [PATCH 3.4 49/65] agp/intel: Fix typo in needs_ilk_vtd_wa() lizf
2015-10-20 0:48 ` [PATCH 3.4 50/65] crush: fix a bug in tree bucket decode lizf
2015-10-20 0:48 ` [PATCH 3.4 51/65] fuse: initialize fc->release before calling it lizf
2015-10-20 0:48 ` [PATCH 3.4 52/65] ACPICA: Tables: Fix an issue that FACS initialization is performed twice lizf
2015-10-20 13:35 ` Moore, Robert
2015-10-21 1:24 ` Zheng, Lv
2015-10-20 0:48 ` [PATCH 3.4 53/65] KVM: x86: make vapics_in_nmi_mode atomic lizf
2015-10-20 0:48 ` [PATCH 3.4 54/65] KVM: x86: properly restore LVT0 lizf
2015-10-20 0:48 ` [PATCH 3.4 55/65] 9p: forgetting to cancel request on interrupted zero-copy RPC lizf
2015-10-20 0:48 ` [PATCH 3.4 56/65] Revert "drm/i915: Don't skip request retirement if the active list is empty" lizf
2015-10-20 0:48 ` [PATCH 3.4 57/65] Revert "drm/radeon: Use drm_calloc_ab for CS relocs" lizf
2015-10-20 0:48 ` [PATCH 3.4 58/65] drm/radeon: partially revert "fix VM_CONTEXT*_PAGE_TABLE_END_ADDR handling" lizf
2015-10-20 0:48 ` [PATCH 3.4 59/65] crypto: s390/ghash: Fix incorrect backport of a1cae34e23b1 lizf
2015-10-20 0:48 ` [PATCH 3.4 60/65] ARM: Fix incorrect backport of 0b59d8806a31 lizf
2015-10-20 0:48 ` [PATCH 3.4 61/65] usb: dwc3: Reset the transfer resource index on SET_INTERFACE lizf
2015-10-20 0:48 ` [PATCH 3.4 62/65] jbd2: avoid infinite loop when destroying aborted journal lizf
2015-10-20 0:48 ` lizf [this message]
2015-10-20 0:48 ` [PATCH 3.4 64/65] dcache: Handle escaped paths in prepend_path lizf
2015-10-20 0:48 ` [PATCH 3.4 65/65] vfs: Test for and handle paths that are unreachable from their mnt_root lizf
2015-10-20 1:06 ` [PATCH 3.4 00/65] 3.4.110-rc1 review Zefan Li
2015-10-20 2:17 ` Guenter Roeck
2015-10-20 7:05 ` Geert Uytterhoeven
2015-10-20 8:23 ` Zefan Li
2015-10-20 8:59 ` Geert Uytterhoeven
2015-10-20 12:56 ` Guenter Roeck
2015-10-20 13:15 ` Guenter Roeck
2015-10-21 8:05 ` Zefan Li
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=1445302095-4695-63-git-send-email-lizf@kernel.org \
--to=lizf@kernel.org \
--cc=ben@decadent.org.uk \
--cc=dledford@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=lizefan@huawei.com \
--cc=mike.marciniszyn@intel.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 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).