linux_dsm_epyc7002/drivers/net/ehea/ehea_qmr.c

1031 lines
23 KiB
C
Raw Normal View History

/*
* linux/drivers/net/ehea/ehea_qmr.c
*
* eHEA ethernet device driver for IBM eServer System p
*
* (C) Copyright IBM Corp. 2006
*
* Authors:
* Christoph Raisch <raisch@de.ibm.com>
* Jan-Bernd Themann <themann@de.ibm.com>
* Thomas Klein <tklein@de.ibm.com>
*
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2, or (at your option)
* any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
#include <linux/mm.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 15:04:11 +07:00
#include <linux/slab.h>
#include "ehea.h"
#include "ehea_phyp.h"
#include "ehea_qmr.h"
struct ehea_bmap *ehea_bmap = NULL;
static void *hw_qpageit_get_inc(struct hw_queue *queue)
{
void *retvalue = hw_qeit_get(queue);
queue->current_q_offset += queue->pagesize;
if (queue->current_q_offset > queue->queue_length) {
queue->current_q_offset -= queue->pagesize;
retvalue = NULL;
} else if (((u64) retvalue) & (EHEA_PAGESIZE-1)) {
ehea_error("not on pageboundary");
retvalue = NULL;
}
return retvalue;
}
static int hw_queue_ctor(struct hw_queue *queue, const u32 nr_of_pages,
const u32 pagesize, const u32 qe_size)
{
int pages_per_kpage = PAGE_SIZE / pagesize;
int i, k;
if ((pagesize > PAGE_SIZE) || (!pages_per_kpage)) {
ehea_error("pagesize conflict! kernel pagesize=%d, "
"ehea pagesize=%d", (int)PAGE_SIZE, (int)pagesize);
return -EINVAL;
}
queue->queue_length = nr_of_pages * pagesize;
queue->queue_pages = kmalloc(nr_of_pages * sizeof(void *), GFP_KERNEL);
if (!queue->queue_pages) {
ehea_error("no mem for queue_pages");
return -ENOMEM;
}
/*
* allocate pages for queue:
* outer loop allocates whole kernel pages (page aligned) and
* inner loop divides a kernel page into smaller hea queue pages
*/
i = 0;
while (i < nr_of_pages) {
u8 *kpage = (u8 *)get_zeroed_page(GFP_KERNEL);
if (!kpage)
goto out_nomem;
for (k = 0; k < pages_per_kpage && i < nr_of_pages; k++) {
(queue->queue_pages)[i] = (struct ehea_page *)kpage;
kpage += pagesize;
i++;
}
}
queue->current_q_offset = 0;
queue->qe_size = qe_size;
queue->pagesize = pagesize;
queue->toggle_state = 1;
return 0;
out_nomem:
for (i = 0; i < nr_of_pages; i += pages_per_kpage) {
if (!(queue->queue_pages)[i])
break;
free_page((unsigned long)(queue->queue_pages)[i]);
}
return -ENOMEM;
}
static void hw_queue_dtor(struct hw_queue *queue)
{
int pages_per_kpage = PAGE_SIZE / queue->pagesize;
int i, nr_pages;
if (!queue || !queue->queue_pages)
return;
nr_pages = queue->queue_length / queue->pagesize;
for (i = 0; i < nr_pages; i += pages_per_kpage)
free_page((unsigned long)(queue->queue_pages)[i]);
kfree(queue->queue_pages);
}
struct ehea_cq *ehea_create_cq(struct ehea_adapter *adapter,
int nr_of_cqe, u64 eq_handle, u32 cq_token)
{
struct ehea_cq *cq;
struct h_epa epa;
u64 *cq_handle_ref, hret, rpage;
u32 act_nr_of_entries, act_pages, counter;
int ret;
void *vpage;
cq = kzalloc(sizeof(*cq), GFP_KERNEL);
if (!cq) {
ehea_error("no mem for cq");
goto out_nomem;
}
cq->attr.max_nr_of_cqes = nr_of_cqe;
cq->attr.cq_token = cq_token;
cq->attr.eq_handle = eq_handle;
cq->adapter = adapter;
cq_handle_ref = &cq->fw_handle;
act_nr_of_entries = 0;
act_pages = 0;
hret = ehea_h_alloc_resource_cq(adapter->handle, &cq->attr,
&cq->fw_handle, &cq->epas);
if (hret != H_SUCCESS) {
ehea_error("alloc_resource_cq failed");
goto out_freemem;
}
ret = hw_queue_ctor(&cq->hw_queue, cq->attr.nr_pages,
EHEA_PAGESIZE, sizeof(struct ehea_cqe));
if (ret)
goto out_freeres;
for (counter = 0; counter < cq->attr.nr_pages; counter++) {
vpage = hw_qpageit_get_inc(&cq->hw_queue);
if (!vpage) {
ehea_error("hw_qpageit_get_inc failed");
goto out_kill_hwq;
}
rpage = virt_to_abs(vpage);
hret = ehea_h_register_rpage(adapter->handle,
0, EHEA_CQ_REGISTER_ORIG,
cq->fw_handle, rpage, 1);
if (hret < H_SUCCESS) {
ehea_error("register_rpage_cq failed ehea_cq=%p "
"hret=%llx counter=%i act_pages=%i",
cq, hret, counter, cq->attr.nr_pages);
goto out_kill_hwq;
}
if (counter == (cq->attr.nr_pages - 1)) {
vpage = hw_qpageit_get_inc(&cq->hw_queue);
if ((hret != H_SUCCESS) || (vpage)) {
ehea_error("registration of pages not "
"complete hret=%llx\n", hret);
goto out_kill_hwq;
}
} else {
if (hret != H_PAGE_REGISTERED) {
ehea_error("CQ: registration of page failed "
"hret=%llx\n", hret);
goto out_kill_hwq;
}
}
}
hw_qeit_reset(&cq->hw_queue);
epa = cq->epas.kernel;
ehea_reset_cq_ep(cq);
ehea_reset_cq_n1(cq);
return cq;
out_kill_hwq:
hw_queue_dtor(&cq->hw_queue);
out_freeres:
ehea_h_free_resource(adapter->handle, cq->fw_handle, FORCE_FREE);
out_freemem:
kfree(cq);
out_nomem:
return NULL;
}
u64 ehea_destroy_cq_res(struct ehea_cq *cq, u64 force)
{
u64 hret;
u64 adapter_handle = cq->adapter->handle;
/* deregister all previous registered pages */
hret = ehea_h_free_resource(adapter_handle, cq->fw_handle, force);
if (hret != H_SUCCESS)
return hret;
hw_queue_dtor(&cq->hw_queue);
kfree(cq);
return hret;
}
int ehea_destroy_cq(struct ehea_cq *cq)
{
u64 hret, aer, aerr;
if (!cq)
return 0;
hcp_epas_dtor(&cq->epas);
hret = ehea_destroy_cq_res(cq, NORMAL_FREE);
if (hret == H_R_STATE) {
ehea_error_data(cq->adapter, cq->fw_handle, &aer, &aerr);
hret = ehea_destroy_cq_res(cq, FORCE_FREE);
}
if (hret != H_SUCCESS) {
ehea_error("destroy CQ failed");
return -EIO;
}
return 0;
}
struct ehea_eq *ehea_create_eq(struct ehea_adapter *adapter,
const enum ehea_eq_type type,
const u32 max_nr_of_eqes, const u8 eqe_gen)
{
int ret, i;
u64 hret, rpage;
void *vpage;
struct ehea_eq *eq;
eq = kzalloc(sizeof(*eq), GFP_KERNEL);
if (!eq) {
ehea_error("no mem for eq");
return NULL;
}
eq->adapter = adapter;
eq->attr.type = type;
eq->attr.max_nr_of_eqes = max_nr_of_eqes;
eq->attr.eqe_gen = eqe_gen;
spin_lock_init(&eq->spinlock);
hret = ehea_h_alloc_resource_eq(adapter->handle,
&eq->attr, &eq->fw_handle);
if (hret != H_SUCCESS) {
ehea_error("alloc_resource_eq failed");
goto out_freemem;
}
ret = hw_queue_ctor(&eq->hw_queue, eq->attr.nr_pages,
EHEA_PAGESIZE, sizeof(struct ehea_eqe));
if (ret) {
ehea_error("can't allocate eq pages");
goto out_freeres;
}
for (i = 0; i < eq->attr.nr_pages; i++) {
vpage = hw_qpageit_get_inc(&eq->hw_queue);
if (!vpage) {
ehea_error("hw_qpageit_get_inc failed");
hret = H_RESOURCE;
goto out_kill_hwq;
}
rpage = virt_to_abs(vpage);
hret = ehea_h_register_rpage(adapter->handle, 0,
EHEA_EQ_REGISTER_ORIG,
eq->fw_handle, rpage, 1);
if (i == (eq->attr.nr_pages - 1)) {
/* last page */
vpage = hw_qpageit_get_inc(&eq->hw_queue);
if ((hret != H_SUCCESS) || (vpage))
goto out_kill_hwq;
} else {
if (hret != H_PAGE_REGISTERED)
goto out_kill_hwq;
}
}
hw_qeit_reset(&eq->hw_queue);
return eq;
out_kill_hwq:
hw_queue_dtor(&eq->hw_queue);
out_freeres:
ehea_h_free_resource(adapter->handle, eq->fw_handle, FORCE_FREE);
out_freemem:
kfree(eq);
return NULL;
}
struct ehea_eqe *ehea_poll_eq(struct ehea_eq *eq)
{
struct ehea_eqe *eqe;
unsigned long flags;
spin_lock_irqsave(&eq->spinlock, flags);
eqe = (struct ehea_eqe *)hw_eqit_eq_get_inc_valid(&eq->hw_queue);
spin_unlock_irqrestore(&eq->spinlock, flags);
return eqe;
}
u64 ehea_destroy_eq_res(struct ehea_eq *eq, u64 force)
{
u64 hret;
unsigned long flags;
spin_lock_irqsave(&eq->spinlock, flags);
hret = ehea_h_free_resource(eq->adapter->handle, eq->fw_handle, force);
spin_unlock_irqrestore(&eq->spinlock, flags);
if (hret != H_SUCCESS)
return hret;
hw_queue_dtor(&eq->hw_queue);
kfree(eq);
return hret;
}
int ehea_destroy_eq(struct ehea_eq *eq)
{
u64 hret, aer, aerr;
if (!eq)
return 0;
hcp_epas_dtor(&eq->epas);
hret = ehea_destroy_eq_res(eq, NORMAL_FREE);
if (hret == H_R_STATE) {
ehea_error_data(eq->adapter, eq->fw_handle, &aer, &aerr);
hret = ehea_destroy_eq_res(eq, FORCE_FREE);
}
if (hret != H_SUCCESS) {
ehea_error("destroy EQ failed");
return -EIO;
}
return 0;
}
/**
* allocates memory for a queue and registers pages in phyp
*/
int ehea_qp_alloc_register(struct ehea_qp *qp, struct hw_queue *hw_queue,
int nr_pages, int wqe_size, int act_nr_sges,
struct ehea_adapter *adapter, int h_call_q_selector)
{
u64 hret, rpage;
int ret, cnt;
void *vpage;
ret = hw_queue_ctor(hw_queue, nr_pages, EHEA_PAGESIZE, wqe_size);
if (ret)
return ret;
for (cnt = 0; cnt < nr_pages; cnt++) {
vpage = hw_qpageit_get_inc(hw_queue);
if (!vpage) {
ehea_error("hw_qpageit_get_inc failed");
goto out_kill_hwq;
}
rpage = virt_to_abs(vpage);
hret = ehea_h_register_rpage(adapter->handle,
0, h_call_q_selector,
qp->fw_handle, rpage, 1);
if (hret < H_SUCCESS) {
ehea_error("register_rpage_qp failed");
goto out_kill_hwq;
}
}
hw_qeit_reset(hw_queue);
return 0;
out_kill_hwq:
hw_queue_dtor(hw_queue);
return -EIO;
}
static inline u32 map_wqe_size(u8 wqe_enc_size)
{
return 128 << wqe_enc_size;
}
struct ehea_qp *ehea_create_qp(struct ehea_adapter *adapter,
u32 pd, struct ehea_qp_init_attr *init_attr)
{
int ret;
u64 hret;
struct ehea_qp *qp;
u32 wqe_size_in_bytes_sq, wqe_size_in_bytes_rq1;
u32 wqe_size_in_bytes_rq2, wqe_size_in_bytes_rq3;
qp = kzalloc(sizeof(*qp), GFP_KERNEL);
if (!qp) {
ehea_error("no mem for qp");
return NULL;
}
qp->adapter = adapter;
hret = ehea_h_alloc_resource_qp(adapter->handle, init_attr, pd,
&qp->fw_handle, &qp->epas);
if (hret != H_SUCCESS) {
ehea_error("ehea_h_alloc_resource_qp failed");
goto out_freemem;
}
wqe_size_in_bytes_sq = map_wqe_size(init_attr->act_wqe_size_enc_sq);
wqe_size_in_bytes_rq1 = map_wqe_size(init_attr->act_wqe_size_enc_rq1);
wqe_size_in_bytes_rq2 = map_wqe_size(init_attr->act_wqe_size_enc_rq2);
wqe_size_in_bytes_rq3 = map_wqe_size(init_attr->act_wqe_size_enc_rq3);
ret = ehea_qp_alloc_register(qp, &qp->hw_squeue, init_attr->nr_sq_pages,
wqe_size_in_bytes_sq,
init_attr->act_wqe_size_enc_sq, adapter,
0);
if (ret) {
ehea_error("can't register for sq ret=%x", ret);
goto out_freeres;
}
ret = ehea_qp_alloc_register(qp, &qp->hw_rqueue1,
init_attr->nr_rq1_pages,
wqe_size_in_bytes_rq1,
init_attr->act_wqe_size_enc_rq1,
adapter, 1);
if (ret) {
ehea_error("can't register for rq1 ret=%x", ret);
goto out_kill_hwsq;
}
if (init_attr->rq_count > 1) {
ret = ehea_qp_alloc_register(qp, &qp->hw_rqueue2,
init_attr->nr_rq2_pages,
wqe_size_in_bytes_rq2,
init_attr->act_wqe_size_enc_rq2,
adapter, 2);
if (ret) {
ehea_error("can't register for rq2 ret=%x", ret);
goto out_kill_hwr1q;
}
}
if (init_attr->rq_count > 2) {
ret = ehea_qp_alloc_register(qp, &qp->hw_rqueue3,
init_attr->nr_rq3_pages,
wqe_size_in_bytes_rq3,
init_attr->act_wqe_size_enc_rq3,
adapter, 3);
if (ret) {
ehea_error("can't register for rq3 ret=%x", ret);
goto out_kill_hwr2q;
}
}
qp->init_attr = *init_attr;
return qp;
out_kill_hwr2q:
hw_queue_dtor(&qp->hw_rqueue2);
out_kill_hwr1q:
hw_queue_dtor(&qp->hw_rqueue1);
out_kill_hwsq:
hw_queue_dtor(&qp->hw_squeue);
out_freeres:
ehea_h_disable_and_get_hea(adapter->handle, qp->fw_handle);
ehea_h_free_resource(adapter->handle, qp->fw_handle, FORCE_FREE);
out_freemem:
kfree(qp);
return NULL;
}
u64 ehea_destroy_qp_res(struct ehea_qp *qp, u64 force)
{
u64 hret;
struct ehea_qp_init_attr *qp_attr = &qp->init_attr;
ehea_h_disable_and_get_hea(qp->adapter->handle, qp->fw_handle);
hret = ehea_h_free_resource(qp->adapter->handle, qp->fw_handle, force);
if (hret != H_SUCCESS)
return hret;
hw_queue_dtor(&qp->hw_squeue);
hw_queue_dtor(&qp->hw_rqueue1);
if (qp_attr->rq_count > 1)
hw_queue_dtor(&qp->hw_rqueue2);
if (qp_attr->rq_count > 2)
hw_queue_dtor(&qp->hw_rqueue3);
kfree(qp);
return hret;
}
int ehea_destroy_qp(struct ehea_qp *qp)
{
u64 hret, aer, aerr;
if (!qp)
return 0;
hcp_epas_dtor(&qp->epas);
hret = ehea_destroy_qp_res(qp, NORMAL_FREE);
if (hret == H_R_STATE) {
ehea_error_data(qp->adapter, qp->fw_handle, &aer, &aerr);
hret = ehea_destroy_qp_res(qp, FORCE_FREE);
}
if (hret != H_SUCCESS) {
ehea_error("destroy QP failed");
return -EIO;
}
return 0;
}
static inline int ehea_calc_index(unsigned long i, unsigned long s)
{
return (i >> s) & EHEA_INDEX_MASK;
}
static inline int ehea_init_top_bmap(struct ehea_top_bmap *ehea_top_bmap,
int dir)
{
if (!ehea_top_bmap->dir[dir]) {
ehea_top_bmap->dir[dir] =
kzalloc(sizeof(struct ehea_dir_bmap), GFP_KERNEL);
if (!ehea_top_bmap->dir[dir])
return -ENOMEM;
}
return 0;
}
static inline int ehea_init_bmap(struct ehea_bmap *ehea_bmap, int top, int dir)
{
if (!ehea_bmap->top[top]) {
ehea_bmap->top[top] =
kzalloc(sizeof(struct ehea_top_bmap), GFP_KERNEL);
if (!ehea_bmap->top[top])
return -ENOMEM;
}
return ehea_init_top_bmap(ehea_bmap->top[top], dir);
}
static DEFINE_MUTEX(ehea_busmap_mutex);
static unsigned long ehea_mr_len;
#define EHEA_BUSMAP_ADD_SECT 1
#define EHEA_BUSMAP_REM_SECT 0
static void ehea_rebuild_busmap(void)
{
u64 vaddr = EHEA_BUSMAP_START;
int top, dir, idx;
for (top = 0; top < EHEA_MAP_ENTRIES; top++) {
struct ehea_top_bmap *ehea_top;
int valid_dir_entries = 0;
if (!ehea_bmap->top[top])
continue;
ehea_top = ehea_bmap->top[top];
for (dir = 0; dir < EHEA_MAP_ENTRIES; dir++) {
struct ehea_dir_bmap *ehea_dir;
int valid_entries = 0;
if (!ehea_top->dir[dir])
continue;
valid_dir_entries++;
ehea_dir = ehea_top->dir[dir];
for (idx = 0; idx < EHEA_MAP_ENTRIES; idx++) {
if (!ehea_dir->ent[idx])
continue;
valid_entries++;
ehea_dir->ent[idx] = vaddr;
vaddr += EHEA_SECTSIZE;
}
if (!valid_entries) {
ehea_top->dir[dir] = NULL;
kfree(ehea_dir);
}
}
if (!valid_dir_entries) {
ehea_bmap->top[top] = NULL;
kfree(ehea_top);
}
}
}
static int ehea_update_busmap(unsigned long pfn, unsigned long nr_pages, int add)
{
unsigned long i, start_section, end_section;
if (!nr_pages)
return 0;
if (!ehea_bmap) {
ehea_bmap = kzalloc(sizeof(struct ehea_bmap), GFP_KERNEL);
if (!ehea_bmap)
return -ENOMEM;
}
start_section = (pfn * PAGE_SIZE) / EHEA_SECTSIZE;
end_section = start_section + ((nr_pages * PAGE_SIZE) / EHEA_SECTSIZE);
/* Mark entries as valid or invalid only; address is assigned later */
for (i = start_section; i < end_section; i++) {
u64 flag;
int top = ehea_calc_index(i, EHEA_TOP_INDEX_SHIFT);
int dir = ehea_calc_index(i, EHEA_DIR_INDEX_SHIFT);
int idx = i & EHEA_INDEX_MASK;
if (add) {
int ret = ehea_init_bmap(ehea_bmap, top, dir);
if (ret)
return ret;
flag = 1; /* valid */
ehea_mr_len += EHEA_SECTSIZE;
} else {
if (!ehea_bmap->top[top])
continue;
if (!ehea_bmap->top[top]->dir[dir])
continue;
flag = 0; /* invalid */
ehea_mr_len -= EHEA_SECTSIZE;
}
ehea_bmap->top[top]->dir[dir]->ent[idx] = flag;
}
ehea_rebuild_busmap(); /* Assign contiguous addresses for mr */
return 0;
}
int ehea_add_sect_bmap(unsigned long pfn, unsigned long nr_pages)
{
int ret;
mutex_lock(&ehea_busmap_mutex);
ret = ehea_update_busmap(pfn, nr_pages, EHEA_BUSMAP_ADD_SECT);
mutex_unlock(&ehea_busmap_mutex);
return ret;
}
int ehea_rem_sect_bmap(unsigned long pfn, unsigned long nr_pages)
{
int ret;
mutex_lock(&ehea_busmap_mutex);
ret = ehea_update_busmap(pfn, nr_pages, EHEA_BUSMAP_REM_SECT);
mutex_unlock(&ehea_busmap_mutex);
return ret;
}
static int ehea_is_hugepage(unsigned long pfn)
{
int page_order;
if (pfn & EHEA_HUGEPAGE_PFN_MASK)
return 0;
page_order = compound_order(pfn_to_page(pfn));
if (page_order + PAGE_SHIFT != EHEA_HUGEPAGESHIFT)
return 0;
return 1;
}
static int ehea_create_busmap_callback(unsigned long initial_pfn,
unsigned long total_nr_pages, void *arg)
{
int ret;
unsigned long pfn, start_pfn, end_pfn, nr_pages;
if ((total_nr_pages * PAGE_SIZE) < EHEA_HUGEPAGE_SIZE)
return ehea_update_busmap(initial_pfn, total_nr_pages,
EHEA_BUSMAP_ADD_SECT);
/* Given chunk is >= 16GB -> check for hugepages */
start_pfn = initial_pfn;
end_pfn = initial_pfn + total_nr_pages;
pfn = start_pfn;
while (pfn < end_pfn) {
if (ehea_is_hugepage(pfn)) {
/* Add mem found in front of the hugepage */
nr_pages = pfn - start_pfn;
ret = ehea_update_busmap(start_pfn, nr_pages,
EHEA_BUSMAP_ADD_SECT);
if (ret)
return ret;
/* Skip the hugepage */
pfn += (EHEA_HUGEPAGE_SIZE / PAGE_SIZE);
start_pfn = pfn;
} else
pfn += (EHEA_SECTSIZE / PAGE_SIZE);
}
/* Add mem found behind the hugepage(s) */
nr_pages = pfn - start_pfn;
return ehea_update_busmap(start_pfn, nr_pages, EHEA_BUSMAP_ADD_SECT);
}
int ehea_create_busmap(void)
{
int ret;
mutex_lock(&ehea_busmap_mutex);
ehea_mr_len = 0;
ret = walk_system_ram_range(0, 1ULL << MAX_PHYSMEM_BITS, NULL,
ehea_create_busmap_callback);
mutex_unlock(&ehea_busmap_mutex);
return ret;
}
void ehea_destroy_busmap(void)
{
int top, dir;
mutex_lock(&ehea_busmap_mutex);
if (!ehea_bmap)
goto out_destroy;
for (top = 0; top < EHEA_MAP_ENTRIES; top++) {
if (!ehea_bmap->top[top])
continue;
for (dir = 0; dir < EHEA_MAP_ENTRIES; dir++) {
if (!ehea_bmap->top[top]->dir[dir])
continue;
kfree(ehea_bmap->top[top]->dir[dir]);
}
kfree(ehea_bmap->top[top]);
}
kfree(ehea_bmap);
ehea_bmap = NULL;
out_destroy:
mutex_unlock(&ehea_busmap_mutex);
}
u64 ehea_map_vaddr(void *caddr)
{
int top, dir, idx;
unsigned long index, offset;
if (!ehea_bmap)
return EHEA_INVAL_ADDR;
index = virt_to_abs(caddr) >> SECTION_SIZE_BITS;
top = (index >> EHEA_TOP_INDEX_SHIFT) & EHEA_INDEX_MASK;
if (!ehea_bmap->top[top])
return EHEA_INVAL_ADDR;
dir = (index >> EHEA_DIR_INDEX_SHIFT) & EHEA_INDEX_MASK;
if (!ehea_bmap->top[top]->dir[dir])
return EHEA_INVAL_ADDR;
idx = index & EHEA_INDEX_MASK;
if (!ehea_bmap->top[top]->dir[dir]->ent[idx])
return EHEA_INVAL_ADDR;
offset = (unsigned long)caddr & (EHEA_SECTSIZE - 1);
return ehea_bmap->top[top]->dir[dir]->ent[idx] | offset;
}
static inline void *ehea_calc_sectbase(int top, int dir, int idx)
{
unsigned long ret = idx;
ret |= dir << EHEA_DIR_INDEX_SHIFT;
ret |= top << EHEA_TOP_INDEX_SHIFT;
return abs_to_virt(ret << SECTION_SIZE_BITS);
}
static u64 ehea_reg_mr_section(int top, int dir, int idx, u64 *pt,
struct ehea_adapter *adapter,
struct ehea_mr *mr)
{
void *pg;
u64 j, m, hret;
unsigned long k = 0;
u64 pt_abs = virt_to_abs(pt);
void *sectbase = ehea_calc_sectbase(top, dir, idx);
for (j = 0; j < (EHEA_PAGES_PER_SECTION / EHEA_MAX_RPAGE); j++) {
for (m = 0; m < EHEA_MAX_RPAGE; m++) {
pg = sectbase + ((k++) * EHEA_PAGESIZE);
pt[m] = virt_to_abs(pg);
}
hret = ehea_h_register_rpage_mr(adapter->handle, mr->handle, 0,
0, pt_abs, EHEA_MAX_RPAGE);
if ((hret != H_SUCCESS) &&
(hret != H_PAGE_REGISTERED)) {
ehea_h_free_resource(adapter->handle, mr->handle,
FORCE_FREE);
ehea_error("register_rpage_mr failed");
return hret;
}
}
return hret;
}
static u64 ehea_reg_mr_sections(int top, int dir, u64 *pt,
struct ehea_adapter *adapter,
struct ehea_mr *mr)
{
u64 hret = H_SUCCESS;
int idx;
for (idx = 0; idx < EHEA_MAP_ENTRIES; idx++) {
if (!ehea_bmap->top[top]->dir[dir]->ent[idx])
continue;
hret = ehea_reg_mr_section(top, dir, idx, pt, adapter, mr);
if ((hret != H_SUCCESS) && (hret != H_PAGE_REGISTERED))
return hret;
}
return hret;
}
static u64 ehea_reg_mr_dir_sections(int top, u64 *pt,
struct ehea_adapter *adapter,
struct ehea_mr *mr)
{
u64 hret = H_SUCCESS;
int dir;
for (dir = 0; dir < EHEA_MAP_ENTRIES; dir++) {
if (!ehea_bmap->top[top]->dir[dir])
continue;
hret = ehea_reg_mr_sections(top, dir, pt, adapter, mr);
if ((hret != H_SUCCESS) && (hret != H_PAGE_REGISTERED))
return hret;
}
return hret;
}
int ehea_reg_kernel_mr(struct ehea_adapter *adapter, struct ehea_mr *mr)
{
int ret;
u64 *pt;
u64 hret;
u32 acc_ctrl = EHEA_MR_ACC_CTRL;
unsigned long top;
ehea: use get_zeroed_page for alignment-sensitive allocation With slub debug enabled, I see the following errors and crash with 2.6.28-rc9: IBM eHEA ethernet device driver (Release EHEA_0095) ehea: Error in ehea_h_register_rpage_mr: not on pageboundary ehea: Error in ehea_reg_mr_section: register_rpage_mr failed ehea: Error in ehea_reg_kernel_mr: registering mr failed ehea: Error in ehea_setup_ports: creating MR failed ehea 23c00100.lhea: setup_ports failed Unable to handle kernel paging request for data at address 0x6b6b6b6b6b6bbdcb Faulting instruction address: 0xd000000000064a24 cpu 0x0: Vector: 300 (Data Access) at [c0000000740e7190] pc: d000000000064a24: .ehea_update_firmware_handles+0x84/0x47c [ehea] lr: d00000000006df34: .ehea_probe_adapter+0x35c/0x39c [ehea] sp: c0000000740e7410 msr: 8000000000009032 dar: 6b6b6b6b6b6bbdcb dsisr: 40000000 current = 0xc000000074233780 paca = 0xc0000000008a3300 pid = 2046, comm = modprobe enter ? for help [c0000000740e74f0] d00000000006df34 .ehea_probe_adapter+0x35c/0x39c [ehea] [c0000000740e75a0] c00000000041d5a4 .of_platform_device_probe+0x78/0xb0 [c0000000740e7630] c0000000002d8b38 .driver_probe_device+0x13c/0x200 [c0000000740e76c0] c0000000002d8c90 .__driver_attach+0x94/0xd8 [c0000000740e7750] c0000000002d7d64 .bus_for_each_dev+0x80/0xd8 [c0000000740e7800] c0000000002d889c .driver_attach+0x28/0x40 [c0000000740e7880] c0000000002d8340 .bus_add_driver+0xd4/0x284 [c0000000740e7920] c0000000002d90a0 .driver_register+0xc4/0x198 [c0000000740e79d0] c00000000041d45c .of_register_driver+0x4c/0x60 [c0000000740e7a50] c000000000020ef8 .ibmebus_register_driver+0x30/0x4c [c0000000740e7ae0] d00000000006e108 .ehea_module_init+0x194/0x208c [ehea] [c0000000740e7b90] c000000000009028 .do_one_initcall+0x90/0x1ac [c0000000740e7d90] c00000000008619c .sys_init_module+0xc4/0x200 [c0000000740e7e30] c0000000000084ac syscall_exit+0x0/0x40 (When slub debug is disabled it works fine.) PAGE_SIZE allocations via slab are not guaranteed to be page-aligned; use get_zeroed_page for the 'pt' buffer (I don't really know what this is, only that it is passed to firmware and that the first error message complains about its alignment). This allows the system to boot. Signed-off-by: Nathan Lynch <ntl@pobox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-12-22 15:42:11 +07:00
pt = (void *)get_zeroed_page(GFP_KERNEL);
if (!pt) {
ehea_error("no mem");
ret = -ENOMEM;
goto out;
}
hret = ehea_h_alloc_resource_mr(adapter->handle, EHEA_BUSMAP_START,
ehea_mr_len, acc_ctrl, adapter->pd,
&mr->handle, &mr->lkey);
if (hret != H_SUCCESS) {
ehea_error("alloc_resource_mr failed");
ret = -EIO;
goto out;
}
if (!ehea_bmap) {
ehea_h_free_resource(adapter->handle, mr->handle, FORCE_FREE);
ehea_error("no busmap available");
ret = -EIO;
goto out;
}
for (top = 0; top < EHEA_MAP_ENTRIES; top++) {
if (!ehea_bmap->top[top])
continue;
hret = ehea_reg_mr_dir_sections(top, pt, adapter, mr);
if((hret != H_PAGE_REGISTERED) && (hret != H_SUCCESS))
break;
}
if (hret != H_SUCCESS) {
ehea_h_free_resource(adapter->handle, mr->handle, FORCE_FREE);
ehea_error("registering mr failed");
ret = -EIO;
goto out;
}
mr->vaddr = EHEA_BUSMAP_START;
mr->adapter = adapter;
ret = 0;
out:
ehea: use get_zeroed_page for alignment-sensitive allocation With slub debug enabled, I see the following errors and crash with 2.6.28-rc9: IBM eHEA ethernet device driver (Release EHEA_0095) ehea: Error in ehea_h_register_rpage_mr: not on pageboundary ehea: Error in ehea_reg_mr_section: register_rpage_mr failed ehea: Error in ehea_reg_kernel_mr: registering mr failed ehea: Error in ehea_setup_ports: creating MR failed ehea 23c00100.lhea: setup_ports failed Unable to handle kernel paging request for data at address 0x6b6b6b6b6b6bbdcb Faulting instruction address: 0xd000000000064a24 cpu 0x0: Vector: 300 (Data Access) at [c0000000740e7190] pc: d000000000064a24: .ehea_update_firmware_handles+0x84/0x47c [ehea] lr: d00000000006df34: .ehea_probe_adapter+0x35c/0x39c [ehea] sp: c0000000740e7410 msr: 8000000000009032 dar: 6b6b6b6b6b6bbdcb dsisr: 40000000 current = 0xc000000074233780 paca = 0xc0000000008a3300 pid = 2046, comm = modprobe enter ? for help [c0000000740e74f0] d00000000006df34 .ehea_probe_adapter+0x35c/0x39c [ehea] [c0000000740e75a0] c00000000041d5a4 .of_platform_device_probe+0x78/0xb0 [c0000000740e7630] c0000000002d8b38 .driver_probe_device+0x13c/0x200 [c0000000740e76c0] c0000000002d8c90 .__driver_attach+0x94/0xd8 [c0000000740e7750] c0000000002d7d64 .bus_for_each_dev+0x80/0xd8 [c0000000740e7800] c0000000002d889c .driver_attach+0x28/0x40 [c0000000740e7880] c0000000002d8340 .bus_add_driver+0xd4/0x284 [c0000000740e7920] c0000000002d90a0 .driver_register+0xc4/0x198 [c0000000740e79d0] c00000000041d45c .of_register_driver+0x4c/0x60 [c0000000740e7a50] c000000000020ef8 .ibmebus_register_driver+0x30/0x4c [c0000000740e7ae0] d00000000006e108 .ehea_module_init+0x194/0x208c [ehea] [c0000000740e7b90] c000000000009028 .do_one_initcall+0x90/0x1ac [c0000000740e7d90] c00000000008619c .sys_init_module+0xc4/0x200 [c0000000740e7e30] c0000000000084ac syscall_exit+0x0/0x40 (When slub debug is disabled it works fine.) PAGE_SIZE allocations via slab are not guaranteed to be page-aligned; use get_zeroed_page for the 'pt' buffer (I don't really know what this is, only that it is passed to firmware and that the first error message complains about its alignment). This allows the system to boot. Signed-off-by: Nathan Lynch <ntl@pobox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2008-12-22 15:42:11 +07:00
free_page((unsigned long)pt);
return ret;
}
int ehea_rem_mr(struct ehea_mr *mr)
{
u64 hret;
if (!mr || !mr->adapter)
return -EINVAL;
hret = ehea_h_free_resource(mr->adapter->handle, mr->handle,
FORCE_FREE);
if (hret != H_SUCCESS) {
ehea_error("destroy MR failed");
return -EIO;
}
return 0;
}
int ehea_gen_smr(struct ehea_adapter *adapter, struct ehea_mr *old_mr,
struct ehea_mr *shared_mr)
{
u64 hret;
hret = ehea_h_register_smr(adapter->handle, old_mr->handle,
old_mr->vaddr, EHEA_MR_ACC_CTRL,
adapter->pd, shared_mr);
if (hret != H_SUCCESS)
return -EIO;
shared_mr->adapter = adapter;
return 0;
}
void print_error_data(u64 *data)
{
int length;
u64 type = EHEA_BMASK_GET(ERROR_DATA_TYPE, data[2]);
u64 resource = data[1];
length = EHEA_BMASK_GET(ERROR_DATA_LENGTH, data[0]);
if (length > EHEA_PAGESIZE)
length = EHEA_PAGESIZE;
if (type == EHEA_AER_RESTYPE_QP)
ehea_error("QP (resource=%llX) state: AER=0x%llX, AERR=0x%llX, "
"port=%llX", resource, data[6], data[12], data[22]);
else if (type == EHEA_AER_RESTYPE_CQ)
ehea_error("CQ (resource=%llX) state: AER=0x%llX", resource,
data[6]);
else if (type == EHEA_AER_RESTYPE_EQ)
ehea_error("EQ (resource=%llX) state: AER=0x%llX", resource,
data[6]);
ehea_dump(data, length, "error data");
}
u64 ehea_error_data(struct ehea_adapter *adapter, u64 res_handle,
u64 *aer, u64 *aerr)
{
unsigned long ret;
u64 *rblock;
u64 type = 0;
rblock = (void *)get_zeroed_page(GFP_KERNEL);
if (!rblock) {
ehea_error("Cannot allocate rblock memory.");
goto out;
}
ret = ehea_h_error_data(adapter->handle, res_handle, rblock);
if (ret == H_SUCCESS) {
type = EHEA_BMASK_GET(ERROR_DATA_TYPE, rblock[2]);
*aer = rblock[6];
*aerr = rblock[12];
print_error_data(rblock);
} else if (ret == H_R_STATE) {
ehea_error("No error data available: %llX.", res_handle);
} else
ehea_error("Error data could not be fetched: %llX", res_handle);
free_page((unsigned long)rblock);
out:
return type;
}