* [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry
@ 2023-03-16 14:09 Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 1/3] dma-debug: small dma_debug_entry's comment and variable name updates Desnes Nunes
` (4 more replies)
0 siblings, 5 replies; 6+ messages in thread
From: Desnes Nunes @ 2023-03-16 14:09 UTC (permalink / raw)
To: iommu, linux-scsi, storagedev, linux-kernel
Cc: hch, martin.petersen, don.brace, m.szyprowski, robin.murphy,
jejb, jsnitsel, Desnes Nunes
In summary, this series fixes an overlapping mappings asymmetry on the
smartpqi driver due to a missing pqi_pci_unmap() call, while also adding
the cacheline on debug messages of dma-debug debugging functions.
Other minor non-functional updates are also provided.
Desnes Nunes (3):
dma-debug: small dma_debug_entry's comment and variable name updates
dma-debug: add cacheline to user/kernel space dump messages
scsi: smartpqi: fix overlapping mappings asymmetry on DMA
drivers/scsi/smartpqi/smartpqi_init.c | 2 +
kernel/dma/debug.c | 133 ++++++++++++++------------
2 files changed, 72 insertions(+), 63 deletions(-)
--
2.39.1
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v2 1/3] dma-debug: small dma_debug_entry's comment and variable name updates
2023-03-16 14:09 [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
@ 2023-03-16 14:09 ` Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 2/3] dma-debug: add cacheline to user/kernel space dump messages Desnes Nunes
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Desnes Nunes @ 2023-03-16 14:09 UTC (permalink / raw)
To: iommu, linux-scsi, storagedev, linux-kernel
Cc: hch, martin.petersen, don.brace, m.szyprowski, robin.murphy,
jejb, jsnitsel, Desnes Nunes
Small update on dma_debug_entry's struct commentary and also standardize
the usage of 'dma_addr' variable name from debug_dma_map_page() on
debug_dma_unmap_page(), and similarly on debug_dma_free_coherent()
Signed-off-by: Desnes Nunes <desnesn@redhat.com>
---
V1 -> V2: Updated comment properly
kernel/dma/debug.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/kernel/dma/debug.c b/kernel/dma/debug.c
index 18c93c2276ca..9aa5100d0d7e 100644
--- a/kernel/dma/debug.c
+++ b/kernel/dma/debug.c
@@ -53,6 +53,7 @@ enum map_err_types {
* struct dma_debug_entry - track a dma_map* or dma_alloc_coherent mapping
* @list: node on pre-allocated free_entries list
* @dev: 'dev' argument to dma_map_{page|single|sg} or dma_alloc_coherent
+ * @dev_addr: dma address
* @size: length of the mapping
* @type: single, page, sg, coherent
* @direction: enum dma_data_direction
@@ -1262,13 +1263,13 @@ void debug_dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
}
EXPORT_SYMBOL(debug_dma_mapping_error);
-void debug_dma_unmap_page(struct device *dev, dma_addr_t addr,
+void debug_dma_unmap_page(struct device *dev, dma_addr_t dma_addr,
size_t size, int direction)
{
struct dma_debug_entry ref = {
.type = dma_debug_single,
.dev = dev,
- .dev_addr = addr,
+ .dev_addr = dma_addr,
.size = size,
.direction = direction,
};
@@ -1403,13 +1404,13 @@ void debug_dma_alloc_coherent(struct device *dev, size_t size,
}
void debug_dma_free_coherent(struct device *dev, size_t size,
- void *virt, dma_addr_t addr)
+ void *virt, dma_addr_t dma_addr)
{
struct dma_debug_entry ref = {
.type = dma_debug_coherent,
.dev = dev,
.offset = offset_in_page(virt),
- .dev_addr = addr,
+ .dev_addr = dma_addr,
.size = size,
.direction = DMA_BIDIRECTIONAL,
};
--
2.39.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v2 2/3] dma-debug: add cacheline to user/kernel space dump messages
2023-03-16 14:09 [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 1/3] dma-debug: small dma_debug_entry's comment and variable name updates Desnes Nunes
@ 2023-03-16 14:09 ` Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 3/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: Desnes Nunes @ 2023-03-16 14:09 UTC (permalink / raw)
To: iommu, linux-scsi, storagedev, linux-kernel
Cc: hch, martin.petersen, don.brace, m.szyprowski, robin.murphy,
jejb, jsnitsel, Desnes Nunes
Having the cacheline also printed on the debug_dma_dump_mappings() and
dump_show() is useful for debugging. Furthermore, this also standardizes
the messages shown on both dump functions.
Signed-off-by: Desnes Nunes <desnesn@redhat.com>
---
kernel/dma/debug.c | 122 ++++++++++++++++++++++++---------------------
1 file changed, 64 insertions(+), 58 deletions(-)
diff --git a/kernel/dma/debug.c b/kernel/dma/debug.c
index e0ad8db1ec25..19cc59829ffd 100644
--- a/kernel/dma/debug.c
+++ b/kernel/dma/debug.c
@@ -396,37 +396,6 @@ static unsigned long long phys_addr(struct dma_debug_entry *entry)
return page_to_phys(pfn_to_page(entry->pfn)) + entry->offset;
}
-/*
- * Dump mapping entries for debugging purposes
- */
-void debug_dma_dump_mappings(struct device *dev)
-{
- int idx;
-
- for (idx = 0; idx < HASH_SIZE; idx++) {
- struct hash_bucket *bucket = &dma_entry_hash[idx];
- struct dma_debug_entry *entry;
- unsigned long flags;
-
- spin_lock_irqsave(&bucket->lock, flags);
-
- list_for_each_entry(entry, &bucket->list, list) {
- if (!dev || dev == entry->dev) {
- dev_info(entry->dev,
- "%s idx %d P=%Lx N=%lx D=%Lx L=%Lx %s %s\n",
- type2name[entry->type], idx,
- phys_addr(entry), entry->pfn,
- entry->dev_addr, entry->size,
- dir2name[entry->direction],
- maperr2str[entry->map_err_type]);
- }
- }
-
- spin_unlock_irqrestore(&bucket->lock, flags);
- cond_resched();
- }
-}
-
/*
* For each mapping (initial cacheline in the case of
* dma_alloc_coherent/dma_map_page, initial cacheline in each page of a
@@ -547,6 +516,70 @@ static void active_cacheline_remove(struct dma_debug_entry *entry)
spin_unlock_irqrestore(&radix_lock, flags);
}
+/*
+ * Dump mappings entries on kernel space for debugging purposes
+ */
+void debug_dma_dump_mappings(struct device *dev)
+{
+ int idx;
+ phys_addr_t cln;
+
+ for (idx = 0; idx < HASH_SIZE; idx++) {
+ struct hash_bucket *bucket = &dma_entry_hash[idx];
+ struct dma_debug_entry *entry;
+ unsigned long flags;
+
+ spin_lock_irqsave(&bucket->lock, flags);
+ list_for_each_entry(entry, &bucket->list, list) {
+ if (!dev || dev == entry->dev) {
+ cln = to_cacheline_number(entry);
+ dev_info(entry->dev,
+ "%s idx %d P=%llx N=%lx D=%llx L=%llx cln=%llx %s %s\n",
+ type2name[entry->type], idx,
+ phys_addr(entry), entry->pfn,
+ entry->dev_addr, entry->size,
+ cln, dir2name[entry->direction],
+ maperr2str[entry->map_err_type]);
+ }
+ }
+ spin_unlock_irqrestore(&bucket->lock, flags);
+
+ cond_resched();
+ }
+}
+
+/*
+ * Dump mappings entries on user space via debugfs
+ */
+static int dump_show(struct seq_file *seq, void *v)
+{
+ int idx;
+ phys_addr_t cln;
+
+ for (idx = 0; idx < HASH_SIZE; idx++) {
+ struct hash_bucket *bucket = &dma_entry_hash[idx];
+ struct dma_debug_entry *entry;
+ unsigned long flags;
+
+ spin_lock_irqsave(&bucket->lock, flags);
+ list_for_each_entry(entry, &bucket->list, list) {
+ cln = to_cacheline_number(entry);
+ seq_printf(seq,
+ "%s %s %s idx %d P=%llx N=%lx D=%llx L=%llx cln=%llx %s %s\n",
+ dev_driver_string(entry->dev),
+ dev_name(entry->dev),
+ type2name[entry->type], idx,
+ phys_addr(entry), entry->pfn,
+ entry->dev_addr, entry->size,
+ cln, dir2name[entry->direction],
+ maperr2str[entry->map_err_type]);
+ }
+ spin_unlock_irqrestore(&bucket->lock, flags);
+ }
+ return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(dump);
+
/*
* Wrapper function for adding an entry to the hash.
* This function takes care of locking itself.
@@ -765,33 +798,6 @@ static const struct file_operations filter_fops = {
.llseek = default_llseek,
};
-static int dump_show(struct seq_file *seq, void *v)
-{
- int idx;
-
- for (idx = 0; idx < HASH_SIZE; idx++) {
- struct hash_bucket *bucket = &dma_entry_hash[idx];
- struct dma_debug_entry *entry;
- unsigned long flags;
-
- spin_lock_irqsave(&bucket->lock, flags);
- list_for_each_entry(entry, &bucket->list, list) {
- seq_printf(seq,
- "%s %s %s idx %d P=%llx N=%lx D=%llx L=%llx %s %s\n",
- dev_name(entry->dev),
- dev_driver_string(entry->dev),
- type2name[entry->type], idx,
- phys_addr(entry), entry->pfn,
- entry->dev_addr, entry->size,
- dir2name[entry->direction],
- maperr2str[entry->map_err_type]);
- }
- spin_unlock_irqrestore(&bucket->lock, flags);
- }
- return 0;
-}
-DEFINE_SHOW_ATTRIBUTE(dump);
-
static int __init dma_debug_fs_init(void)
{
struct dentry *dentry = debugfs_create_dir("dma-api", NULL);
--
2.39.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v2 3/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry
2023-03-16 14:09 [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 1/3] dma-debug: small dma_debug_entry's comment and variable name updates Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 2/3] dma-debug: add cacheline to user/kernel space dump messages Desnes Nunes
@ 2023-03-16 14:09 ` Desnes Nunes
2023-03-28 1:37 ` [PATCH v2 0/3] " Christoph Hellwig
2023-03-29 7:15 ` Geert Uytterhoeven
4 siblings, 0 replies; 6+ messages in thread
From: Desnes Nunes @ 2023-03-16 14:09 UTC (permalink / raw)
To: iommu, linux-scsi, storagedev, linux-kernel
Cc: hch, martin.petersen, don.brace, m.szyprowski, robin.murphy,
jejb, jsnitsel, Desnes Nunes
Currently, pqi_keep_device_offline() calls pqi_build_raid_path_request()
(i.e., a wrapper to pqi_map_single()), but there isn't a pqi_pci_unmap()
call in pqi_keep_device_offline(). Hence, this is flaged as an API viola-
tion due to the overlapping mapping of the physical address.
This patch fixes following warning:
DMA-API: smartpqi 0000:43:00.0: cacheline tracking EEXIST, overlapping
mappings aren't supported
WARNING: CPU: 0 PID: 15 at kernel/dma/debug.c:570
add_dma_entry+0x1f6/0x2f0
CPU: 0 PID: 15 Comm: kworker/0:1 Not tainted 6.3.0-rc1-linux #1
Call Trace:
<TASK>
dma_map_page_attrs+0x6f/0xa0
pqi_map_single+0xaa/0x160 [smartpqi]
pqi_keep_device_offline+0xaf/0x160 [smartpqi]
? dma_unmap_page_attrs+0x1a1/0x1e0
? pqi_identify_physical_device.constprop.0+0xd3/0xe0 [smartpqi]
pqi_update_scsi_devices+0x284/0x7e0 [smartpqi]
pqi_scan_scsi_devices+0x87/0xe0 [smartpqi]
pqi_ctrl_init+0x725/0xa60 [smartpqi]
pqi_pci_probe+0xa7/0x130 [smartpqi]
local_pci_probe+0x46/0xa0
work_for_cpu_fn+0x16/0x20
process_one_work+0x2bf/0x670
? __pfx_worker_thread+0x10/0x10
worker_thread+0x1d1/0x3b0
? __pfx_worker_thread+0x10/0x10
kthread+0xf3/0x120
? __pfx_kthread+0x10/0x10
ret_from_fork+0x2c/0x50
</TASK>
Fixes: be76f90668d8 ("scsi: smartpqi: Add TEST UNIT READY check for SANITIZE operation")
Signed-off-by: Desnes Nunes <desnesn@redhat.com>
---
drivers/scsi/smartpqi/smartpqi_init.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c
index 49a8f91810b6..6fca497ca605 100644
--- a/drivers/scsi/smartpqi/smartpqi_init.c
+++ b/drivers/scsi/smartpqi/smartpqi_init.c
@@ -1726,6 +1726,8 @@ static bool pqi_keep_device_offline(struct pqi_ctrl_info *ctrl_info,
rc = pqi_submit_raid_request_synchronous(ctrl_info, &request.header, 0, &error_info);
+ pqi_pci_unmap(ctrl_info->pci_dev, request.sg_descriptors, 1, dir);
+
if (rc)
goto out; /* Assume not offline */
--
2.39.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry
2023-03-16 14:09 [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
` (2 preceding siblings ...)
2023-03-16 14:09 ` [PATCH v2 3/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
@ 2023-03-28 1:37 ` Christoph Hellwig
2023-03-29 7:15 ` Geert Uytterhoeven
4 siblings, 0 replies; 6+ messages in thread
From: Christoph Hellwig @ 2023-03-28 1:37 UTC (permalink / raw)
To: Desnes Nunes
Cc: iommu, linux-scsi, storagedev, linux-kernel, hch,
martin.petersen, don.brace, m.szyprowski, robin.murphy, jejb,
jsnitsel
Thanks,
I've applied patches 1 and 2 to the dma mapping tree. You might want
to resent patch 1 to the scsi maintainers and list to better get it
onto their radar.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry
2023-03-16 14:09 [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
` (3 preceding siblings ...)
2023-03-28 1:37 ` [PATCH v2 0/3] " Christoph Hellwig
@ 2023-03-29 7:15 ` Geert Uytterhoeven
4 siblings, 0 replies; 6+ messages in thread
From: Geert Uytterhoeven @ 2023-03-29 7:15 UTC (permalink / raw)
To: Desnes Nunes
Cc: iommu, linux-scsi, storagedev, linux-kernel, hch,
martin.petersen, don.brace, m.szyprowski, robin.murphy, jejb,
jsnitsel
Hi Desnes,
On Thu, Mar 16, 2023 at 3:13 PM Desnes Nunes <desnesn@redhat.com> wrote:
> In summary, this series fixes an overlapping mappings asymmetry on the
> smartpqi driver due to a missing pqi_pci_unmap() call, while also adding
> the cacheline on debug messages of dma-debug debugging functions.
>
> Other minor non-functional updates are also provided.
Thanks for your series!
Unfortunately this v2 does not fix the build issues reported by the
bots, thus breaking linux-next on plain 32-bit.
I have sent a fix
https://lore.kernel.org/r/20230329071405.1568021-1-geert+renesas@glider.be
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2023-03-29 7:16 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-03-16 14:09 [PATCH v2 0/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 1/3] dma-debug: small dma_debug_entry's comment and variable name updates Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 2/3] dma-debug: add cacheline to user/kernel space dump messages Desnes Nunes
2023-03-16 14:09 ` [PATCH v2 3/3] scsi: smartpqi: fix DMA overlapping mappings asymmetry Desnes Nunes
2023-03-28 1:37 ` [PATCH v2 0/3] " Christoph Hellwig
2023-03-29 7:15 ` Geert Uytterhoeven
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.