Add persistent durable identifier to storage log messages
Subsystems (6)
v5 cover Add persistent durable identifier to storage log messages
Today users have no easy way to correlate kernel log messages for storage devices across reboots, device dynamic add/remove, or when the device is physically or logically moved from from system to system. This is due to the existing log IDs which identify how the device is attached and not a unique ID of what is attached. Additionally, even when the attachment hasn't changed, it's not always obvious which messages belong to the device as the different areas in the storage stack use different identifiers, eg. (sda, sata1.00, sd 0:0:0:0).
This change addresses this by adding a unique ID to each log message. It couples the existing structured key/value logging capability and VPD 0x83 device identification. The structured key/value data is not visible in normal viewing and is not seen in the dmesg output or journal output unless you go looking for it by dumping the output as JSON.
Some examples of logs filtered for a specific device utilizing this patch series.
$ journalctl -b _KERNEL_DURABLE_NAME="`cat /sys/block/sdb/device/wwid`"
| cut -c 25- | fmt -t
l: scsi 1:0:0:0: Attached scsi generic sg1 type 0
l: sd 1:0:0:0: [sdb] 209715200 512-byte logical blocks: (107 GB/100 GiB)
l: sd 1:0:0:0: [sdb] Write Protect is off
l: sd 1:0:0:0: [sdb] Mode Sense: 00 3a 00 00
l: sd 1:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't
support DPO or FUA
l: sd 1:0:0:0: [sdb] Attached SCSI disk
l: ata2.00: exception Emask 0x0 SAct 0x8 SErr 0x8 action 0x6 frozen
l: ata2.00: failed command: READ FPDMA QUEUED
l: ata2.00: cmd 60/01:18:10:27:00/00:00:00:00:00/40 tag 3 ncq dma 512
in res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
l: ata2.00: status: { DRDY }
l: ata2.00: configured for UDMA/100
l: ata2.00: device reported invalid CHS sector 0
l: ata2.00: exception Emask 0x0 SAct 0x4000 SErr 0x4000 action 0x6 frozen
l: ata2.00: failed command: READ FPDMA QUEUED
l: ata2.00: cmd 60/01:70:10:27:00/00:00:00:00:00/40 tag 14 ncq dma 512
in res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
l: ata2.00: status: { DRDY }
l: ata2.00: configured for UDMA/100
l: ata2.00: device reported invalid CHS sector 0
l: ata2.00: exception Emask 0x0 SAct 0x80000000 SErr 0x80000000 action
0x6 frozen
l: ata2.00: failed command: READ FPDMA QUEUED
l: ata2.00: cmd 60/01:f8:10:27:00/00:00:00:00:00/40 tag 31 ncq dma 512
in res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
l: ata2.00: status: { DRDY }
l: ata2.00: configured for UDMA/100
l: ata2.00: NCQ disabled due to excessive errors
l: ata2.00: exception Emask 0x0 SAct 0x40000 SErr 0x40000 action 0x6
frozen
l: ata2.00: failed command: READ FPDMA QUEUED
l: ata2.00: cmd 60/01:90:10:27:00/00:00:00:00:00/40 tag 18 ncq dma 512
in res 40/00:01:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
l: ata2.00: status: { DRDY }
l: ata2.00: configured for UDMA/100$ journalctl -b _KERNEL_DURABLE_NAME="`cat /sys/block/nvme0n1/wwid`" | cut -c 25- | fmt -t l: blk_update_request: critical medium error, dev nvme0n1, sector 10000 op 0x0:(READ) flags 0x80700 phys_seg 4 prio class 0 l: blk_update_request: critical medium error, dev nvme0n1, sector 10000 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0 l: Buffer I/O error on dev nvme0n1, logical block 1250, async page read
$ journalctl -b _KERNEL_DURABLE_NAME="`cat /sys/block/sdc/device/wwid`" | cut -c 25- | fmt -t l: sd 8:0:0:0: Power-on or device reset occurred l: sd 8:0:0:0: [sdc] 16777216 512-byte logical blocks: (8.59 GB/8.00 GiB) l: sd 8:0:0:0: Attached scsi generic sg2 type 0 l: sd 8:0:0:0: [sdc] Write Protect is off l: sd 8:0:0:0: [sdc] Mode Sense: 63 00 00 08 l: sd 8:0:0:0: [sdc] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA l: sd 8:0:0:0: [sdc] Attached SCSI disk l: sd 8:0:0:0: [sdc] tag#255 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE cmd_age=0s l: sd 8:0:0:0: [sdc] tag#255 Sense Key : Medium Error [current] l: sd 8:0:0:0: [sdc] tag#255 Add. Sense: Unrecovered read error l: sd 8:0:0:0: [sdc] tag#255 CDB: Read(10) 28 00 00 00 27 10 00 00 01 00 l: blk_update_request: critical medium error, dev sdc, sector 10000 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
There should be no changes to the log message content with this patch series. I ran release kernel and this patch series and did a compare while forcing the kernel through the same errors paths to verify.
The first 6 commits in the patch series utilize changes needed for dev_printk code path. The last 6 commits in the patch add the needed changes to utilize durable_name_printk. The function durable_name_printk is nothing more than a printk that adds structured key/value durable name to unmodified printk output. I structured it this way so only a subset of the patch series could be theoretically applied if we cannot get agreement on complete patch series.
v2:
- Incorporated changes suggested by James Bottomley
- Removed string function which removed leading/trailing/duplicate adjacent spaces from generated id, value matches /sys/block/<device>/device/wwid
- Remove xfs patch, limiting changes to lower block layers
- Moved callback from struct device_type to struct device. Struct device_type is typically static const and with a number of different areas using shared implementation of genhd unable to modify for each of the different areas.
v3:
- Increase the size of the buffers for NVMe id generation and dev_vprintk_emit
v4:
- Back out dev_printk for those locations that weren't using it before, so that we don't change the content of the user visible log message by using a function durable_name_printk.
- Remove RFC from patch series.
v5:
- Reduced stack usage for nvme wwid
- Make function ata_scsi_durable_name static, found by kernel test robot
- Incorporated suggested changes from Andy Shevchenko and Sergei Shtylyov
* Remove unneeded line spacing
* Correct spelling
* Remove unneeded () in conditional operator
* Re-worked expressions to follow common kernel patterns, added function dev_to_scsi_device
- Re-based for v5.8 branch
Tony Asleson (12):
struct device: Add function callback durable_name
create_syslog_header: Add durable name
dev_vprintk_emit: Increase hdr size
scsi: Add durable_name for dev_printk
nvme: Add durable name for dev_printk
libata: Add ata_scsi_durable_name
libata: Make ata_scsi_durable_name static Add durable_name_printk
libata: use durable_name_printk
Add durable_name_printk_ratelimited
print_req_error: Use durable_name_printk_ratelimited
buffer_io_error: Use durable_name_printk_ratelimited
block/blk-core.c | 5 ++++- drivers/ata/libata-core.c | 17 +++++++------- drivers/ata/libata-scsi.c | 18 ++++++++++++--- drivers/base/core.c | 46 +++++++++++++++++++++++++++++++++++++- drivers/nvme/host/core.c | 18 +++++++++++++++ drivers/scsi/scsi_lib.c | 9 ++++++++ drivers/scsi/scsi_sysfs.c | 35 +++++++++++++++++++++++------ drivers/scsi/sd.c | 2 ++ fs/buffer.c | 15 +++++++++---- include/linux/dev_printk.h | 14 ++++++++++++ include/linux/device.h | 4 ++++ include/scsi/scsi_device.h | 3 +++ 12 files changed, 162 insertions(+), 24 deletions(-)
base-commit: bcf876870b95592b52519ed4aafcf9d95999bc9c
Signature · 1 line
2.26.2
PATCH v5 1/12 struct device: Add function callback durable_name
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
--- drivers/base/core.c | 24 ++++++++++++++++++++++++ include/linux/device.h | 4 ++++ 2 files changed, 28 insertions(+)
drivers/base/core.c +24 -0
@@ -2489,6 +2489,30 @@ int dev_set_name(struct device *dev, const char *fmt, ...) }EXPORT_SYMBOL_GPL (dev_set_name); +/** +* dev_durable_name - Write "DURABLE_NAME"=<durable name> in buffer +* @dev: device +* @buffer: character buffer to write results +* @len: length of buffer +* @return: Number of bytes written to buffer +*/ +int dev_durable_name (const struct device *dev,char *buffer,size_t len) +{ +int tmp, dlen; + +if (dev && dev->durable_name) { + tmp =snprintf (buffer, len,"DURABLE_NAME=" ); +if (tmp < len) { + dlen = dev->durable_name (dev, buffer + tmp, + len - tmp); +if (dlen >0 && ((dlen + tmp) < len)) +return dlen + tmp; + } + } +return 0 ; +} +EXPORT_SYMBOL_GPL (dev_durable_name); +/** * device_to_dev_kobj - select a /sys/dev/ directory for the device * @dev: device
include/linux/device.h +4 -0
@@ -614,6 +614,8 @@ struct device {struct iommu_group *iommu_group;struct dev_iommu *iommu; +int (*durable_name)(const struct device *dev,char *buff,size_t len); + bool offline_disabled:1; bool offline:1; bool of_node_reused:1;
@@ -655,6 +657,8 @@ static inline const char *dev_name(const struct device *dev) extern __printf(2, 3) int dev_set_name(struct device *dev, const char *name, ...); +int dev_durable_name(const struct device *d, char *buffer, size_t len); + #ifdef CONFIG_NUMA static inline int dev_to_node(struct device *dev) {
Signature · 1 line
2.26.2
Re: [v5 01/12] struct device: Add function callback durable_name
Quoted 16 lines · Tony Asleson
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
Signed-off-by Tony Asleson <[email protected]>--- drivers/base/core.c | 24 ++++++++++++++++++++++++ include/linux/device.h | 4 ++++ 2 files changed, 28 insertions(+)drivers/base/core.c +0
-0
drivers/base/core.c int dev_set_name(struct device *dev, const char *fmt, ...) · line 2497
Quoted 9 lines
drivers/base/core.c +6
-0@@ -2489,6 +2489,30 @@ int dev_set_name(struct device *dev, const char *fmt, ...) }EXPORT_SYMBOL_GPL (dev_set_name); +/** +* dev_durable_name - Write "DURABLE_NAME"=<durable name> in buffer +* @dev: device +* @buffer: character buffer to write results +* @len: length of buffer +* @return: Number of bytes written to buffer
This is not how the kernel-doc commenta describe the function result, IIRC...
Quoted 11 lines
drivers/base/core.c +11
-0+*/ +int dev_durable_name (const struct device *dev,char *buffer,size_t len) +{ +int tmp, dlen; + +if (dev && dev->durable_name) { + tmp =snprintf (buffer, len,"DURABLE_NAME=" ); +if (tmp < len) { + dlen = dev->durable_name (dev, buffer + tmp, + len - tmp); +if (dlen >0 && ((dlen + tmp) < len))
drivers/base/core.c int dev_set_name(struct device *dev, const char *fmt, ...) · line 2518
Quoted 10 lines
drivers/base/core.c +7
-0+return dlen + tmp; + } + } +return 0 ; +} +EXPORT_SYMBOL_GPL (dev_durable_name); +/** * device_to_dev_kobj - select a /sys/dev/ directory for the device * @dev: device
[...]
MBR, Sergei
Re: [v5 01/12] struct device: Add function callback durable_name
Quoted 22 lines · Sergei Shtylyov
Quoted 20 lines · Tony Asleson
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
Signed-off-by Tony Asleson <[email protected]>---
drivers/base/core.c | 24 ++++++++++++++++++++++++include/linux/device.h| 4 ++++2 files changed, 28 insertions(+)drivers/base/core.c +0
-0@@ -2489,6 +2489,30 @@ int dev_set_name(struct device *dev, const char *fmt, ...) } EXPORT_SYMBOL_GPL(dev_set_name);+/**
Quoted 8 lines
Quoted 5 lines
+ *
dev_durable_name- Write "DURABLE_NAME"=<durable name> in buffer
+ * @dev: device
+ * @buffer: character buffer to write results
+ * @len: length of buffer
+ * @return: Number of bytes written to bufferThis is not how the kernel-doc commenta describe the function result, IIRC...
I did my compile with make W=1 and there isn't any warnings/error with source documentation, but the documentation does indeed outline a different syntax. It's interesting how common the @return syntax is in the existing code base.
I'll re-work the function documentation return.
Thanks
Re: [v5 01/12] struct device: Add function callback durable_name
Quoted 29 lines · Tony Asleson
Quoted 20 lines
Quoted 20 lines
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
Signed-off-by Tony Asleson <[email protected]>--- drivers/base/core.c | 24 ++++++++++++++++++++++++ include/linux/device.h | 4 ++++ 2 files changed, 28 insertions(+)drivers/base/core.c +0
-0@@ -2489,6 +2489,30 @@ int dev_set_name(struct device *dev, const char *fmt, ...) } EXPORT_SYMBOL_GPL(dev_set_name); +/**Quoted 8 lines
drivers/base/core.c int dev_set_name(struct device *dev, const char *fmt, ...) · line 2497
Quoted 5 lines
drivers/base/core.c +5
-0+* dev_durable_name - Write "DURABLE_NAME"=<durable name> in buffer +* @dev: device +* @buffer: character buffer to write results +* @len: length of buffer +* @return: Number of bytes written to buffer This is not how the kernel-doc commenta describe the function result,
IIRC...
Quoted 2 lines
I did my compile with
make W=1and there isn't any warnings/error with source documentation, but the documentation does indeed outline a
IIRC, you only get the warnings when you try to build the kernel-docs.
Quoted 2 lines
different syntax. It's interesting how common the @return syntax is in the existing code base.
FWIW, I'm seeing @return: for the 1st time in my Linux tenure (since 2004).
Quoted 1 line
I'll re-work the function documentation return.
OK, thanks. :-)
Quoted 1 line
Thanks
MBR, Sergei
Re: [v5 01/12] struct device: Add function callback durable_name
In reply to Tony AslesonIndepend of my opinion on the whole scheme that I shared last time, we really should not bloat struct device with function pointers.
Quoted 71 lines · Tony Asleson
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
Signed-off-by Tony Asleson <[email protected]>--- drivers/base/core.c | 24 ++++++++++++++++++++++++ include/linux/device.h | 4 ++++ 2 files changed, 28 insertions(+)drivers/base/core.c +24
-0@@ -2489,6 +2489,30 @@ int dev_set_name(struct device *dev, const char *fmt, ...) }EXPORT_SYMBOL_GPL (dev_set_name); +/** +* dev_durable_name - Write "DURABLE_NAME"=<durable name> in buffer +* @dev: device +* @buffer: character buffer to write results +* @len: length of buffer +* @return: Number of bytes written to buffer +*/ +int dev_durable_name (const struct device *dev,char *buffer,size_t len) +{ +int tmp, dlen; + +if (dev && dev->durable_name) { + tmp =snprintf (buffer, len,"DURABLE_NAME=" ); +if (tmp < len) { + dlen = dev->durable_name (dev, buffer + tmp, + len - tmp); +if (dlen >0 && ((dlen + tmp) < len)) +return dlen + tmp; + } + } +return 0 ; +} +EXPORT_SYMBOL_GPL (dev_durable_name); +/** * device_to_dev_kobj - select a /sys/dev/ directory for the device * @dev: device include/linux/device.h +4
-0@@ -614,6 +614,8 @@ struct device {struct iommu_group *iommu_group;struct dev_iommu *iommu; +int (*durable_name)(const struct device *dev,char *buff,size_t len); + bool offline_disabled:1; bool offline:1; bool of_node_reused:1;@@ -655,6 +657,8 @@ static inline const char *dev_name(const struct device *dev) extern __printf(2, 3) int dev_set_name(struct device *dev, const char *name, ...); +int dev_durable_name(const struct device *d, char *buffer, size_t len); + #ifdef CONFIG_NUMA static inline int dev_to_node(struct device *dev) {Signature · 1 line
2.26.2
---end quoted text---
Greg Kroah-Hartman
maintainerRe: [v5 01/12] struct device: Add function callback durable_name
Quoted 14 lines · Christoph Hellwig
Independ of my opinion on the whole scheme that I shared last time, we really should not bloat struct device with function pointers.
Quoted 10 lines · Tony Asleson
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
Signed-off-by Tony Asleson <[email protected]>--- drivers/base/core.c | 24 ++++++++++++++++++++++++ include/linux/device.h | 4 ++++ 2 files changed, 28 insertions(+)
I can't find this patch anywhere in my archives, why was I not cc:ed on it? It's a v5 and no one thought to ask the driver core developers/maintainers about it???
{sigh}
And for log messages, what about the dynamic debug developers, why not include them as well? Since when is this a storage-only thing?
Quoted 44 lines
Quoted 44 lines
drivers/base/core.c +24
-0@@ -2489,6 +2489,30 @@ int dev_set_name(struct device *dev, const char *fmt, ...) }EXPORT_SYMBOL_GPL (dev_set_name); +/** +* dev_durable_name - Write "DURABLE_NAME"=<durable name> in buffer +* @dev: device +* @buffer: character buffer to write results +* @len: length of buffer +* @return: Number of bytes written to buffer +*/ +int dev_durable_name (const struct device *dev,char *buffer,size_t len) +{ +int tmp, dlen; + +if (dev && dev->durable_name) { + tmp =snprintf (buffer, len,"DURABLE_NAME=" ); +if (tmp < len) { + dlen = dev->durable_name (dev, buffer + tmp, + len - tmp); +if (dlen >0 && ((dlen + tmp) < len)) +return dlen + tmp; + } + } +return 0 ; +} +EXPORT_SYMBOL_GPL (dev_durable_name); +/** * device_to_dev_kobj - select a /sys/dev/ directory for the device * @dev: device include/linux/device.h +0
-0@@ -614,6 +614,8 @@ struct device {struct iommu_group *iommu_group;struct dev_iommu *iommu;
Quoted 1 line
include/linux/device.h struct device { · line 617
Quoted 1 line
include/linux/device.h +1
-0+int (*durable_name)(const struct device *dev,char *buff,size_t len);
No, that's not ok at all, why is this even a thing?
Who is setting this? Why can't the bus do this without anything "special" needed from the driver core?
We have a mapping of 'struct device' to a unique hardware device at a specific point in time, why are you trying to create another one?
What is wrong with what we have today?
So this is a HARD NAK on this patch for now.
thanks,
greg k-h
Re: [v5 01/12] struct device: Add function callback durable_name
Quoted 3 lines · Greg Kroah-Hartman
Quoted 2 lines · Christoph Hellwig
Independ of my opinion on the whole scheme that I shared last time, we really should not bloat struct device with function pointers.
Christoph, thank you for sharing a bit more of your concerns and bringing Greg into the discussion. It's more productive.
Quoted 13 lines
Quoted 7 lines · Tony Asleson
Quoted 5 lines
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
Quoted 5 lines
Quoted 5 lines
Signed-off-by Tony Asleson <[email protected]>--- drivers/base/core.c | 24 ++++++++++++++++++++++++ include/linux/device.h | 4 ++++ 2 files changed, 28 insertions(+)
Quoted 3 lines
I can't find this patch anywhere in my archives, why was I not cc:ed on it? It's a v5 and no one thought to ask the driver core developers/maintainers about it???
You were CC'd into v3, ref.
https://lore.kernel.org/linux-ide/[email protected]/
I should have continued to CC you on it, sorry about that.
Quoted 4 lines
{sigh}
And for log messages, what about the dynamic debug developers, why not include them as well? Since when is this a storage-only thing?
Hannes Reinecke has been involved in the discussion some and he's involved in dynamic debug AFAIK.
If others have a need to identify a specific piece of hardware in a potential sea of identical hardware that is encountering errors and logging messages and can optionally be added and removed at run-time, then yes they should be included too. There is nothing with this patch series that is preventing any device from registering a callback which provides this information when asked.
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
include/linux/device.h +0
-0@@ -614,6 +614,8 @@ struct device {struct iommu_group *iommu_group;struct dev_iommu *iommu;
Quoted 6 lines
Quoted 1 line
include/linux/device.h struct device { · line 617
Quoted 1 line
include/linux/device.h +1
-0+int (*durable_name)(const struct device *dev,char *buff,size_t len);No, that's not ok at all, why is this even a thing?
Who is setting this? Why can't the bus do this without anything "special" needed from the driver core?
I'm setting it in the different storage subsystems. The intent is that when we go through a common logging path eg. dev_printk you can ask the device what it's ID is so that it can be logged as structured data with the log message. I was trying to avoid having separate logging functions all over the place which enforces a consistent meta data to log messages, but maybe that would be better than adding a function pointer to struct device? My first patch tried adding a call back to struct device_type, but that ran into the issue where struct device_type is static const in a number of areas.
Basically for any piece of hardware with a serial number, call this function to get it. That was the intent anyway.
Quoted 2 lines
We have a mapping of 'struct device' to a unique hardware device at a specific point in time, why are you trying to create another one?
I don't think I'm creating anything new. Can you clarify this a bit more? I'm trying to leverage what is already in place.
Quoted 1 line
What is wrong with what we have today?
I'm trying to figure out a way to positively identify which storage device an error belongs to over time. Logging how something is attached and not what is attached, only solves for right now, this point in time.
Additionally when the only identifier is in the actual message itself,
it makes user space break every time the message content changes. Also what we have today in dev_printk doesn't tag the meta-data consistently for journald to leverage.
This patch series adds structured data to the log entry for positive identification. It doesn't change when the content of the log message changes. It's true over time and it's true if a device gets moved to a different system.
Quoted 1 line
So this is a HARD NAK on this patch for now.
Thank you for supplying some feedback and asking questions. I've been asking for suggestions and would very much like to have a discussion on how this issue is best solved. I'm not attached to what I've provided. I'm just trying to get towards a solution.
We've looked at user space quite a bit and there is an inherit race condition with trying to fetch the unique hardware id for a message when it gets emitted from the kernel as udev rules haven't even run (assuming we even have the meta-data to make the association). The last thing people want to do is delay writing the log message to disk until the device it belongs to can be identified. Of course this patch series still has a window from where the device is first discovered by the kernel and fetches the needed vpd data from the device. Any kernel messages logged during that time have no id to associate with it.
Thanks,
Tony
Greg Kroah-Hartman
maintainerRe: [v5 01/12] struct device: Add function callback durable_name
Quoted 21 lines · Tony Asleson
Quoted 3 lines · Greg Kroah-Hartman
Quoted 2 lines · Christoph Hellwig
Independ of my opinion on the whole scheme that I shared last time, we really should not bloat struct device with function pointers.
Christoph, thank you for sharing a bit more of your concerns and bringing Greg into the discussion. It's more productive.
Quoted 13 lines
Quoted 12 lines · Tony Asleson
Quoted 10 lines
Function callback and function to be used to write a persistent durable name to the supplied character buffer. This will be used to add structured key-value data to log messages for hardware related errors which allows end users to correlate message and specific hardware.
Signed-off-by Tony Asleson <[email protected]>--- drivers/base/core.c | 24 ++++++++++++++++++++++++ include/linux/device.h | 4 ++++ 2 files changed, 28 insertions(+)
Quoted 7 lines
Quoted 3 lines
I can't find this patch anywhere in my archives, why was I not cc:ed on it? It's a v5 and no one thought to ask the driver core developers/maintainers about it???
You were CC'd into v3, ref.
Yeah, and I rejected that patch too, no wonder you didn't include me in further patch reviews, you didn't want me to reject them either :)
Quoted 7 lines
Quoted 4 lines
{sigh}
And for log messages, what about the dynamic debug developers, why not include them as well? Since when is this a storage-only thing?
Hannes Reinecke has been involved in the discussion some and he's involved in dynamic debug AFAIK.
From the maintainers file:
DYNAMIC DEBUG
M: Jason Baron <[email protected]>
S: Maintained
F: include/linux/dynamic_debug.h
F: lib/dynamic_debug.c
Come on, you know this, don't try to avoid the people who have to maintain the code you are wanting to change, that's not ok.
Quoted 6 lines
If others have a need to identify a specific piece of hardware in a potential sea of identical hardware that is encountering errors and logging messages and can optionally be added and removed at run-time, then yes they should be included too. There is nothing with this patch series that is preventing any device from registering a callback which provides this information when asked.
But that's not what the kernel is supposed to be doing, it doesn't care about tracking things outside of the lifetime from when it can see a device. That's what userspace is for.
Quoted 25 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
include/linux/device.h +0
-0@@ -614,6 +614,8 @@ struct device {struct iommu_group *iommu_group;struct dev_iommu *iommu;Quoted 6 lines
Quoted 1 line
include/linux/device.h struct device { · line 617
Quoted 1 line
include/linux/device.h +1
-0+int (*durable_name)(const struct device *dev,char *buff,size_t len);No, that's not ok at all, why is this even a thing?
Who is setting this? Why can't the bus do this without anything "special" needed from the driver core?
I'm setting it in the different storage subsystems. The intent is that when we go through a common logging path eg.
dev_printkyou can ask the device what it's ID is so that it can be logged as structured data with the log message. I was trying to avoid having separate logging functions all over the place which enforces a consistent meta data to log messages, but maybe that would be better than adding a function pointer to struct device? My first patch tried adding a call back to structdevice_type, but that ran into the issue where structdevice_typeis static const in a number of areas.
Quoted 2 lines
Basically for any piece of hardware with a serial number, call this function to get it. That was the intent anyway.
That's not ok, again, if you really want something crazy like this, then modify your logging messages to provide it. But I would strongly argue that you do not really need this, as you have this information at runtime, in userspace, already.
Quoted 10 lines
Quoted 2 lines
We have a mapping of 'struct device' to a unique hardware device at a specific point in time, why are you trying to create another one?
I don't think I'm creating anything new. Can you clarify this a bit more? I'm trying to leverage what is already in place.
Quoted 1 line
What is wrong with what we have today?
I'm trying to figure out a way to positively identify which storage device an error belongs to over time.
"over time" is not the kernel's responsibility.
This comes up every 5 years or so. The kernel provides you, at runtime, a mapping between a hardware device and a "logical" device. It can provide information to userspace about this mapping, but once that device goes away, the kernel is free to reuse that logical device again.
If you want to track what logical devices match up to what physical device, then do it in userspace, by parsing the log files. All of the needed information is already there, you don't have to add anything new to it.
Again, this comes up every few years for some reason, because people feel it is eaasier to modify the kernel than work with userspace programs. Odd...
Quoted 6 lines
Quoted 1 line
So this is a HARD NAK on this patch for now.
Thank you for supplying some feedback and asking questions. I've been asking for suggestions and would very much like to have a discussion on how this issue is best solved. I'm not attached to what I've provided. I'm just trying to get towards a solution.
Again, solve this in userspace, you have the information there at runtime, why not use it?
Quoted 4 lines
We've looked at user space quite a bit and there is an inherit race condition with trying to fetch the unique hardware id for a message when it gets emitted from the kernel as udev rules haven't even run (assuming we even have the meta-data to make the association).
But one moment later you do have the information, so you can properly correlate it, right?
Quoted 6 lines
The last thing
people want to do is delay writing the log message to disk until the device it belongs to can be identified. Of course this patch series still has a window from where the device is first discovered by the kernel and fetches the needed vpd data from the device. Any kernel messages logged during that time have no id to associate with it.
No need to delay logging, you are looking at things way later in time, so there's no issue of race conditions or anything else.
thanks,
greg k-h
Greg Kroah-Hartman
maintainerRe: [v5 01/12] struct device: Add function callback durable_name
Quoted 15 lines · Greg Kroah-Hartman
Quoted 7 lines
Quoted 4 lines
{sigh}
And for log messages, what about the dynamic debug developers, why not include them as well? Since when is this a storage-only thing?
Hannes Reinecke has been involved in the discussion some and he's involved in dynamic debug AFAIK.
From the maintainers file:
DYNAMIC DEBUG
M: Jason Baron <[email protected]>
S: Maintained
F:include/linux/dynamic_debug.h
F:lib/dynamic_debug.c
Quoted 2 lines
Come on, you know this, don't try to avoid the people who have to maintain the code you are wanting to change, that's not ok.
Also, why are you not using scripts/get_maintainer.pl on your patches? It would have told you about this...
greg k-h
Re: [v5 01/12] struct device: Add function callback durable_name
In reply to Greg Kroah-HartmanQuoted 6 lines · Greg Kroah-Hartman
Quoted 2 lines · Tony Asleson
I'm trying to figure out a way to positively identify which storage device an error belongs to over time.
"over time" is not the kernel's responsibility.
Quoted 7 lines
This comes up every 5 years or so. The kernel provides you, at runtime, a mapping between a hardware device and a "logical" device. It can provide information to userspace about this mapping, but once that device goes away, the kernel is free to reuse that logical device again.
If you want to track what logical devices match up to what physical device, then do it in userspace, by parsing the log files.
I don't understand why people think it's acceptable to ask user space to parse text that is subject to change.
Quoted 7 lines
Quoted 4 lines
Thank you for supplying some feedback and asking questions. I've been asking for suggestions and would very much like to have a discussion on how this issue is best solved. I'm not attached to what I've provided. I'm just trying to get towards a solution.
Again, solve this in userspace, you have the information there at runtime, why not use it?
We usually don't have the needed information if you remove the expectation that user space should parse the human readable portion of the error message.
Quoted 7 lines
Quoted 4 lines
We've looked at user space quite a bit and there is an inherit race condition with trying to fetch the unique hardware id for a message when it gets emitted from the kernel as udev rules haven't even run (assuming we even have the meta-data to make the association).
But one moment later you do have the information, so you can properly correlate it, right?
We could have the information if all the storage paths went through dev_printk. Here is what we get today when we encounter a read error which uses printk in the block layer:
{
"_HOSTNAME" : "pn",
"_TRANSPORT" : "kernel",
"__MONOTONIC_TIMESTAMP" : "1806379233",
"SYSLOG_IDENTIFIER" : "kernel",
"_SOURCE_MONOTONIC_TIMESTAMP" : "1805611354",
"SYSLOG_FACILITY" : "0",
"MESSAGE" : "blk_update_request: critical medium error, dev
nvme0n1, sector 10000 op 0x0:(READ) flags 0x80700 phys_seg 3 prio class 0",
"PRIORITY" : "3",
"_MACHINE_ID" : "3f31a0847cea4c95b7a9cec13d07deeb",
"__REALTIME_TIMESTAMP" : "1601471260802301",
"_BOOT_ID" : "b03ed610f21d46ab8243a495ba5a0058",
"__CURSOR" :"s=a063a22bbb384da0b0412e8f652deabb;i=23c2;b=b03ed610f21d46ab8243a495ba5a0058;m=6bab28e1;t=5b087959e3cfd;x=20528862f8f765c9" }
Unless you parse the message text you cannot make the association. If the same message was changed to dev_printk we would get:
{
"__REALTIME_TIMESTAMP" : "1589401901093443",
"__CURSOR" :
"s=caac9703b34a48fd92f7875adae55a2f;i=1c713;b=e2ae14a9def345aa803a13648b95429c;m=7d25b4f;t=5a58d77b85243;x=b034c2d3fb853870",
"SYSLOG_IDENTIFIER" : "kernel",
"_KERNEL_DEVICE" : "b259:917504",
"__MONOTONIC_TIMESTAMP" : "131226447",
"_UDEV_SYSNAME" : "nvme0n1",
"PRIORITY" : "3",
"_KERNEL_SUBSYSTEM" : "block",
"_SOURCE_MONOTONIC_TIMESTAMP" : "130941917",
"_TRANSPORT" : "kernel",
"_MACHINE_ID" : "3f31a0847cea4c95b7a9cec13d07deeb",
"_HOSTNAME" : "pn",
"SYSLOG_FACILITY" : "0",
"_BOOT_ID" : "e2ae14a9def345aa803a13648b95429c",
"_UDEV_DEVLINK" : [
"/dev/disk/by-uuid/22fc262a-d621-452a-a951-7761d9fcf0dc",
"/dev/disk/by-path/pci-0000:00:05.0-nvme-1","/dev/disk/by-id/nvme-nvme.8086-4445414442454546-51454d55204e564d65204374726c-00000001",
"/dev/disk/by-id/nvme-QEMU_NVMe_Ctrl_DEADBEEF"
],
"MESSAGE" : "block nvme0n1: blk_update_request: critical medium
error, dev nvme0n1, sector 10000 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0",
"_UDEV_DEVNODE" : "/dev/nvme0n1" }
Journald already knows how to utilize the dev_printk meta data.
One idea that I've suggested along the way is creating a dev_printk function that doesn't change the message text. We then avoid breaking people that are parsing. Is this something that would be acceptable to folks? It doesn't solve early boot where udev rules haven't even run, but it's better.
Thanks,
Tony
Greg Kroah-Hartman
maintainerRe: [v5 01/12] struct device: Add function callback durable_name
Quoted 12 lines · Tony Asleson
Quoted 11 lines · Greg Kroah-Hartman
Quoted 2 lines · Tony Asleson
I'm trying to figure out a way to positively identify which storage device an error belongs to over time.
"over time" is not the kernel's responsibility.
This comes up every 5 years or so. The kernel provides you, at runtime, a mapping between a hardware device and a "logical" device. It can provide information to userspace about this mapping, but once that device goes away, the kernel is free to reuse that logical device again.
Quoted 5 lines
Quoted 2 lines
If you want to track what logical devices match up to what physical device, then do it in userspace, by parsing the log files.
I don't understand why people think it's acceptable to ask user space to parse text that is subject to change.
What text is changing? The format of of the prefix of dev_*() is well known and has been stable for 15+ years now, right? What is difficult in parsing it?
Quoted 11 lines
Quoted 7 lines
Quoted 4 lines
Thank you for supplying some feedback and asking questions. I've been asking for suggestions and would very much like to have a discussion on how this issue is best solved. I'm not attached to what I've provided. I'm just trying to get towards a solution.
Again, solve this in userspace, you have the information there at runtime, why not use it?
We usually don't have the needed information if you remove the expectation that user space should parse the human readable portion of the error message.
I don't expect that userspace should have to parse any human readable portion, if they don't want to. But if you do want it to, it is pretty trivial to parse what you have today:
scsi 2:0:0:0: Direct-Access Generic STORAGE DEVICE 1531 PQ: 0 ANSI: 6
If you really have a unique identifier, then great, parse it today:
usb 4-1.3.1: Product: USB3.0 Card Reader usb 4-1.3.1: Manufacturer: Generic usb 4-1.3.1: SerialNumber: 000000001531
What's keeping that from working now?
In fact, I would argue that it does seem to work, as there are many commercial tools out there that seem to handle it just fine...
Quoted 26 lines
Quoted 7 lines
Quoted 4 lines
We've looked at user space quite a bit and there is an inherit race condition with trying to fetch the unique hardware id for a message when it gets emitted from the kernel as udev rules haven't even run (assuming we even have the meta-data to make the association).
But one moment later you do have the information, so you can properly correlate it, right?
We could have the information if all the storage paths went through
dev_printk. Here is what we get today when we encounter a read error which uses printk in the block layer:{ "_HOSTNAME" : "pn", "_TRANSPORT" : "kernel", "__MONOTONIC_TIMESTAMP" : "1806379233", "SYSLOG_IDENTIFIER" : "kernel", "_SOURCE_MONOTONIC_TIMESTAMP" : "1805611354", "SYSLOG_FACILITY" : "0", "MESSAGE" : "blk_update_request: critical medium error, dev nvme0n1, sector 10000 op 0x0:(READ) flags 0x80700 phys_seg 3 prio class 0", "PRIORITY" : "3", "_MACHINE_ID" : "3f31a0847cea4c95b7a9cec13d07deeb", "__REALTIME_TIMESTAMP" : "1601471260802301", "_BOOT_ID" : "b03ed610f21d46ab8243a495ba5a0058", "__CURSOR" :
Quoted 2 lines
"s=a063a22bbb384da0b0412e8f652deabb;i=23c2;b=b03ed610f21d46ab8243a495ba5a0058;m=6bab28e1;t=5b087959e3cfd;x=20528862f8f765c9" }
Ok, messy stuff, don't do that :)
Quoted 24 lines
Unless you parse the message text you cannot make the association. If the same message was changed to
dev_printkwe would get:{ "__REALTIME_TIMESTAMP" : "1589401901093443", "__CURSOR" : "s=caac9703b34a48fd92f7875adae55a2f;i=1c713;b=e2ae14a9def345aa803a13648b95429c;m=7d25b4f;t=5a58d77b85243;x=b034c2d3fb853870", "SYSLOG_IDENTIFIER" : "kernel", "_KERNEL_DEVICE" : "b259:917504", "__MONOTONIC_TIMESTAMP" : "131226447", "_UDEV_SYSNAME" : "nvme0n1", "PRIORITY" : "3", "_KERNEL_SUBSYSTEM" : "block", "_SOURCE_MONOTONIC_TIMESTAMP" : "130941917", "_TRANSPORT" : "kernel", "_MACHINE_ID" : "3f31a0847cea4c95b7a9cec13d07deeb", "_HOSTNAME" : "pn", "SYSLOG_FACILITY" : "0", "_BOOT_ID" : "e2ae14a9def345aa803a13648b95429c", "_UDEV_DEVLINK" : [ "/dev/disk/by-uuid/22fc262a-d621-452a-a951-7761d9fcf0dc", "/dev/disk/by-path/pci-0000:00:05.0-nvme-1",
Quoted 8 lines
"
/dev/disk/by-id/nvme-nvme.8086-4445414442454546-51454d55204e564d65204374726c-00000001",
"/dev/disk/by-id/nvme-QEMU_NVMe_Ctrl_DEADBEEF"
],
"MESSAGE" : "block nvme0n1:blk_update_request: critical medium
error, dev nvme0n1, sector 10000 op 0x0:(READ) flags0x0phys_seg1 prio class 0","_UDEV_DEVNODE" : "/dev/nvme0n1" }
Great, you have a udev sysname, a kernel subsystem and a way to associate that with a real device, what more are you wanting?
Quoted 1 line
Journald already knows how to utilize the
dev_printkmeta data.
And if you talk to the printk developers (which you seem to be keeping out of the loop here), they are ripping out the meta data facility as fast as possible. So don't rely on extending that please.
Quoted 5 lines
One idea that I've suggested along the way is creating a
dev_printkfunction that doesn't change the message text. We then avoid breaking people that are parsing. Is this something that would be acceptable to folks? It doesn't solve early boot where udev rules haven't even run, but it's better.
I still fail to understand the root problem here.
Ok, no, I think I understand what you think the problem is, I just don't see why it is up to the kernel to change what we have today when there are lots of tools out there working just fine without any kernel changes needed.
So try explaining the problem as you see it please, so we all know where to work from.
But again, cutting out the developers of the subsystems you were wanting to modify might just be making me really grumpy about this whole thing...
greg k-h
Re: [v5 01/12] struct device: Add function callback durable_name
Quoted 16 lines · Greg Kroah-Hartman
Quoted 15 lines · Tony Asleson
Quoted 13 lines · Greg Kroah-Hartman
Quoted 2 lines · Tony Asleson
I'm trying to figure out a way to positively identify which storage device an error belongs to over time.
"over time" is not the kernel's responsibility.
This comes up every 5 years or so. The kernel provides you, at runtime, a mapping between a hardware device and a "logical" device. It can provide information to userspace about this mapping, but once that device goes away, the kernel is free to reuse that logical device again.
If you want to track what logical devices match up to what physical device, then do it in userspace, by parsing the log files.
Quoted 6 lines
Quoted 2 lines
I don't understand why people think it's acceptable to ask user space to parse text that is subject to change.
What text is changing? The format of of the prefix of dev_*() is well known and has been stable for 15+ years now, right? What is difficult in parsing it?
Many of the storage layer messages are using printk, not dev_printk.
Quoted 18 lines
Quoted 11 lines
Quoted 7 lines
Quoted 4 lines
Thank you for supplying some feedback and asking questions. I've been asking for suggestions and would very much like to have a discussion on how this issue is best solved. I'm not attached to what I've provided. I'm just trying to get towards a solution.
Again, solve this in userspace, you have the information there at runtime, why not use it?
We usually don't have the needed information if you remove the expectation that user space should parse the human readable portion of the error message.
I don't expect that userspace should have to parse any human readable portion, if they don't want to. But if you do want it to, it is pretty trivial to parse what you have today:
scsi 2:0:0:0: Direct-Access Generic STORAGE DEVICE 1531 PQ: 0 ANSI: 6
Quoted 7 lines
If you really have a unique identifier, then great, parse it today:
usb 4-1.3.1: Product: USB3.0 Card Reader usb 4-1.3.1: Manufacturer: Generic usb 4-1.3.1: SerialNumber: 000000001531What's keeping that from working now?
I believe these examples are using dev_printk. With dev_printk we don't need to parse the text, we can use the meta data.
Quoted 2 lines
In fact, I would argue that it does seem to work, as there are many commercial tools out there that seem to handle it just fine...
I'm trying to get something that's works for journalctl.
Quoted 64 lines
Quoted 26 lines
Quoted 7 lines
Quoted 4 lines
We've looked at user space quite a bit and there is an inherit race condition with trying to fetch the unique hardware id for a message when it gets emitted from the kernel as udev rules haven't even run (assuming we even have the meta-data to make the association).
But one moment later you do have the information, so you can properly correlate it, right?
We could have the information if all the storage paths went through
dev_printk. Here is what we get today when we encounter a read error which uses printk in the block layer:{ "_HOSTNAME" : "pn", "_TRANSPORT" : "kernel", "__MONOTONIC_TIMESTAMP" : "1806379233", "SYSLOG_IDENTIFIER" : "kernel", "_SOURCE_MONOTONIC_TIMESTAMP" : "1805611354", "SYSLOG_FACILITY" : "0", "MESSAGE" : "blk_update_request: critical medium error, dev nvme0n1, sector 10000 op 0x0:(READ) flags 0x80700 phys_seg 3 prio class 0", "PRIORITY" : "3", "_MACHINE_ID" : "3f31a0847cea4c95b7a9cec13d07deeb", "__REALTIME_TIMESTAMP" : "1601471260802301", "_BOOT_ID" : "b03ed610f21d46ab8243a495ba5a0058", "__CURSOR" :Quoted 2 lines
"s=a063a22bbb384da0b0412e8f652deabb;i=23c2;b=b03ed610f21d46ab8243a495ba5a0058;m=6bab28e1;t=5b087959e3cfd;x=20528862f8f765c9" }
Ok, messy stuff, don't do that :)
Quoted 24 lines
Unless you parse the message text you cannot make the association. If the same message was changed to
dev_printkwe would get:{ "__REALTIME_TIMESTAMP" : "1589401901093443", "__CURSOR" : "s=caac9703b34a48fd92f7875adae55a2f;i=1c713;b=e2ae14a9def345aa803a13648b95429c;m=7d25b4f;t=5a58d77b85243;x=b034c2d3fb853870", "SYSLOG_IDENTIFIER" : "kernel", "_KERNEL_DEVICE" : "b259:917504", "__MONOTONIC_TIMESTAMP" : "131226447", "_UDEV_SYSNAME" : "nvme0n1", "PRIORITY" : "3", "_KERNEL_SUBSYSTEM" : "block", "_SOURCE_MONOTONIC_TIMESTAMP" : "130941917", "_TRANSPORT" : "kernel", "_MACHINE_ID" : "3f31a0847cea4c95b7a9cec13d07deeb", "_HOSTNAME" : "pn", "SYSLOG_FACILITY" : "0", "_BOOT_ID" : "e2ae14a9def345aa803a13648b95429c", "_UDEV_DEVLINK" : [ "/dev/disk/by-uuid/22fc262a-d621-452a-a951-7761d9fcf0dc", "/dev/disk/by-path/pci-0000:00:05.0-nvme-1",Quoted 8 lines
"
/dev/disk/by-id/nvme-nvme.8086-4445414442454546-51454d55204e564d65204374726c-00000001",
"/dev/disk/by-id/nvme-QEMU_NVMe_Ctrl_DEADBEEF"
],
"MESSAGE" : "block nvme0n1:blk_update_request: critical medium
error, dev nvme0n1, sector 10000 op 0x0:(READ) flags0x0phys_seg1 prio class 0","_UDEV_DEVNODE" : "/dev/nvme0n1" }
Quoted 2 lines
Great, you have a udev sysname, a kernel subsystem and a way to associate that with a real device, what more are you wanting?
Did you miss in my example where it's currently a printk? I showed what it would look like if it was a dev_printk.
Journald is using _KERNEL_DEVICE to add the _UDEV_DEVLINK information to the journal entry, it's not parsing the prefix of the message.
The above json is outputted from journalctl when you specify "-o json-pretty".
Quoted 5 lines
Quoted 1 line
Journald already knows how to utilize the
dev_printkmeta data.And if you talk to the printk developers (which you seem to be keeping out of the loop here), they are ripping out the meta data facility as fast as possible. So don't rely on extending that please.
Again, I'm not trying to keep anyone out of the loop. Last I knew the meta data capability wasn't being removed, maybe this has changed?
Ref.
https://lore.kernel.org/lkml/[email protected]/
Quoted 7 lines
Quoted 5 lines
One idea that I've suggested along the way is creating a
dev_printkfunction that doesn't change the message text. We then avoid breaking people that are parsing. Is this something that would be acceptable to folks? It doesn't solve early boot where udev rules haven't even run, but it's better.I still fail to understand the root problem here.
IMHO one of the root problems is that many storage messages are still using printk. Changing messages to dev_printk has been met with resistance.
Quoted 7 lines
Ok, no, I think I understand what you think the problem is, I just don't see why it is up to the kernel to change what we have today when there are lots of tools out there working just fine without any kernel changes needed.
So try explaining the problem as you see it please, so we all know where to work from.
To me the problem today is the kernel logs information to identify how a storage device is attached, not what is attached. I think you agree with this statement. The log information is not helpful without the information to correlate to the actual device. I think you also agree with this too as you have mentioned it's user space's responsibility to collect this so that the correlation can be done.
If the following are *both* true, we have a usable message that has the correlated data with it in the journal.
1. The storage related kernel message goes through dev_printk
2. At the time of the message being emitted the device symlinks are present.
When those two things are both true, journalctl can do the following (today):
$ journalctl /dev/disk/by-id/wwn-0x5002538844584d30
However, it usually can't because the above two things are rarely both true at the same time for a given message for journald when it logs to the journal.
You keep saying this is a user space issue, but I believe we still need a bit of help from the kernel at the very least by migrating to dev_printk or something similar that adds the same meta data without changing the message text.
Yes, my patch series went one step further and added the device ID as structured data to the log message, but I was also trying to minimize the race condition between the kernel emitting a message and journald not having the information to associate it to the hardware device.
If people have other suggestions please let them be known.
Quoted 3 lines
But again, cutting out the developers of the subsystems you were wanting to modify might just be making me really grumpy about this whole thing...
Again, I'm sorry I didn't reach out to the correct people. Hopefully I've CC'd everyone that is appropriate for this discussion.
Thanks,
Tony
Greg Kroah-Hartman
maintainerRe: [v5 01/12] struct device: Add function callback durable_name
Quoted 17 lines · Tony Asleson
Quoted 16 lines · Greg Kroah-Hartman
Quoted 15 lines · Tony Asleson
Quoted 13 lines · Greg Kroah-Hartman
Quoted 2 lines · Tony Asleson
I'm trying to figure out a way to positively identify which storage device an error belongs to over time.
"over time" is not the kernel's responsibility.
This comes up every 5 years or so. The kernel provides you, at runtime, a mapping between a hardware device and a "logical" device. It can provide information to userspace about this mapping, but once that device goes away, the kernel is free to reuse that logical device again.
If you want to track what logical devices match up to what physical device, then do it in userspace, by parsing the log files.
Quoted 8 lines
Quoted 6 lines
Quoted 2 lines
I don't understand why people think it's acceptable to ask user space to parse text that is subject to change.
What text is changing? The format of of the prefix of dev_*() is well known and has been stable for 15+ years now, right? What is difficult in parsing it?
Many of the storage layer messages are using printk, not
dev_printk.
Ok, then stop right there. Fix that up. Don't try to route around the standard way of displaying log messages by creating a totally different way of doing things.
Just use the dev_*() calls, and all will be fine. Kernel log messages are not "ABI" in that they have to be preserved in any specific way, so adding a prefix to them as dev_*() does, will be fine.
thanks,
greg k-h
Re: [v5 01/12] struct device: Add function callback durable_name
Greg,
Quoted 9 lines
Quoted 5 lines
Quoted 3 lines
What text is changing? The format of of the prefix of dev_*() is well known and has been stable for 15+ years now, right? What is difficult in parsing it?
Many of the storage layer messages are using printk, not
dev_printk.Ok, then stop right there. Fix that up. Don't try to route around the standard way of displaying log messages by creating a totally different way of doing things.
Couldn't agree more!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Re: [v5 01/12] struct device: Add function callback durable_name
In reply to Tony AslesonQuoted 17 lines · Tony Asleson
Quoted 16 lines · Greg Kroah-Hartman
Quoted 15 lines · Tony Asleson
Quoted 13 lines · Greg Kroah-Hartman
Quoted 2 lines · Tony Asleson
I'm trying to figure out a way to positively identify which storage device an error belongs to over time.
"over time" is not the kernel's responsibility.
This comes up every 5 years or so. The kernel provides you, at runtime, a mapping between a hardware device and a "logical" device. It can provide information to userspace about this mapping, but once that device goes away, the kernel is free to reuse that logical device again.
If you want to track what logical devices match up to what physical device, then do it in userspace, by parsing the log files.
Quoted 9 lines
Quoted 6 lines
Quoted 2 lines
I don't understand why people think it's acceptable to ask user space to parse text that is subject to change.
What text is changing? The format of of the prefix of dev_*() is well known and has been stable for 15+ years now, right? What is difficult in parsing it?
Many of the storage layer messages are using printk, not
dev_printk.
So that would be the immediate angle of attack ...
Quoted 24 lines
Quoted 24 lines
Quoted 11 lines
Quoted 7 lines
Quoted 4 lines
Thank you for supplying some feedback and asking questions. I've been asking for suggestions and would very much like to have a discussion on how this issue is best solved. I'm not attached to what I've provided. I'm just trying to get towards a solution.
Again, solve this in userspace, you have the information there at runtime, why not use it?
We usually don't have the needed information if you remove the expectation that user space should parse the human readable portion of the error message.
I don't expect that userspace should have to parse any human readable portion, if they don't want to. But if you do want it to, it is pretty trivial to parse what you have today:
scsi 2:0:0:0: Direct-Access Generic STORAGE DEVICE 1531 PQ: 0 ANSI: 6
If you really have a unique identifier, then great, parse it today:
usb 4-1.3.1: Product: USB3.0 Card Reader usb 4-1.3.1: Manufacturer: Generic usb 4-1.3.1: SerialNumber: 000000001531
Quoted 5 lines
Quoted 1 line
What's keeping that from working now?
I believe these examples are using
dev_printk. Withdev_printkwe don't need to parse the text, we can use the meta data.
So it looks as most of your usecase would be solved by moving to
dev_printk().
Why not work on that instead?
I do presume this will have immediate benefits for everybody, and will have approval from everyone.
Cheers,
Hannes
Signature · 4 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), Geschäftsführer: Felix Imendörffer
Re: [v5 01/12] struct device: Add function callback durable_name
In reply to Tony AslesonQuoted 2 lines · Tony Asleson
The log information is not helpful without the information to correlate to the actual device.
Log messages that associate one entity with another can be generated whenever such an association comes into existence, which is probably when devices get probed.
E.g. a host:channel:target:lun identifier gets associated with a block device name by the dev_printk() calls in sd_probe():
[ 3.600000] sd 0:0:0:0: [sda] Attached SCSI disk
BTW, if you think of {"0:0:0:0","sda"} as a row in some normalized table and squint a bit, this problem is not unlike the replication of database tables over a message queue.
PATCH v5 2/12 create_syslog_header: Add durable name
In reply to Tony AslesonThis gets us a persistent durable name for code that logs messages in the block layer that have the appropriate callbacks setup for durable name.
--- drivers/base/core.c | 5 +++++ 1 file changed, 5 insertions(+)
drivers/base/core.c +5 -0
@@ -3864,6 +3864,7 @@ create_syslog_header(const struct device *dev, char *hdr, size_t hdrlen) {const char *subsys;size_t pos =0 ; +int dlen;if (dev->class) subsys = dev->class->name;
@@ -3906,6 +3907,10 @@ create_syslog_header(const struct device *dev, char *hdr, size_t hdrlen) "DEVICE=+%s:%s", subsys, dev_name(dev)); } + dlen =dev_durable_name (dev, hdr + (pos +1 ), hdrlen - (pos +1 )); +if (dlen) + pos += dlen +1 ; +if (pos >= hdrlen) goto overflow;
Signature · 1 line
2.26.2
PATCH v5 3/12 dev_vprintk_emit: Increase hdr size
In reply to Tony AslesonWith the addition of the device persistent id we have the possibility of adding 154 more bytes to the hdr. Thus if we assume the previous size of 128 was sufficient we can simply add the 2 amounts and round up.
--- drivers/base/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/base/core.c +1 -1
@@ -3924,7 +3924,7 @@ create_syslog_header(const struct device *dev, char *hdr, size_t hdrlen)int dev_vprintk_emit (int level,const struct device *dev,const char *fmt,va_list args) { -char hdr[128 ]; +char hdr[288 ];size_t hdrlen; hdrlen =create_syslog_header (dev, hdr,sizeof (hdr));
Signature · 1 line
2.26.2
PATCH v5 4/12 scsi: Add durable_name for dev_printk
In reply to Tony AslesonAdd the needed functions to fill out the durable_name function call back for scsi based storage devices. This allows calls into dev_printk for scsi devices to have a persistent id associated with them.
--- drivers/scsi/scsi_lib.c | 9 +++++++++ drivers/scsi/scsi_sysfs.c | 35 ++++++++++++++++++++++++++++------- drivers/scsi/sd.c | 2 ++ include/scsi/scsi_device.h | 3 +++ 4 files changed, 42 insertions(+), 7 deletions(-)
drivers/scsi/scsi_lib.c +9 -0
@@ -3151,3 +3151,12 @@ int scsi_vpd_tpg_id(struct scsi_device *sdev, int *rel_id)return group_id; }EXPORT_SYMBOL (scsi_vpd_tpg_id); + +int scsi_durable_name (struct scsi_device *sdev,char *buf,size_t len) +{ +int vpd_len =scsi_vpd_lun_id (sdev, buf, len); +if (vpd_len >0 && vpd_len < len) +return vpd_len +1 ; +return 0 ; +} +EXPORT_SYMBOL (scsi_durable_name);
drivers/scsi/scsi_sysfs.c +28 -7
@@ -504,15 +504,19 @@ static struct class sdev_class = { .dev_release = scsi_device_cls_release, }; +static struct scsi_device *dev_to_scsi_device (const struct device *dev) +{ +return dev->type == &scsi_dev_type ?to_scsi_device (dev) :NULL ; +} +/* all probing is done in the individual ->probe routines */ static int scsi_bus_match (struct device *dev,struct device_driver *gendrv) { -struct scsi_device *sdp; +struct scsi_device *sdp =dev_to_scsi_device (dev); -if (dev->type != &scsi_dev_type) +if (!sdp)return 0 ; - sdp =to_scsi_device (dev);if (sdp->no_uld_attach)return 0 ;return (sdp->inq_periph_qual ==SCSI_INQ_PQ_CON )?1 :0 ;
@@ -520,13 +524,11 @@ static int scsi_bus_match(struct device *dev, struct device_driver *gendrv)static int scsi_bus_uevent (struct device *dev,struct kobj_uevent_env *env) { -struct scsi_device *sdev; +struct scsi_device *sdev =dev_to_scsi_device (dev); -if (dev->type != &scsi_dev_type) +if (!sdev)return 0 ; - sdev =to_scsi_device (dev); -add_uevent_var (env,"MODALIAS=" SCSI_DEVICE_MODALIAS_FMT , sdev->type);return 0 ; }
@@ -1582,6 +1584,24 @@ static struct device_type scsi_dev_type = { .groups = scsi_sdev_attr_groups, }; +int dev_to_scsi_durable_name (const struct device *dev,char *buf,size_t len) +{ +/* +* When we go through dev_printk in the scsi layer, dev is embedded +* in a struct scsi_device. When we go through the block layer, +* dev is embedded in struct genhd, thus we need different paths to +* retrieve the struct scsi_device to call scsi_durable_name. +*/ +struct scsi_device *sdev =dev_to_scsi_device (dev); +if (!sdev) + sdev =dev_to_scsi_device (dev->parent); +if (!sdev) +return 0 ; + +return scsi_durable_name (sdev, buf, len); +} +EXPORT_SYMBOL (dev_to_scsi_durable_name); +void scsi_sysfs_device_initialize (struct scsi_device *sdev) {unsigned long flags;
@@ -1591,6 +1611,7 @@ void scsi_sysfs_device_initialize(struct scsi_device *sdev)device_initialize (&sdev->sdev_gendev); sdev->sdev_gendev.bus = &scsi_bus_type; sdev->sdev_gendev.type = &scsi_dev_type; + sdev->sdev_gendev.durable_name = dev_to_scsi_durable_name;dev_set_name (&sdev->sdev_gendev,"%d:%d:%d:%llu" , sdev->host->host_no, sdev->channel, sdev->id, sdev->lun);
drivers/scsi/sd.c +2 -0
@@ -3386,6 +3386,8 @@ static int sd_probe(struct device *dev) gd->private_data = &sdkp->driver; gd->queue = sdkp->device->request_queue; +disk_to_dev (gd)->durable_name = dev_to_scsi_durable_name; +/* defaults, until the device tells us otherwise */ sdp->sector_size =512 ; sdkp->capacity =0 ;
include/scsi/scsi_device.h +3 -0
@@ -464,6 +464,9 @@ extern void sdev_disable_disk_events(struct scsi_device *sdev);extern void sdev_enable_disk_events (struct scsi_device *sdev);extern int scsi_vpd_lun_id (struct scsi_device *,char *,size_t );extern int scsi_vpd_tpg_id (struct scsi_device *,int *); +extern int dev_to_scsi_durable_name (const struct device *dev,char *buf, +size_t len); +extern int scsi_durable_name (struct scsi_device *sdev,char *buf,size_t len);#ifdef CONFIG_PM extern int scsi_autopm_get_device (struct scsi_device *);
Signature · 1 line
2.26.2
PATCH v5 5/12 nvme: Add durable name for dev_printk
In reply to Tony AslesonChanged the comment from // to /* and re-worked buffer space needed for formatting wwid as requested by Keith Busch.
ref.
https://lore.kernel.org/linux-block/[email protected]/
--- drivers/nvme/host/core.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)
drivers/nvme/host/core.c +18 -0
@@ -2734,6 +2734,22 @@ static bool nvme_validate_cntlid(struct nvme_subsystem *subsys,return true; } +static ssize_t wwid_show (struct device *dev,struct device_attribute *attr, +char *buf); + +static int dev_to_nvme_durable_name (const struct device *dev,char *buf,size_t len) +{ +/* +* Max 141 needed for wwid_show, make sure we have the space available +* in our buffer before we format the wwid directly into it. +*/ +if (len >=141 ) { +ssize_t wwid_len =wwid_show ((struct device *)dev,NULL , buf); +return wwid_len >0 ? wwid_len -1 :0 ;/* remove '\n' */ + } +return 0 ; +} +static int nvme_init_subsystem (struct nvme_ctrl *ctrl,struct nvme_id_ctrl *id) {struct nvme_subsystem *subsys, *found;
@@ -3663,6 +3679,8 @@ static void nvme_alloc_ns(struct nvme_ctrl *ctrl, unsigned nsid) disk->queue = ns->queue; disk->flags = flags;memcpy (disk->disk_name, disk_name,DISK_NAME_LEN ); +disk_to_dev (disk)->durable_name = dev_to_nvme_durable_name; + ns->disk = disk;if (__nvme_revalidate_disk (disk, id))
Signature · 1 line
2.26.2
PATCH v5 6/12 libata: Add ata_scsi_durable_name
In reply to Tony AslesonFunction used to create the durable name for ata scsi.
--- drivers/ata/libata-scsi.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-)
drivers/ata/libata-scsi.c +15 -3
@@ -1086,6 +1086,13 @@ int ata_scsi_dev_config(struct scsi_device *sdev, struct ata_device *dev)return 0 ; } +int ata_scsi_durable_name (const struct device *dev,char *buf,size_t len) +{ + struct ata_device *ata_dev = container_of(dev, struct ata_device, tdev); + +return scsi_durable_name (ata_dev->sdev, buf, len); +} +/** * ata_scsi_slave_config - Set SCSI device attributes * @sdev: SCSI device to examine
@@ -1102,14 +1109,19 @@ int ata_scsi_slave_config(struct scsi_device *sdev) {struct ata_port *ap =ata_shost_to_port (sdev->host);struct ata_device *dev =__ata_scsi_find_dev (ap, sdev); -int rc =0 ; +int rc;ata_scsi_sdev_config (sdev); -if (dev) +if (dev) { rc =ata_scsi_dev_config (sdev, dev); +if (rc) +return rc; -return rc; + dev->tdev.durable_name = ata_scsi_durable_name; + } + +return 0 ; }EXPORT_SYMBOL_GPL (ata_scsi_slave_config);
Signature · 1 line
2.26.2
PATCH v5 7/12 libata: Make ata_scsi_durable_name static
In reply to Tony Asleson--- drivers/ata/libata-scsi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/ata/libata-scsi.c +1 -1
@@ -1086,7 +1086,7 @@ int ata_scsi_dev_config(struct scsi_device *sdev, struct ata_device *dev)return 0 ; } -int ata_scsi_durable_name (const struct device *dev,char *buf,size_t len) +static int ata_scsi_durable_name (const struct device *dev,char *buf,size_t len) { struct ata_device *ata_dev = container_of(dev, struct ata_device, tdev);
Signature · 1 line
2.26.2
Re: [v5 07/12] libata: Make ata_scsi_durable_name static
Hello!
Quoted 11 lines · Tony Asleson
Signed-off-by Tony Asleson <[email protected]>Signed-off-by kernel test robot <[email protected]>--- drivers/ata/libata-scsi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/ata/libata-scsi.c +0
-0
drivers/ata/libata-scsi.c int ata_scsi_dev_config(struct scsi_device *sdev, struct ata_device *dev) · line 1089
Quoted 5 lines
drivers/ata/libata-scsi.c +1
-1@@ -1086,7 +1086,7 @@ int ata_scsi_dev_config(struct scsi_device *sdev, struct ata_device *dev)return 0 ; } -int ata_scsi_durable_name (const struct device *dev,char *buf,size_t len) +static int ata_scsi_durable_name (const struct device *dev,char *buf,size_t len)
Why not do it in patch #6 -- when introducing the function?
[...]
MBR, Sergei
Re: [v5 07/12] libata: Make ata_scsi_durable_name static
Quoted 10 lines · Sergei Shtylyov
Hello!
Quoted 6 lines · Tony Asleson
Signed-off-by Tony Asleson <[email protected]>Signed-off-by kernel test robot <[email protected]>---
drivers/ata/libata-scsi.c| 2 +-1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 14 lines
drivers/ata/libata-scsi.c int ata_scsi_dev_config(struct scsi_device · line 1092
Quoted 12 lines
drivers/ata/libata-scsi.c +1
-0@@ -1086,7 +1086,7 @@ int ata_scsi_dev_config(struct scsi_device *sdev, struct ata_device *dev) return 0; } -int ata_scsi_durable_name(const struct device *dev, char *buf, size_t len) +static int ata_scsi_durable_name(const struct device *dev, char *buf,size_t len)Why not do it in patch #6 -- when introducing the function?
This issue was found by the intel kernel test robot in v4 patch series. I thought it was better to have a separate commit with the correction that matched it's signed off. Maybe that's not the correct approach?
James Bottomley
maintainerRe: [v5 07/12] libata: Make ata_scsi_durable_name static
Quoted 28 lines · Tony Asleson
Quoted 10 lines · Sergei Shtylyov
Hello!
Quoted 6 lines · Tony Asleson
Signed-off-by Tony Asleson <[email protected]>Signed-off-by kernel test robot <[email protected]>--- drivers/ata/libata-scsi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)Quoted 16 lines
Quoted 13 lines
drivers/ata/libata- renamed from drivers/ata/libata-scsi.c +0
-0
scsi.c
index194dac7dbdca..13a58ed7184c100644drivers/ata/libata-scsi.c +1
-0@@ -1086,7 +1086,7 @@ int ata_scsi_dev_config(struct scsi_device *sdev, struct ata_device *dev)return 0 ; } -int ata_scsi_durable_name(const struct device *dev, char *buf, size_t len) +static int ata_scsi_durable_name(const struct device *dev, char *buf,Quoted 1 line
size_tlen)Why not do it in patch #6 -- when introducing the function?
Quoted 4 lines
This issue was found by the intel kernel test robot in v4 patch series. I thought it was better to have a separate commit with the correction that matched it's signed off. Maybe that's not the correct approach?
No ... while a patch is being reviewed the purpose of review is to make it better by folding in all the comments. It then gets a changelog put below the
---
v5: made X function static
So people can follow how it has evolved. This is all actually described in Documentation/submitting-patches.
James
Re: [v5 07/12] libata: Make ata_scsi_durable_name static
In reply to Sergei ShtylyovQuoted 2 lines · Sergei Shtylyov
Why not do it in patch #6 -- when introducing the function?
Re-working patch series, will do as you suggest and as outlined in the submitting-patches.rst referenced by James.
Thanks
PATCH v5 8/12 Add durable_name_printk
In reply to Tony AslesonIdeally block related code would standardize on using dev_printk, but dev_printk does change the user visible messages which is questionable. Adding this function which adds the structured key/value durable name to the log entry. It has the same signature as dev_printk. In the future, code that is using this could easily transition to dev_printk when that becomes workable.
--- drivers/base/core.c | 15 +++++++++++++++ include/linux/dev_printk.h | 5 +++++ 2 files changed, 20 insertions(+)
drivers/base/core.c +15 -0
@@ -3975,6 +3975,21 @@ void dev_printk(const char *level, const struct device *dev, }EXPORT_SYMBOL (dev_printk); +void durable_name_printk (const char *level,const struct device *dev, +const char *fmt, ...) +{ +size_t dictlen; +va_list args; +char dict[288 ]; + + dictlen =dev_durable_name (dev, dict,sizeof (dict)); + +va_start (args, fmt); +vprintk_emit (0 , level[1 ] -'0' , dict, dictlen, fmt, args); +va_end (args); +} +EXPORT_SYMBOL (durable_name_printk); +#define define_dev_printk_level (func, kern_level) \ void func(const struct device *dev, const char *fmt, ...) \ { \
include/linux/dev_printk.h +5 -0
@@ -32,6 +32,11 @@ int dev_printk_emit(int level, const struct device *dev, const char *fmt, ...); __printf(3, 4) __coldvoid dev_printk (const char *level,const struct device *dev,const char *fmt, ...); + +__printf(3, 4) __cold +void durable_name_printk (const char *level,const struct device *dev, +const char *fmt, ...); + __printf(2, 3) __coldvoid _dev_emerg (const struct device *dev,const char *fmt, ...);__printf (2 ,3 )__cold
Signature · 1 line
2.26.2
Re: [v5 08/12] Add durable_name_printk
Quoted 8 lines · Tony Asleson
Ideally block related code would standardize on using
dev_printk, butdev_printkdoes change the user visible messages which is questionable. Adding this function which adds the structured key/value durable name to the log entry. It has the same signature asdev_printk. In the future, code that is using this could easily transition todev_printkwhen that becomes workable.
Quoted 5 lines
Signed-off-by Tony Asleson <[email protected]>--- drivers/base/core.c | 15 +++++++++++++++ include/linux/dev_printk.h | 5 +++++ 2 files changed, 20 insertions(+)
Hi,
I suggest that these 2 new function names should be
printk_durable_name()
and
printk_durable_name_ratelimited()
Those names would be closer to the printk* family of function names. Of course, you can find exceptions to this, like dev_printk(), but that is in the dev_*() family of function names.
Quoted 33 lines
drivers/base/core.c +15
-0@@ -3975,6 +3975,21 @@ void dev_printk(const char *level, const struct device *dev, }EXPORT_SYMBOL (dev_printk); +void durable_name_printk (const char *level,const struct device *dev, +const char *fmt, ...) +{ +size_t dictlen; +va_list args; +char dict[288 ]; + + dictlen =dev_durable_name (dev, dict,sizeof (dict)); + +va_start (args, fmt); +vprintk_emit (0 , level[1 ] -'0' , dict, dictlen, fmt, args); +va_end (args); +} +EXPORT_SYMBOL (durable_name_printk); +#define define_dev_printk_level (func, kern_level) \ void func(const struct device *dev, const char *fmt, ...) \ { \include/linux/dev_printk.h +0
-0@@ -32,6 +32,11 @@ int dev_printk_emit(int level, const struct device *dev, const char *fmt, ...);__printf (3 ,4 )__cold
include/linux/dev_printk.h int dev_printk_emit(int level, const struct device *dev, const char *fmt, ...); · line 42
Quoted 10 lines
include/linux/dev_printk.h +5
-0void dev_printk (const char *level,const struct device *dev,const char *fmt, ...); + +__printf(3, 4) __cold +void durable_name_printk (const char *level,const struct device *dev, +const char *fmt, ...); + __printf(2, 3) __coldvoid _dev_emerg (const struct device *dev,const char *fmt, ...);__printf (2 ,3 )__cold
Thanks.
Signature · 1 line
~Randy
Re: [v5 08/12] Add durable_name_printk
Quoted 9 lines · Randy Dunlap
I suggest that these 2 new function names should be
printk_durable_name()
and
printk_durable_name_ratelimited()Those names would be closer to the printk* family of function names. Of course, you can find exceptions to this, like
dev_printk(), but that is in the dev_*() family of function names.
durable_name_printk has the same argument signature as dev_printk with it's intention that in the future it might be a candidate to get changed to dev_printk. The reason I'm not using dev_printk is to avoid changing the content of the message users see.
With this clarification, do you still suggest the rename or maybe suggest something different?
dev_id_printkid_printk
...
I'm also thinking that maybe we should add a new function do everything dev_printk does, but without prepending the device driver name and device name to the message. So we can get the metadata adds without having the content of the message change.
Thanks
Re: [v5 08/12] Add durable_name_printk
Quoted 11 lines · Tony Asleson
Quoted 9 lines · Randy Dunlap
I suggest that these 2 new function names should be
printk_durable_name()
and
printk_durable_name_ratelimited()Those names would be closer to the printk* family of function names. Of course, you can find exceptions to this, like
dev_printk(), but that is in the dev_*() family of function names.
Quoted 7 lines
durable_name_printkhas the same argument signature asdev_printkwith it's intention that in the future it might be a candidate to get changed todev_printk. The reason I'm not usingdev_printkis to avoid changing the content of the message users see.With this clarification, do you still suggest the rename or maybe suggest something different?
Since you seem to bring it up, "durable_name" is a bit long IMO.
But yes, I still prefer printk_durable_name() etc. The other order seems backwards to me. But that's still just an opinion.
Quoted 8 lines
dev_id_printkid_printk
...I'm also thinking that maybe we should add a new function do everything
dev_printkdoes, but without prepending the device driver name and device name to the message. So we can get the metadata adds without having the content of the message change.
thanks.
Signature · 2 lines
~Randy Reported-by: Randy Dunlap <[email protected]>
PATCH v5 9/12 libata: use durable_name_printk
In reply to Tony AslesonUtilize durable_name_printk to associate the durable name with the log message via structured data. The user visible portion of the log message is unchanged.
--- drivers/ata/libata-core.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-)
drivers/ata/libata-core.c +9 -8
@@ -6443,7 +6443,8 @@ void ata_port_printk(const struct ata_port *ap, const char *level, vaf.fmt = fmt; vaf.va = &args; -printk ("%sata%u: %pV" , level, ap->print_id, &vaf); +durable_name_printk (level, &ap->tdev,"ata%u: %pV" , + ap->print_id, &vaf);va_end (args); }
@@ -6461,11 +6462,11 @@ void ata_link_printk(const struct ata_link *link, const char *level, vaf.va = &args;if (sata_pmp_attached (link->ap) || link->ap->slave_link) -printk ("%sata%u.%02u: %pV" , - level, link->ap->print_id, link->pmp, &vaf); +durable_name_printk (level, &link->tdev,"ata%u.%02u: %pV" , + link->ap->print_id, link->pmp, &vaf);else -printk ("%sata%u: %pV" , - level, link->ap->print_id, &vaf); +durable_name_printk (level, &link->tdev,"ata%u: %pV" , + link->ap->print_id, &vaf);va_end (args); }
@@ -6482,9 +6483,9 @@ void ata_dev_printk(const struct ata_device *dev, const char *level, vaf.fmt = fmt; vaf.va = &args; -printk ("%sata%u.%02u: %pV" , - level, dev->link->ap->print_id, dev->link->pmp + dev->devno, - &vaf); +durable_name_printk (level, &dev->tdev,"ata%u.%02u: %pV" , + dev->link->ap->print_id, dev->link->pmp + dev->devno, + &vaf);va_end (args); }
Signature · 1 line
2.26.2
PATCH v5 10/12 Add durable_name_printk_ratelimited
In reply to Tony AslesonCreate a rate limited macro for durable_name_printk so that we can use this for printk_ratelimited usage in the block layers and add the durable name key/value to the log message.
--- include/linux/dev_printk.h | 9 +++++++++ 1 file changed, 9 insertions(+)
include/linux/dev_printk.h +9 -0
@@ -37,6 +37,15 @@ __printf(3, 4) __coldvoid durable_name_printk (const char *level,const struct device *dev,const char *fmt, ...); +#define durable_name_printk_ratelimited (level, dev, fmt, ...) \ +do { \ + static DEFINE_RATELIMIT_STATE(_rs, \ + DEFAULT_RATELIMIT_INTERVAL, \ + DEFAULT_RATELIMIT_BURST); \ + if (__ratelimit(&_rs)) \ + durable_name_printk(level, dev, fmt, ##__VA_ARGS__); \ +} while (0) + __printf(2, 3) __coldvoid _dev_emerg (const struct device *dev,const char *fmt, ...);__printf (2 ,3 )__cold
Signature · 1 line
2.26.2
PATCH v5 11/12 print_req_error: Use durable_name_printk_ratelimited
In reply to Tony AslesonReplace printk_ratelimited with one that adds the key/value durable name to log entry.
--- block/blk-core.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
block/blk-core.c +4 -1
@@ -218,12 +218,15 @@ EXPORT_SYMBOL_GPL(blk_status_to_errno);static void print_req_error (struct request *req,blk_status_t status,const char *caller) { +struct device *dev;int idx = (__force int )status;if (WARN_ON_ONCE (idx >=ARRAY_SIZE (blk_errors)))return ; - printk_ratelimited(KERN_ERR + dev = req->rq_disk ?disk_to_dev (req->rq_disk) :NULL ; + + durable_name_printk_ratelimited(KERN_ERR, dev, "%s: %s error, dev %s, sector %llu op 0x%x:(%s) flags 0x%x " "phys_seg %u prio class %u\n", caller, blk_errors[idx].name,
Signature · 1 line
2.26.2
PATCH v5 12/12 buffer_io_error: Use durable_name_printk_ratelimited
In reply to Tony AslesonReplace printk_ratelimited with one that adds the key/value durable name to log entry.
--- fs/buffer.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-)
fs/buffer.c +11 -4
@@ -125,10 +125,17 @@ EXPORT_SYMBOL(__wait_on_buffer);static void buffer_io_error (struct buffer_head *bh,char *msg) { -if (!test_bit (BH_Quiet, &bh->b_state)) -printk_ratelimited (KERN_ERR -"Buffer I/O error on dev %pg, logical block %llu%s\n" , - bh->b_bdev, (unsigned long long )bh->b_blocknr, msg); +struct device *gendev; + +if (test_bit (BH_Quiet, &bh->b_state)) +return ; + + gendev = bh->b_bdev->bd_disk ? +disk_to_dev (bh->b_bdev->bd_disk) :NULL ; + +durable_name_printk_ratelimited (KERN_ERR , gendev, +"Buffer I/O error on dev %pg, logical block %llu%s\n" , + bh->b_bdev, (unsigned long long )bh->b_blocknr, msg); }/*
Signature · 1 line
2.26.2