FAILED: patch "[PATCH] s390/dasd: Use correct lock while counting channel queue" failed to apply to 5.4-stable tree
Subsystems (3)
[email protected]
maintainerThe patch below does not apply to the 5.4-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.
To reproduce the conflict and resubmit, you may use the following commands:
git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.4.y
git checkout FETCH_HEAD
git cherry-pick -x ccc45cb4e7271c74dbb27776ae8f73d84557f5c6
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2023061111-tracing-shakiness-9054@gregkh' --subject-prefix 'PATCH 5.4.y' HEAD^..
Possible dependencies:
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
From ccc45cb4e7271c74dbb27776ae8f73d84557f5c6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jan=20H=C3=B6ppner?= <[email protected]>
Date: Fri, 9 Jun 2023 17:37:50 +0200
Subject: [PATCH] s390/dasd: Use correct lock while counting channel queue
length
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The lock around counting the channel queue length in the BIODASDINFO ioctl was incorrectly changed to the dasd_block->queue_lock with commit 583d6535cb9d ("dasd: remove dead code"). This can lead to endless list iterations and a subsequent crash.
The queue_lock is supposed to be used only for queue lists belonging to dasd_block. For dasd_device related queue lists the ccwdev lock must be used.
Fix the mentioned issues by correctly using the ccwdev lock instead of the queue lock.
drivers/s390/block/dasd_ioctl.c +2 -2
@@ -552,10 +552,10 @@ static int __dasd_ioctl_information(struct dasd_block *block,memcpy (dasd_info->type, base->discipline->name,4 ); -spin_lock_irqsave (&block->queue_lock, flags); +spin_lock_irqsave (get_ccwdev_lock (base->cdev), flags);list_for_each (l, &base->ccw_queue) dasd_info->chanq_len++; -spin_unlock_irqrestore (&block->queue_lock, flags); +spin_unlock_irqrestore (get_ccwdev_lock (base->cdev), flags);return 0 ; }
Jan Höppner
maintainerPATCH 1/1 s390/dasd: Use correct lock while counting channel queue length
The lock around counting the channel queue length in the BIODASDINFO ioctl was incorrectly changed to the dasd_block->queue_lock with commit 583d6535cb9d ("dasd: remove dead code"). This can lead to endless list iterations and a subsequent crash.
The queue_lock is supposed to be used only for queue lists belonging to dasd_block. For dasd_device related queue lists the ccwdev lock must be used.
Fix the mentioned issues by correctly using the ccwdev lock instead of the queue lock.
--- drivers/s390/block/dasd_ioctl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/s390/block/dasd_ioctl.c +2 -2
@@ -516,10 +516,10 @@ static int dasd_ioctl_information(struct dasd_block *block,memcpy (dasd_info->type, base->discipline->name,4 ); -spin_lock_irqsave (&block->queue_lock, flags); +spin_lock_irqsave (get_ccwdev_lock (base->cdev), flags);list_for_each (l, &base->ccw_queue) dasd_info->chanq_len++; -spin_unlock_irqrestore (&block->queue_lock, flags); +spin_unlock_irqrestore (get_ccwdev_lock (base->cdev), flags); rc =0 ; if (copy_to_user(argp, dasd_info,
Signature · 1 line
2.39.2
Sasha Levin
maintainerRe: [PATCH 5.4.y] s390/dasd: Use correct lock while counting channel queue length
[ Sasha's backport helper bot ]
Hi,
Summary of potential issues:
❌ Build failures detected
⚠️ Found matching upstream commit but patch is missing proper reference to it
Found matching upstream commit: ccc45cb4e7271c74dbb27776ae8f73d84557f5c6
Status in newer kernel trees:
6.13.y | Present (exact SHA1)
6.12.y | Present (exact SHA1)
6.6.y | Present (exact SHA1)
6.1.y | Present (different SHA1: 1e1e2ee0cf7f)
Note: The patch differs from the upstream commit:
---
Failed to apply patch cleanly.
---
Results of testing on various branches:
| Branch | Patch Apply | Build Test | |---------------------------|-------------|------------| | stable/linux-5.4.y | Failed | N/A |
Build Errors:
Patch failed to apply on stable/linux-5.4.y. Reject:
diff a/drivers/s390/block/dasd_ioctl.c b/drivers/s390/block/dasd_ioctl.c (rejected hunks)
@@ -516,10 +516,10 @@ static int dasd_ioctl_information(struct dasd_block *block,
memcpy(dasd_info->type, base->discipline->name, 4);
- spin_lock_irqsave(&block->queue_lock, flags);
+ spin_lock_irqsave(get_ccwdev_lock(base->cdev), flags);
list_for_each(l, &base->ccw_queue)
dasd_info->chanq_len++;
- spin_unlock_irqrestore(&block->queue_lock, flags);
+ spin_unlock_irqrestore(get_ccwdev_lock(base->cdev), flags);
rc = 0;
if (copy_to_user(argp, dasd_info,
Jan Höppner
maintainerPATCH 1/1 WIP
In reply to [email protected]--- arch/s390/include/asm/idals.h | 51 ++++ drivers/s390/char/tape.h | 9 +- drivers/s390/char/tape_34xx.c | 10 + drivers/s390/char/tape_char.c | 436 +++++++++++++++++++++++++++++---- drivers/s390/char/tape_class.c | 2 +- drivers/s390/char/tape_core.c | 5 +- drivers/s390/char/tape_std.c | 146 +++++------ drivers/s390/char/tape_std.h | 1 + 8 files changed, 512 insertions(+), 148 deletions(-)
arch/s390/include/asm/idals.h +51 -0
@@ -137,6 +137,7 @@ static inline struct idal_buffer *idal_buffer_alloc(size_t size, int page_order) nr_ptrs = (size +IDA_BLOCK_SIZE -1 ) >>IDA_SIZE_SHIFT ; nr_chunks = (PAGE_SIZE << page_order) >>IDA_SIZE_SHIFT ; +pr_warn ("DEBUG: ib struct_size: %lu\n" ,struct_size (ib, data, nr_ptrs)); ib =kmalloc (struct_size (ib, data, nr_ptrs),GFP_DMA |GFP_KERNEL );if (!ib)return ERR_PTR (-ENOMEM );
@@ -247,6 +248,8 @@ static inline size_t idal_buffer_from_user(struct idal_buffer *ib, const void __BUG_ON (count > ib->size);for (i =0 ; count >IDA_BLOCK_SIZE ; i++) { vaddr =dma64_to_virt (ib->data[i]); +pr_warn ("DEBUG(idal_buf_from_user): count(%d): %lu idaws: %llx vaddr: %p\n" , + i, count, ib->data[i], vaddr); left =copy_from_user (vaddr, from,IDA_BLOCK_SIZE );if (left)return left + count -IDA_BLOCK_SIZE ;
@@ -257,4 +260,52 @@ static inline size_t idal_buffer_from_user(struct idal_buffer *ib, const void __return copy_from_user (vaddr, from, count); } +static inline size_t idal_to_user (dma64_t *idaws,void __user *to,size_t count) +{ +size_t left; +void *vaddr; + +/* BUG_ON(count > ib->size); */ +while (count >IDA_BLOCK_SIZE ) { + vaddr =dma64_to_virt (*idaws++); + left =copy_to_user (to, vaddr,IDA_BLOCK_SIZE ); +pr_warn ("DEBUG(idal_to_user): vaddr %p left/count: %lu/%lu\n" , + vaddr, left, count); +if (left) +return left + count -IDA_BLOCK_SIZE ; + to = (void __user *)to +IDA_BLOCK_SIZE ; + count -=IDA_BLOCK_SIZE ; + } + vaddr =dma64_to_virt (*idaws); + left =copy_to_user (to, vaddr, count); +pr_warn ("DEBUG(idal_to_user last): vaddr %p left/count: %lu/%lu\n" , + vaddr, left, count); +return left; +/* return copy_to_user(to, vaddr, count); */ +} + +static inline size_t idal_from_user (dma64_t *idaws,const void __user *from, +size_t count) +{ +size_t left; +void *vaddr; + +/* BUG_ON(count > ib->size); */ +while (count >IDA_BLOCK_SIZE ) { + vaddr =dma64_to_virt (*idaws++); + left =copy_from_user (vaddr, from,IDA_BLOCK_SIZE ); +pr_warn ("DEBUG(idal_from_user): vaddr %p left/count: %lu/%lu\n" , + vaddr, left, count); +if (left) +return left + count -IDA_BLOCK_SIZE ; + from = (void __user *)from +IDA_BLOCK_SIZE ; + count -=IDA_BLOCK_SIZE ; + } + vaddr =dma64_to_virt (*idaws); + left =copy_from_user (vaddr, from, count); +pr_warn ("DEBUG(idal_from_user last): vaddr %p left/count: %lu/%lu\n" , + vaddr, left, count); +return left; +/* return copy_from_user(vaddr, from, count); */ +}#endif
drivers/s390/char/tape.h +6 -3
@@ -130,6 +130,7 @@ struct tape_request {int retries;/* retry counter for error recovery. */ int rescnt;/* residual count from devstat. */ struct timer_list timer;/* timer for std_assign_timeout(). */ +struct irb irb;/* device status */ /* Callback for delivering final status. */ void (*callback)(struct tape_request *,void *);
@@ -346,8 +347,9 @@ tape_ccw_repeat(struct ccw1 *ccw, __u8 cmd_code, int count)return ccw; } -static inline struct ccw1 * -tape_ccw_ida_idal (struct ccw1 *ccw,__u8 cmd_code,struct idal_buffer *idal) +static inline struct ccw1 *tape_ccw_ida_idal (struct ccw1 *ccw,__u8 cmd_code, +dma64_t *idaws,void *vaddr, +int length) { ccw->cmd_code = cmd_code; ccw->flags =CCW_FLAG_IDA ;
@@ -355,7 +357,8 @@ tape_ccw_ida_idal(struct ccw1 *ccw, __u8 cmd_code, struct idal_buffer *idal)// Create IDA Words to chain entire data list // idaws: cqr->data == idal buffer?! == ib->data?! // dst: vaddr buffer -//idal_create_words(idaws, dst, blksize); + ccw->cda =virt_to_dma32 (idaws); + idaws =idal_create_words (idaws, vaddr, length);return ccw++; }
drivers/s390/char/tape_34xx.c +10 -0
@@ -17,6 +17,9 @@#include <linux/workqueue.h> #include <linux/slab.h> +#include <asm/css_chars.h> +#include <asm/ccwdev.h> +#define TAPE_DBF_AREA tape_34xx_dbf#include "tape.h"
@@ -1059,6 +1062,9 @@ tape_34xx_setup_device(struct tape_device * device)struct list_head * discdata;struct tape_34xx_rdc_data *rdc_data; +int fcx_in_css; +unsigned int mdc; +DBF_EVENT (6 ,"34xx device setup\n" ); rdc_data =kmalloc (sizeof (*rdc_data),GFP_KERNEL |GFP_DMA );if (!rdc_data) {
@@ -1076,6 +1082,10 @@ tape_34xx_setup_device(struct tape_device * device)pr_warn ("DEBUG: max_blksize: %u opt_blksize: %u\n" , rdc_data->max_block_size, rdc_data->opt_block_size); + fcx_in_css = css_general_characteristics.fcx; + mdc =ccw_device_get_mdc (device->cdev,0 ); +pr_warn ("DEBUG: fcx: %d mdc: %u\n" , fcx_in_css, mdc); +if ((rc =tape_std_assign (device)) ==0 ) {if ((rc =tape_34xx_medium_sense (device)) !=0 ) {DBF_LH (3 ,"34xx medium sense returned %d\n" , rc);
drivers/s390/char/tape_char.c +382 -54
535 lines not on this page · raw message
drivers/s390/char/tape_class.c +1 -1
@@ -119,7 +119,7 @@ EXPORT_SYMBOL(unregister_tape_dev);static int __init tape_init (void ) { -pr_warn ("DEBUG: tape classe module reloaded... 1\n" ); +pr_warn ("DEBUG: tape classe module reloaded... 2\n" );return class_register (&tape_class); }
drivers/s390/char/tape_core.c +3 -2
@@ -1098,9 +1098,10 @@ __tape_do_irq (struct ccw_device *cdev, unsigned long intparm, struct irb *irb) }/* May be an unsolicited irq */ -if (request !=NULL ) +if (request !=NULL ) { request->rescnt = irb->scsw.cmd.count; -else if ((irb->scsw.cmd.dstat ==0x85 || irb->scsw.cmd.dstat ==0x80 ) && +memcpy (&request->irb, irb,sizeof (*irb)); + }else if ((irb->scsw.cmd.dstat ==0x85 || irb->scsw.cmd.dstat ==0x80 ) && !list_empty (&device->req_queue)) {/* Not Ready to Ready after long busy ? */ struct tape_request *req;
drivers/s390/char/tape_std.c +58 -88
384 lines not on this page · raw message
drivers/s390/char/tape_std.h +1 -0
@@ -102,6 +102,7 @@ struct tape_request *tape_std_read_block(struct tape_device *, size_t);void tape_std_read_backward (struct tape_device *device,struct tape_request *request);struct tape_request *tape_std_write_block (struct tape_device *,size_t ); +struct tape_request *tape_std_write_block_new (struct tape_device *,const void __user *,size_t );/* Some non-mtop commands. */ int tape_std_assign (struct tape_device *);
Signature · 1 line
2.45.2
Greg KH
maintainerRe: [PATCH 5.4.y] WIP
Quoted 10 lines · Jan Höppner
--- arch/s390/include/asm/idals.h | 51 ++++ drivers/s390/char/tape.h | 9 +- drivers/s390/char/tape_34xx.c | 10 + drivers/s390/char/tape_char.c | 436 +++++++++++++++++++++++++++++---- drivers/s390/char/tape_class.c | 2 +- drivers/s390/char/tape_core.c | 5 +- drivers/s390/char/tape_std.c | 146 +++++------ drivers/s390/char/tape_std.h | 1 + 8 files changed, 512 insertions(+), 148 deletions(-)
I'm confused, is this a patch submission of a backport?
Jan Höppner
maintainerRe: [PATCH 5.4.y] WIP
Quoted 12 lines · Greg KH
Quoted 10 lines · Jan Höppner
--- arch/s390/include/asm/idals.h | 51 ++++ drivers/s390/char/tape.h | 9 +- drivers/s390/char/tape_34xx.c | 10 + drivers/s390/char/tape_char.c | 436 +++++++++++++++++++++++++++++---- drivers/s390/char/tape_class.c | 2 +- drivers/s390/char/tape_core.c | 5 +- drivers/s390/char/tape_std.c | 146 +++++------ drivers/s390/char/tape_std.h | 1 + 8 files changed, 512 insertions(+), 148 deletions(-)
Quoted 2 lines
I'm confused, is this a patch submission of a backport?
Hi Greg,
I'm very sorry, please ignore this completely. I was clearly to dumb to use my bash backward search correctly. I accidentally ran an old command that send this WIP patch, which is not meant to be for the outside world at all yet.
I already put a safety net in place so this won't happen again! My apologies for the inconvenience.
regards,
Jan