wireguard-linux/drivers/scsi, branch update-toolchain

Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

2024-08-17T17:04:01Z

Pull SCSI fixes from James Bottomley: "Two small fixes to the mpi3mr driver. One to avoid oversize allocations in tracing and the other to fix an uninitialized spinlock in the user to driver feature request code (used to trigger dumps and the like)" * tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: scsi: mpi3mr: Avoid MAX_PAGE_ORDER WARNING for buffer allocations scsi: mpi3mr: Add missing spin_lock_init() for mrioc->trigger_lock

scsi: mpi3mr: Avoid MAX_PAGE_ORDER WARNING for buffer allocations

2024-08-13T00:51:37Z

Commit fc4444941140 ("scsi: mpi3mr: HDB allocation and posting for hardware and firmware buffers") added mpi3mr_alloc_diag_bufs() which calls dma_alloc_coherent() to allocate the trace buffer and the firmware buffer. mpi3mr_alloc_diag_bufs() decides the buffer sizes from the driver configuration. In my environment, the sizes are 8MB. With the sizes, dma_alloc_coherent() fails and report this WARNING: WARNING: CPU: 4 PID: 438 at mm/page_alloc.c:4676 __alloc_pages_noprof+0x52f/0x640 The WARNING indicates that the order of the allocation size is larger than MAX_PAGE_ORDER. After this failure, mpi3mr_alloc_diag_bufs() reduces the buffer sizes and retries dma_alloc_coherent(). In the end, the buffer allocations succeed with 4MB size in my environment, which corresponds to MAX_PAGE_ORDER=10. Though the allocations succeed, the WARNING message is misleading and should be avoided. To avoid the WARNING, check the orders of the buffer allocation sizes before calling dma_alloc_coherent(). If the orders are larger than MAX_PAGE_ORDER, fall back to the retry path. Fixes: fc4444941140 ("scsi: mpi3mr: HDB allocation and posting for hardware and firmware buffers") Signed-off-by: Shin'ichiro Kawasaki Link: https://lore.kernel.org/r/20240810042701.661841-3-shinichiro.kawasaki@wdc.com Acked-by: Sathya Prakash Veerichetty Signed-off-by: Martin K. Petersen

scsi: mpi3mr: Add missing spin_lock_init() for mrioc->trigger_lock

2024-08-12T22:22:04Z

Commit fc4444941140 ("scsi: mpi3mr: HDB allocation and posting for hardware and firmware buffers") added the spinlock trigger_lock to the struct mpi3mr_ioc. However, spin_lock_init() call was not added for it, then the lock does not work as expected. Also, the kernel reports the message below when lockdep is enabled. INFO: trying to register non-static key. The code is fine but needs lockdep annotation, or maybe you didn't initialize this object before use? To fix the issue and to avoid the INFO message, add the missing spin_lock_init() call. Fixes: fc4444941140 ("scsi: mpi3mr: HDB allocation and posting for hardware and firmware buffers") Signed-off-by: Shin'ichiro Kawasaki Link: https://lore.kernel.org/r/20240810042701.661841-2-shinichiro.kawasaki@wdc.com Acked-by: Sathya Prakash Veerichetty Signed-off-by: Martin K. Petersen

Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

2024-08-11T16:28:04Z

Pull SCSI fixes from James Bottomley: "Two core fixes: one to prevent discard type changes (seen on iSCSI) during intermittent errors and the other is fixing a lockdep problem caused by the queue limits change. And one driver fix in ufs" * tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: scsi: sd: Keep the discard mode stable scsi: sd: Move sd_read_cpr() out of the q->limits_lock region scsi: ufs: core: Fix hba->last_dme_cmd_tstamp timestamp updating logic

Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi

2024-08-03T22:12:56Z

Pull SCSI fixes from James Bottomley: "One core change that reverts the double message print patch in sd.c (it was causing regressions on embedded systems). The rest are driver fixes in ufs, mpt3sas and mpi3mr" * tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: scsi: ufs: exynos: Don't resume FMP when crypto support is disabled scsi: mpt3sas: Avoid IOMMU page faults on REPORT ZONES scsi: mpi3mr: Avoid IOMMU page faults on REPORT ZONES scsi: ufs: core: Do not set link to OFF state while waking up from hibernation scsi: Revert "scsi: sd: Do not repeat the starting disk message" scsi: ufs: core: Fix deadlock during RTC update scsi: ufs: core: Bypass quick recovery if force reset is needed scsi: ufs: core: Check LSDBS cap when !mcq

scsi: sd: Keep the discard mode stable

2024-08-03T01:58:41Z

There is a scenario where a large number of discard commands are issued when the iscsi initiator connects to the target and then performs a session rescan operation. There is a time window, most of the commands are in UNMAP mode, and some discard commands become WRITE SAME with UNMAP. The discard mode has been negotiated during the SCSI probe. If the mode is temporarily changed from UNMAP to WRITE SAME with UNMAP, an I/O ERROR may occur because the target may not implement WRITE SAME with UNMAP. Keep the discard mode stable to fix this issue. Signed-off-by: Li Feng Link: https://lore.kernel.org/r/20240718080751.313102-2-fengli@smartx.com Reviewed-by: Christoph Hellwig Reviewed-by: Martin K. Petersen Signed-off-by: Martin K. Petersen

scsi: sd: Move sd_read_cpr() out of the q->limits_lock region

2024-08-02T02:02:34Z

Commit 804e498e0496 ("sd: convert to the atomic queue limits API") introduced pairs of function calls to queue_limits_start_update() and queue_limits_commit_update(). These two functions lock and unlock q->limits_lock. In sd_revalidate_disk(), sd_read_cpr() is called after queue_limits_start_update() call and before queue_limits_commit_update() call. sd_read_cpr() locks q->sysfs_dir_lock and &q->sysfs_lock. Then new lock dependencies were created between q->limits_lock, q->sysfs_dir_lock and q->sysfs_lock, as follows: sd_revalidate_disk queue_limits_start_update mutex_lock(&q->limits_lock) sd_read_cpr disk_set_independent_access_ranges mutex_lock(&q->sysfs_dir_lock) mutex_lock(&q->sysfs_lock) mutex_unlock(&q->sysfs_lock) mutex_unlock(&q->sysfs_dir_lock) queue_limits_commit_update mutex_unlock(&q->limits_lock) However, the three locks already had reversed dependencies in other places. Then the new dependencies triggered the lockdep WARN "possible circular locking dependency detected" [1]. This WARN was observed by running the blktests test case srp/002. To avoid the WARN, move the sd_read_cpr() call in sd_revalidate_disk() after the queue_limits_commit_update() call. In other words, move the sd_read_cpr() call out of the q->limits_lock region. [1] https://lore.kernel.org/linux-scsi/vlmv53ni3ltwxplig5qnw4xsl2h6ccxijfbqzekx76vxoim5a5@dekv7q3es3tx/ Fixes: 804e498e0496 ("sd: convert to the atomic queue limits API") Signed-off-by: Shin'ichiro Kawasaki Link: https://lore.kernel.org/r/20240801054234.540532-1-shinichiro.kawasaki@wdc.com Tested-by: Luca Coelho Reviewed-by: Damien Le Moal Reviewed-by: Christoph Hellwig Reviewed-by: Bart Van Assche Signed-off-by: Martin K. Petersen

Merge branch '6.11/scsi-queue' into 6.11/scsi-fixes

2024-07-30T01:46:16Z

Pull outstanding commits from 6.11 queue into fixes. Signed-off-by: Martin K. Petersen

minmax: scsi: fix mis-use of 'clamp()' in sr.c

2024-07-29T00:06:20Z

While working on simplifying the minmax functions, and avoiding excessive macro expansion, it turns out that the sr.c use of the 'clamp()' macro has the arguments the wrong way around. The clamp logic is val = clamp(in, low, high); and it returns the input clamped to the low/high limits. But sr.c ddid speed = clamp(0, speed, 0xffff / 177); which clamps the value '0' to the range '[speed, 0xffff / 177]' and ends up being nonsensical. Happily, I don't think anybody ever cared. Fixes: 9fad9d560af5 ("scsi: sr: Fix unintentional arithmetic wraparound") Cc: Justin Stitt Cc: Kees Cook Cc: Martin K. Petersen Signed-off-by: Linus Torvalds

minmax: make generic MIN() and MAX() macros available everywhere

2024-07-28T22:49:18Z

This just standardizes the use of MIN() and MAX() macros, with the very traditional semantics. The goal is to use these for C constant expressions and for top-level / static initializers, and so be able to simplify the min()/max() macros. These macro names were used by various kernel code - they are very traditional, after all - and all such users have been fixed up, with a few different approaches: - trivial duplicated macro definitions have been removed Note that 'trivial' here means that it's obviously kernel code that already included all the major kernel headers, and thus gets the new generic MIN/MAX macros automatically. - non-trivial duplicated macro definitions are guarded with #ifndef This is the "yes, they define their own versions, but no, the include situation is not entirely obvious, and maybe they don't get the generic version automatically" case. - strange use case #1 A couple of drivers decided that the way they want to describe their versioning is with #define MAJ 1 #define MIN 2 #define DRV_VERSION __stringify(MAJ) "." __stringify(MIN) which adds zero value and I just did my Alexander the Great impersonation, and rewrote that pointless Gordian knot as #define DRV_VERSION "1.2" instead. - strange use case #2 A couple of drivers thought that it's a good idea to have a random 'MIN' or 'MAX' define for a value or index into a table, rather than the traditional macro that takes arguments. These values were re-written as C enum's instead. The new function-line macros only expand when followed by an open parenthesis, and thus don't clash with enum use. Happily, there weren't really all that many of these cases, and a lot of users already had the pattern of using '#ifndef' guarding (or in one case just using '#undef MIN') before defining their own private version that does the same thing. I left such cases alone. Cc: David Laight Cc: Lorenzo Stoakes Signed-off-by: Linus Torvalds