issue(fs): assertion !frame_is_full(ctx->frame) failed at /nanos/src/kernel/mutex.c:109 (IP 0xffffffff800460b4) #1933
Assignees
Comments
rinor
changed the title
francescolavra
added a commit
that referenced
this issue
Sep 5, 2023
A call to storage_sync() may suspend the current context, e.g. to acquire the filsystem mutex. Therefore, this function cannot be called directly from the code that handles a page fault exception; instead, an asynchronus thunk should be used. This commit changes the mmap_anon_page() closure function so that it calls storage_sync() first, and the demand_anonymous_page() function so that it invokes the above closure via async_apply_bh(). Closes #1933.
francescolavra
added a commit
that referenced
this issue
Sep 5, 2023
A call to storage_sync() may suspend the current context, e.g. to acquire the filesystem mutex. Therefore, this function cannot be called directly from the code that handles a page fault exception; instead, an asynchronous thunk should be used. This change modifies the mmap_anon_page() closure function so that it calls storage_sync() first, and the demand_anonymous_page() function so that it invokes the above closure via async_apply_bh(). Closes #1933.
|
tested #1934 and can confirm it works as expected. Thank you |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
While re-testing #1845 after #1904 merged on main branch at 2da53c8 got:
Note: This (similar) was observed also on #1874 but on other conditions and fixed at #1875
config.json{ "BaseVolumeSz": "4m", "Args": [ "-w=10", "-f=1", "-s=128", "-t=0", "-c=false", "-d=/tfs_mnt" ], "Mounts": { "stressdisk": "tfs_mnt" }, "RunConfig": { "Memory": "614m" }, "Env": { "GODEBUG": "gctrace=0,scavtrace=0" }, "Boot": "/usr/local/src/nanos/output/platform/pc/boot/boot.img", "Kernel": "/usr/local/src/nanos/output/platform/pc/bin/kernel.img", "KlibDir": "/usr/local/src/nanos/output/klib/bin", "Debugflags": [ "reboot_on_exit" ] }ops volume create stressdisk -s 1G# extra volume clean storageops run -c config.json stressdisk --smp=4#with clean storageThe text was updated successfully, but these errors were encountered: