Aren/haiku
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

kernel/vfs: Remove read sizing hack in common_file_io_vec_pages.

Browse Source 
All drivers using IORequest should have any problems here already caught
by it, and overruns in other drivers would overrun the buffer and cause
corruptions, so we should no longer silently be ignoring this,
presuming it is still a problem at all.

As the TODO comment indicated that this logic could also be used for
writes if this wasn't a problem, I have added a separate TODO for that.
This commit is contained in:
Augustin Cavalier 2023-02-13 23:12:23 -05:00
parent 478c8c847d
commit 81458ef8a6
1 changed files with 1 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/system/kernel/fs/vfs.cpp
View File

@ -3489,6 +3489,7 @@ common_file_io_vec_pages(struct vnode* vnode, void* cookie,
if (!doWrite && vecOffset == 0) {
// now directly read the data from the device
// the first file_io_vec can be read directly
// TODO: we could also write directly
if (fileVecs[0].length < (off_t)numBytes)
size = fileVecs[0].length;
@ -3506,17 +3507,6 @@ common_file_io_vec_pages(struct vnode* vnode, void* cookie,
if (status != B_OK)
return status;
// TODO: this is a work-around for buggy device drivers!
// When our own drivers honour the length, we can:
// a) also use this direct I/O for writes (otherwise, it would
// overwrite precious data)
// b) panic if the term below is true (at least for writes)
if ((off_t)size > fileVecs[0].length) {
//dprintf("warning: device driver %p doesn't respect total length "
// "in read_pages() call!\n", ref->device);
size = fileVecs[0].length;
}
ASSERT((off_t)size <= fileVecs[0].length);
// If the file portion was contiguous, we're already done now