nexoniarz/SchOS
nexoniarz/SchOS
Watch
1
0
Fork
You've already forked SchOS
 0
Code Issues Pull requests Projects Releases 1 Packages Wiki Activity Actions

dev: raw disks are 64-bit, so a real drive stops wrapping at 4 GB

Browse source 
struct file already carried a uint64 offset and bread already took a
uint64 block number. The narrowing was one boundary in the middle:
devsw's read and write took a 32-bit uint, so fileread passed a 64-bit
f->off into a 32-bit parameter and the top half was discarded.

The effect was not a refusal at 4 GB but silent wraparound past it. A
read at 5 GB returned the bytes at 1 GB, reporting success. disk.c had a
DISK_MAX_BYTES constant documenting the limit, and nothing anywhere
referenced it -- the cap was described but never enforced, which is the
worst of the three options.

The signature is now uint64 across all nine files that carry it, and
diskread/diskwrite no longer narrow the block number on the way to
bread. Widening this is a prerequisite for NVMe and for any real disk:
namespaces start well above 4 GB and the failure mode reads valid data
from the wrong place.

Verified end to end rather than by inspection. A 6 GB sparse image with
"PAST4GB-MARKER-OK" written at byte 5368709120, attached as /dev/hdb to a
ramdisk-root kernel so device 1 was free:

  # dd if=/dev/hdb bs=4096 skip=1310720 count=1 | od -c | head -2
  0000000   P   A   S   T   4   G   B   -   M   A   R   K   E   R   -   O

Under the old code that offset truncates to 1 GB, where the image is
zeroes. The whole path is exercised: f->off, devsw, diskread, bread and
ide.c's LBA48.

Worth recording for next time: `make ramroot` stages the ESP itself, and
running `make esp` afterwards rebuilds the kernel without RAMROOT and
silently undoes it. The tell is a 855 KB kernel where a 25 MB one is
expected.

110 [ok], no failures, on the default build.
This commit is contained in:
Nexoniarz 2026-08-11 12:15:38 +02:00
commit ad2b321d56
9 changed files with 42 additions and 41 deletions
Download patch file Download diff file Expand all files Collapse all files

16
include/defs.h
View file

@ -198,8 +198,8 @@ int schfs_kreaddir_dev(uint dev, uint64 dir,
uint uptime(void);
// procfs.c
int procread(struct inode*, uint, char*, int);
int procwrite(struct inode*, uint, char*, int);
int procread(struct inode*, uint64, char*, int);
int procwrite(struct inode*, uint64, char*, int);
// Shared with the /proc filesystem (kernel/fs/procfs.c): one owns the
// namespace, the other the content.
int proc_gen_global(int id, char *buf, int cap);
@ -208,8 +208,8 @@ void procfs_mount(void);
// input.c -- raw evdev event devices
void inputinit(void);
int inputread(struct inode*, uint, char*, int);
int inputwrite(struct inode*, uint, char*, int);
int inputread(struct inode*, uint64, char*, int);
int inputwrite(struct inode*, uint64, char*, int);
int inputready(struct inode*);
void inputclose(struct inode*);
void inputopen(struct inode*);
@ -431,13 +431,13 @@ void fbclose(struct inode*);
// Called from sys_mmap: a mapping of /dev/fb0 is a licence to draw with no
// further syscalls, so it is the last the kernel hears about it.
void fb_mapped(void);
int fbread(struct inode*, uint, char*, int);
int fbwrite(struct inode*, uint, char*, int);
int fbread(struct inode*, uint64, char*, int);
int fbwrite(struct inode*, uint64, char*, int);
int fb_ioctl(int request, addr_t arg);
// dev/disk.c -- /dev/hdN, raw seekable access to a whole disk. The minor is the
// IDE device number. Writes bypass every filesystem above them.
int diskread(struct inode*, uint, char*, int);
int diskwrite(struct inode*, uint, char*, int);
int diskread(struct inode*, uint64, char*, int);
int diskwrite(struct inode*, uint64, char*, int);
// drivers/net/e1000 -- raw Ethernet. No protocol stack above it yet, so these
// are the whole interface: queue a frame, take a frame, and the MAC.
int e1000_transmit(const void *frame, int len);