old-updates.txt

FILE ./threads-cv/1.txt
Fixed typo and other text in citation [LR80] at end of chapter

Jersen Deng (UCLA)




FILE ./file-implementation/2.txt
replace ~ in URL in fat citation to %7e
(which works with linking)

Nizare Leonetti



FILE ./file-implementation/1.txt
20-KB in size (5 4KB blocks) ->
20KB in size (five 4KB blocks)

Dakota Cookenmaster




FILE ./vm-smalltables/1.txt
At the 11th page, in the first paragraph right under the first picture, change the following sentences:

"If our page directory has 2^14 entries, it spans not one page but 128."

to:

"If our page directory has 2^14 entries, assuming each PDE size also of 4 bytes, it spans not one page but 128."

reason: the PDE size is not mentioned, which will bring misunderstanding to the calculation




FILE ./vm-intro/2.txt
13.5: The VM system is responsible for providing the illusion of a large,
sparse, private address space to programs, which hold all of their instructions and data therein.

->

13.5: The VM system is responsible for providing the illusion of a large,
sparse, private address space to each program, which holds all of their instructions and data therein.



FILE ./vm-intro/1.txt
typo in citation M83, now fixed

Abhishek Bhattacherjee (NITR)



FILE ./cpu-api/1.txt
Version 1.01, Page 4, Footnote 3

execl -> execl()		# at the bottom of the page



FILE ./security-distributed/1.txt
p.13

[NT95]
->
[NT94]

Kyutae Lee



FILE ./intro/3.txt
In the body of the text, [S+68] should be edited to be [S68] to match the item in references.

Kyutae Lee



FILE ./intro/2.txt
./mem &; ./mem &
should be
./mem & ./mem &



FILE ./intro/1.txt
Add "common_threads.h" to threads.c code listing to match code in ostep-code

Patrick Elsen




FILE ./file-dialogue/1.txt
Page-1. 3rd line from bottom.
"Check out some the latest..." -> "Check out some (of) the latest...".



FILE ./security-crypto/1.txt
p.7

And we’d better be careful about it, since Bob is going to assume that
messages encrypted
with the public key he thinks belongs to Alice were actually created by Alice.

->

And we’d better be careful about it, since Bob is going to assume that
messages decrypted
with the public key he thinks belongs to Alice were actually created by Alice.


That is, "encrypted" should be changed to "decrypted".

Kyutae Lee



FILE ./vm-complete/3.txt
p.12

In this way, most applications would be unaffected (and continue to use only 4-KB pages;
->
In this way, most applications would be unaffected (and continue to use only 4-KB pages);

Kyutae Lee



FILE ./vm-complete/2.txt
p.3

One neat aspect of studying VMS is that we can see how a real address space is constructed (Figure 23.1.
->
One neat aspect of studying VMS is that we can see how a real address space is constructed (Figure 23.1).

Kyutae Lee



FILE ./vm-complete/1.txt
page 16, "See meltdownattack.com
and spectreattack.com for papers describing each attack in detail."
meltdownattack.com and spectreattack.com both refer to the same website, there is no need to reference both.

Matthew Prinz (UT)

AND

"...if not, the CPU undoes their effects on architectural state (e.g., registers) tries again" ->
"...if not, the CPU undoes their effects on the architectural state (e.g., registers) and tries again"

Noah Jackson

and

"Even the code in main()..." --> "Even though the code in main...()"

missing "though"



FILE ./preface/2.txt
"The first are systems programming projects" -> "The first is systems programming projects"
"The second type are ..." -> "The second type is ..."

Po-Hao Su (Taiwan)



FILE ./file-integrity/2.txt
p.1

latent-sector errors (LSEs)
->
latent sector errors (LSEs)

Kyutae Lee



FILE ./file-integrity/1.txt
cite [M86] changed to [MK96] (Giovanni Moricca)





FILE ./vm-beyondphys-policy/3.txt
Page 15: "leading to a renaissance in page replacement algoriths."

"algoriths" -> "algorithms"

Jack Xu (UW-Madison)



FILE ./vm-beyondphys-policy/2.txt
Please forgive me if I'm mistaken, but within section 22.3, the last word of the second paragraph ("3") I believe it's supposed to say "1".

Holly Johnson (USF)




FILE ./vm-beyondphys-policy/1.txt
the use bit for [math P] set to 0 (cleared) ->
the use bit for [math P] is set to 0 (cleared)

Jersey Deng (UCLA)



FILE ./dist-intro/1.txt
p.1

they underly
->
they underlie

Kyutae Lee



FILE ./vm-complete-mem-systems/1.txt
Page 10 in "VM - COMPLETE VIRTUAL MEMORY SYSTEMS", line 9
directory memory access -> direct memory access

(Amusingly, the typo is in bold.)
Yuxing Xiang [Peking University]



FILE ./cpu-sched-lottery/8.txt
On p.7, reference to "[K+15]" is missing.

Recent studies have shownthat scheduler efficiency is surprisingly important; specifically,
in a studyof Google datacenters, Kanev et al. show that even after aggressive opti-mization,
scheduling uses about 5% of overall datacenter CPU time.

->

Recent studies have shownthat scheduler efficiency is surprisingly important; specifically,
in a studyof Google datacenters, Kanev et al. show that even after aggressive opti-mization,
scheduling uses about 5% of overall datacenter CPU time [K+15].

Kyutae Lee



FILE ./cpu-sched-lottery/5.txt
43 0 49 ->
43 0 49 12

and

A A A ->
A A A A


Stardustman




FILE ./cpu-sched-lottery/4.txt
Unfairness metric to Fairness metric

U to F throughout

Figure change to match

can be quite severe ->
can be quite low.

Otto Sievert



FILE ./cpu-sched-lottery/6.txt
In page 10, under the "Using Red-Black Trees" section, in the first sentence of the third paragraph, there's a typo.
(OLD) CFS does not keep all 'process' in this structure;
(NEW) CFS does not keep all 'processes' in this structure;

FUN FACT: This page has a footer admitting a deliberate grammatical typo elsewhere in the page to mimic Gollum from LOTR.

Thanks
Phani Karan



FILE ./cpu-sched-lottery/7.txt
In References in p.13

In the description for [B+18],
there is only one closing parenthesis while there are two opening parentheses.
I think there should be one more closing parenthesis somewhere appropriate.

Kyutae Lee



FILE ./cpu-sched-lottery/3.txt
it inversely by the weight of the process ->
it inversely by the weight of the process, by dividing the default weight of 1024 (weight_0) by its weight, weight_i.

Chien Chi



FILE ./cpu-sched-lottery/2.txt
The formula used to do so is as follows: ->
The formula used to do so is as follows, assuming n processes:

Chien Chi



FILE ./cpu-sched-lottery/1.txt
Change name of stride scheduling section to stride scheduling (Hein Meling)



FILE ./file-disks/3.txt
In References on p.15,
on the comments for [JW91]

scooped by Seltzer et al. [S90].
->
scooped by Seltzer et al. [SCO90].

Kyutae Lee



FILE ./file-disks/2.txt
However, when updating the disk, the only guarantee drive manufactures make is that a single 512-byte write is atomic
->
However, when updating the disk, the only guarantee drive manufacturers make is that a single 512-byte write is atomic



FILE ./file-disks/1.txt
Using SSDs well seems as complicated than hard drives, and sometimes more so. ->
Using SSDs well seems as complicated as hard drives, and sometimes more so.

Jersey Deng (UCLA)



FILE ./dist-afs/1.txt
AFSv2 was measured and found to be much more scalable that the original version.
->
AFSv2 was measured and found to be much more scalable than the original version.
(that -> than)


"Figure 50.4 (page 9) shows the results of our qualitative comparison"
should probably remove the page reference "(page 9)". Because the page number won't make sense in a printed book.

Sirui Chen



FILE ./file-ffs/2.txt
Chen Huo (Shippensburg University)

-I 5 -> -i 5

(Homework question #6, wrong option for specifying inode table size)



FILE ./file-ffs/1.txt
Tian Guo (WPI)

3.69 -> 3.6
40.6 -> 39.6

(wrong numbers in "large file" calculation)




FILE ./file-lfs/1.txt
“The Unwritten Contract of of Solid State Drives” by Jun He, Sudarsun Kannan, Andrea C. Arpaci-Dusseau, Remzi H. Arpaci-Dusseau
is referred as [H+17] in this chapter.
But in other chapters (file-disks and file-ssd), it is referred as [HK+17].
So I think it is appropriate to change each [H+17] in this chapter to [HK+17].
(When I searched "[HK+17]" by cmd+f shortcut on the full pdf file purchased on Gumroad,
the reference in this chapter did not appear since it was referred as [H+17].)

[H+17]
->
[HK+17]



FILE ./vm-beyondphys/1.txt
"With a large address space, you don’t have to worry
about if there is room enough in memory for your program’s data struc-
tures;"

Should be

With a large address space, you don’t have to worry
about if there is [enough room] in memory for your program’s data struc-
tures;

Efkan S. Goktepe



FILE ./file-raid/2.txt
Add footnote to explain bad mirrored read/write performance, kind of:

The $1/2$ penalty assumes a naive read/write pattern for mirroring; a more
sophisticated approach that issued large I/O requests to differing parts of
each mirror could potentially achieve full bandwidth. Think about this to see
if you can figure out why.

(many people)




FILE ./file-raid/1.txt
or (RAID 1+0) ->
(or RAID 1+0, stripe of mirrors)

(or RAID 0+1) ->
(or RAID 0+1, mirror of stripes)

Giovanni Di Santi




FILE ./vm-tlbs/1.txt
int sum = 0; ->
int i, sum = 0;


Santiago Marini



FILE ./threads-locks/3.txt
p.18

dequeue the running thread.
->
deschedule the running thread.

Kyutae Lee



FILE ./threads-locks/2.txt
Section 28.14

"The scheduler determines which thread runs next; if
the scheduler makes a bad choice, a thread runs that must either spin
waiting for the lock (our first approach), or yield the CPU immediately
(our second approach)."

I think the intended meaning was:

"The scheduler determines which thread runs next; if
the scheduler makes a bad choice, a thread **that runs** must either spin
waiting for the lock (our first approach), or yield the CPU immediately
(our second approach)."

Efkan S. Goktepe



FILE ./threads-locks/1.txt
Locks, page 13, section 28.12 "Too Much Spinning: What Now?"

The first sentence reads as follows:

"Our simple hardware-based locks are simple (...) and they work (...), which are two excellent properties of any system or code."

I believe the use of "simple" as an adjective twice is redundant, and could be re-written by removing the first occurrence without loss of meaning.

-> "Our hardware-based locks are simple (...) and they work (...), which are two excellent properties of any system or code."

Efkan S. Goktepe



FILE ./threads-sema/2.txt
In footnote 3 on p.20:

The url for the book “Little Book of Semaphores” was changed.
The previous url is unavailable.

http://greenteapress.com/semaphores/downey08semaphores.pdf
->
https://greenteapress.com/semaphores/LittleBookOfSemaphores.pdf

Kyutae Lee



FILE ./threads-sema/1.txt
int i, tmp=0; ->
int tmp=0;
(fig 31.10)

Matías De Pascuale - FaMAF Universidad Nacional De Cordoba



FILE ./threads-bugs/1.txt
In References on p.15,
on the comment for [H93],

ensure than
->
ensure that

Kyutae Lee




FILE ./threads-locks-usage/5.txt
In References on p.15,
typo on the title for [MS98].

“Nonblocking Algorithms and Preemption-safe Locking on by Multiprogrammed Sharedmemory Multiprocessors. ”
->
“Nonblocking Algorithms and Preemption-Safe Locking on Multiprogrammed Shared Memory Multiprocessors”

Kyutae Lee



FILE ./threads-locks-usage/4.txt
p.15 References

On the comment for [C06], a period mark "." is missing.

in Linux This article
->
in Linux. This article

Kyutae Lee



FILE ./threads-locks-usage/3.txt
[C06] [...] LWN has many wonderful articles about the latest in Linux This article is a short [...]
-->
[C06] [...] LWN has many wonderful articles about the latest in Linux. This article is a short [...]



FILE ./threads-locks-usage/2.txt
missing "queue_t" from 29.9

Jason Waterman




FILE ./threads-locks-usage/1.txt
in homework part:

interested -> interesting in question 5

Reynaldo H. Verdejo Pinochet



FILE ./vm-paging/1.txt
1036 jne  0x1024 ->
1036 jne  1024



FILE ./threads-intro/1.txt
while he was at the Technische Hochshule of Eindhoven (THE)
->
while he was at the Technische Hogeschool Eindhoven (THE)

Felix Berief



FILE ./vm-mechanism/2.txt

page 7:
a virtual address that is greater than the bounds ->
a virtual address that is greater than or equal to the bounds

thasinaz

AND

In the 5th page "ASIDE: WHY HARDWARE DOESN’T HANDLE PAGE FAULTS", first sentence is "We know from our experience with the TLB that hardware designers are
loathe to trust the OS to do much of anything", but I think it should be "loath" instead of "loathe".
(but this should be in vm-beyondphys)

AND

page 13:
"In our current approach, although there might be enough physical
memory for more processes, we are currently restricted to placing
an address space in a fixed-sized slot and thus internal fragmentation
can arise"
->
"In our current approach, although there might be enough physical
memory for more processes, we are currently restricted to placing
an address space in a fixed-sized slot, thereby inefficiently
using physical memory and unnecessarily limiting our machines capability."

AND

last page (homework), question 2. "What value do you have set" -> "What value do you have to set"





FILE ./vm-mechanism/1.txt
page 3: "client of the interface" -> "interface of the client" (Khaled Emara)




FILE ./file-devices/2.txt
In References on p.14,
on the comment for [H18]

the better aggregrators
->
the better aggregators

Kyutae Lee



FILE ./file-devices/1.txt
On p.6

TIP: INTERRUPTS NOT ALWAYS BETTER THAN PIO
->
TIP: INTERRUPTS NOT ALWAYS BETTER THAN POLLING

Kyutae Lee



FILE ./cpu-mechanisms/1.txt
In Figure 6.4,

# void swtch(struct context **old, struct context *new);

should be changed to

# void swtch(struct context *old, struct context *new);

-----

The current code in OSTEP Figure 6.4 is buggy.
I found that others already found the same bug many years ago (see the comments at https://stackoverflow.com/q/45929162/8581025).

Original source code of xv6:
https://github.com/mit-pdos/xv6-public/blob/c95bde8163eca7a19c95f3b05156e0ad33312589/swtch.S

-----

Kyutae Lee



FILE ./dist-nfs/4.txt
p.11

The cache consistency problem is best illustrated with two clients and
a single server.
->
The cache consistency problem is best illustrated with three clients and
a single server.

When we read the context, we can find that there are "three" clients,
not "two", in the illustration used.

Kyutae Lee



FILE ./dist-nfs/3.txt
pp.6-7

-----

return the data to the client (or an error if there was a failure).
->
return the data and the file attributes after the read takes place
to the client (or an error if there was a failure).

Reference: https://www.rfc-editor.org/rfc/rfc1094.html#section-2.2.7
-----

and just a success code is returned.
->
and just a success code and the file attributes after the write has completed is returned.

Reference: https://www.rfc-editor.org/rfc/rfc1094.html#section-2.2.9
-----

Kyutae Lee



FILE ./dist-nfs/2.txt
If we compare "Figure 49.4: The NFS Protocol: Examples"
with RFC 1094 (https://www.rfc-editor.org/rfc/rfc1094.html) which is
a specification for NFSv2, we can find Figure 49.4 has some errors.

---

NFSPROC_GETATTR  file handle
returns: attributes
NFSPROC_SETATTR  file handle, attributes
returns: –
NFSPROC_LOOKUP   directory file handle, name of file/dir to look up
returns: file handle
NFSPROC_READ     file handle, offset, count
data, attributes
NFSPROC_WRITE    file handle, offset, count, data
attributes
NFSPROC_CREATE   directory file handle, name of file, attributes
–
NFSPROC_REMOVE   directory file handle, name of file to be removed
–
NFSPROC_MKDIR    directory file handle, name of directory, attributes
file handle
NFSPROC_RMDIR    directory file handle, name of directory to be removed
–
NFSPROC_READDIR  directory handle, count of bytes to read, cookie
returns: directory entries, cookie (to get more entries)


->


NFSPROC_GETATTR  file handle
returns: attributes
NFSPROC_SETATTR  file handle, attributes
returns: attributes
NFSPROC_LOOKUP   directory file handle, name of file/dir to look up
returns: file handle, attributes
NFSPROC_READ     file handle, offset, count
returns: data, attributes
NFSPROC_WRITE    file handle, offset, data
returns: attributes
NFSPROC_CREATE   directory file handle, name of file, attributes
returns: file handle, attributes
NFSPROC_REMOVE   directory file handle, name of file to be removed
returns: –
NFSPROC_MKDIR    directory file handle, name of directory, attributes
returns: file handle, attributes
NFSPROC_RMDIR    directory file handle, name of directory to be removed
returns: –
NFSPROC_READDIR  directory handle, count of bytes to read, cookie
returns: directory entries, cookie (to get more entries),
eof (TRUE if there are no more entries in the directory)


---

I was not sure I was doing right but I also deleted "count" in the arguments
for NFSPROC_WRITE because RFC 1094 says:
'The arguments "beginoffset" and "totalcount" are ignored and are removed
in the next protocol revision.'
Maybe "count" information is included in "nfsdata" in RFC 1094? I don't know..

Kyutae Lee



FILE ./dist-nfs/1.txt
Page 7 of Network File System chapter
Last line of the first paragraph under section 49.6:

"information needed to complete request." -> "information needed to complete the request."

(missing "the")

Daniela Ferreira Franco Moura



FILE ./vm-segmentation/1.txt
p.11
that underly computing systems today.
->
that underlie computing systems today.

Kyutae Lee



FILE ./file-ssd/5.txt
p.24 Homework Question #7

the low watermark (-G M)
->
the low watermark (-g M)

Kyutae Lee



FILE ./file-ssd/4.txt
p.18
"how to build file systems for hard drives" was discussed in chapter 41, chapter 42 and chapter 43.
This chapter (Flash-based SSDs) is chapter 44.

we will learn in subsequent chapters
->
we learned in previous chapters

Kyutae Lee



FILE ./file-ssd/3.txt
p.8
"Log-structured File Systems" is chapter 43.
And this chapter ("Flash-based SSDs") is chapter 44.

as we’ll see in the chapter on log-structured file systems
->
as we saw in the chapter on log-structured file systems

Kyutae Lee



FILE ./file-ssd/2.txt
Homework

7. "the low watermark (-G M) tells"
-> "the low watermark (-g M) tells"

thasinaz
=======

AND

(as we'll see in the chapter on log-structured file systems)
->
(as we saw in the chapter on log-structured file systems)

Since the log-structured file systems chapter comes before the flash-based SSDs chapter

Megan Cutrofello

AND

44.8 (Garbage Collection)

"For example, 2000 and 2001 clearly are still pointed to by the map;
100 and 101 are not and
therefore are candidates for garbage collection."

->

"For example, pages 2 and 3 clearly are still pointed to by the map (by logical blocks
2000 and 2001, respectively); 0 and 1
are no longer pointed to by anything and are therefore candidates for garbage collection."

Kyle Hale (Illinois Institute of Technology)



FILE ./file-ssd/1.txt
"and two lives blocks" -> "and two live blocks"

"results in 1 GB of memory needed the device" -> "results in 1 GB of memory needed by the device"

Jersey Deng (UCLA)




FILE ./file-journaling/1.txt
page 3
Stolen music perhaps?
-> (Missing comma)
Stolen music, perhaps?



FILE ./cpu-sched/1.txt
For example, if we had the schedule from above
->
For example, if we had the schedule from Figure 7.5

Muhammad Mobeen Movania




FILE ./file-intro/5.txt
This is not a typo, just a suggestion.

On p.4, there is a piece of source from xv6 kernel:

struct proc {
...
struct file *ofile[NOFILE]; // Open files
...
};

And I think the explanation around this piece can be more clarified if the following fact is included:
the indices for the array `ofile` are the file descriptors.
(Put it another way: the entries of the array `ofile` are indexed by the file descriptors.)

We can check the fact from the following source:
https://github.com/mit-pdos/xv6-public/blob/12abb1a56164a0d71fb7a76a465c912409f2f60b/sysfile.c#L36-L50
where `proc->ofile` is indexed by `fd` which is a file descriptor.

Kyutae Lee



FILE ./file-intro/4.txt
Suggest changing
"After creating a hard link to a file, to the file system, there is no difference between the original file name"
to
"After creating a hard link to a file, the file system perceives no difference between the original file name"



FILE ./file-intro/6.txt
On p.8,

The xv6 kernel just keeps these as an array as well, with one lock per
entry, as shown here:
struct {
struct spinlock lock;
struct file file[NFILE];
} ftable;

->

The xv6 kernel just keeps these as an array as well, with one lock for
the whole table, as shown here:
struct {
struct spinlock lock;
struct file file[NFILE];
} ftable;



---

As we can see from `struct spinlock lock;`, just one lock is declared "for the whole table", not "per entry".
And we can also check that the one lock is used for the whole table from the following source of xv6 kernel:
https://github.com/mit-pdos/xv6-public/blob/12abb1a56164a0d71fb7a76a465c912409f2f60b/file.c#L30-L38

Kyutae Lee



FILE ./file-intro/3.txt

page 24, in the tip section
More radical approaches, such as using a transactional file system [H+18], would solve the problem, there aren’t many transactional file systems in wide deployment.

-> change 3rd comma to a semi-colon

More radical approaches, such as using a transactional file system [H+18], would solve the problem; there aren’t many transactional file systems in wide deployment.

AND
"second, that the 64-bit offset be used" should be
"second, that the 64-bit offset is used"




FILE ./file-intro/2.txt

To allow someone (in this case, the other author) to access to this directory,  ->
To allow someone (in this case, the other author) to access this directory,


Alex Giorev




FILE ./file-intro/1.txt
radicial -> radical

Jersey Deng (UCLA)



FILE ./cpu-sched-multi/2.txt

Minor grammatical error in section 8.3
"The simple idea here is to boost the priority of all the jobs in system." should be
"The simple idea here is to boost the priority of all the jobs in the system."
(this should actually be in MLFQ chapter but oh well)

AND

On p.6, [A91] should be changed to [A90] to match the code in Reference.

Locks, unfortunately, can greatly reduce performance,
particularly asthe number of CPUs in the systems grows [A91].

->

Locks, unfortunately, can greatly reduce performance,
particularly asthe number of CPUs in the systems grows [A90].

Kyutae Lee



FILE ./cpu-sched-multi/1.txt
The sentence:
"The program then re-reads the value at address A; there is no such data CPU 2's cache",
should probably read:
"The program then re-reads the value at address A; there is no such data in CPU 2's cache"
Joel Hassan (U. Helsinki)


FILE ./preface/2.txt
"The first are systems programming projects" -> "The first is systems programming projects"
"The second type are ..." -> "The second type is ..."

Po-Hao Su (Taiwan)