CVE-2022-48629
Publication date:
05/03/2024
In the Linux kernel, the following vulnerability has been resolved:<br />
<br />
crypto: qcom-rng - ensure buffer for generate is completely filled<br />
<br />
The generate function in struct rng_alg expects that the destination<br />
buffer is completely filled if the function returns 0. qcom_rng_read()<br />
can run into a situation where the buffer is partially filled with<br />
randomness and the remaining part of the buffer is zeroed since<br />
qcom_rng_generate() doesn&#39;t check the return value. This issue can<br />
be reproduced by running the following from libkcapi:<br />
<br />
kcapi-rng -b 9000000 > OUTFILE<br />
<br />
The generated OUTFILE will have three huge sections that contain all<br />
zeros, and this is caused by the code where the test<br />
&#39;val & PRNG_STATUS_DATA_AVAIL&#39; fails.<br />
<br />
Let&#39;s fix this issue by ensuring that qcom_rng_read() always returns<br />
with a full buffer if the function returns success. Let&#39;s also have<br />
qcom_rng_generate() return the correct value.<br />
<br />
Here&#39;s some statistics from the ent project<br />
(https://www.fourmilab.ch/random/) that shows information about the<br />
quality of the generated numbers:<br />
<br />
$ ent -c qcom-random-before<br />
Value Char Occurrences Fraction<br />
0 606748 0.067416<br />
1 33104 0.003678<br />
2 33001 0.003667<br />
...<br />
253 � 32883 0.003654<br />
254 � 33035 0.003671<br />
255 � 33239 0.003693<br />
<br />
Total: 9000000 1.000000<br />
<br />
Entropy = 7.811590 bits per byte.<br />
<br />
Optimum compression would reduce the size<br />
of this 9000000 byte file by 2 percent.<br />
<br />
Chi square distribution for 9000000 samples is 9329962.81, and<br />
randomly would exceed this value less than 0.01 percent of the<br />
times.<br />
<br />
Arithmetic mean value of data bytes is 119.3731 (127.5 = random).<br />
Monte Carlo value for Pi is 3.197293333 (error 1.77 percent).<br />
Serial correlation coefficient is 0.159130 (totally uncorrelated =<br />
0.0).<br />
<br />
Without this patch, the results of the chi-square test is 0.01%, and<br />
the numbers are certainly not random according to ent&#39;s project page.<br />
The results improve with this patch:<br />
<br />
$ ent -c qcom-random-after<br />
Value Char Occurrences Fraction<br />
0 35432 0.003937<br />
1 35127 0.003903<br />
2 35424 0.003936<br />
...<br />
253 � 35201 0.003911<br />
254 � 34835 0.003871<br />
255 � 35368 0.003930<br />
<br />
Total: 9000000 1.000000<br />
<br />
Entropy = 7.999979 bits per byte.<br />
<br />
Optimum compression would reduce the size<br />
of this 9000000 byte file by 0 percent.<br />
<br />
Chi square distribution for 9000000 samples is 258.77, and randomly<br />
would exceed this value 42.24 percent of the times.<br />
<br />
Arithmetic mean value of data bytes is 127.5006 (127.5 = random).<br />
Monte Carlo value for Pi is 3.141277333 (error 0.01 percent).<br />
Serial correlation coefficient is 0.000468 (totally uncorrelated =<br />
0.0).<br />
<br />
This change was tested on a Nexus 5 phone (msm8974 SoC).
Severity CVSS v4.0: Pending analysis
Last modification:
26/02/2025