Background
With the development of space technology and nuclear technology, semiconductor ionizing radiation effect (also called Single event effect) is further classified and studied, such as Single event latch-up (SEL), Single Event Upset (SEU), Single Event Functional Interrupt (SEFI), Single Event Burnout (SEB), and the like. The single event effect can be divided into unrecoverable errors and recoverable errors according to whether the influence of the single event effect on the electronic components can be recovered or not. "unrecoverable error" or "hard error" refers to an error that, once it occurs, can cause fatal permanent damage to a device or system, such as SEB; "recoverable error" or "soft error" refers to a normal error that can be recovered by restarting a device or rewriting data, such as an SEU, a SET, an SED, etc. The single event latch SEL and the single event upset SEU are two single event effects with high occurrence frequency.
On the other hand, the failure caused by the single event effect has become an important threat of the crypto security chip. The cryptographic chip may perform complex encryption and decryption algorithms, such as symmetric algorithms or public key algorithms. The cipher chip has cipher key protecting mechanism to store the cipher key in special memory area without transmission via communication interface. Cryptographic chips are abundantly present in electronic products such as credit cards, mobile phone SIM cards, wireless network cards, RFID, USB keys, TPM (trusted platform module), etc. Therefore, cryptographic chips have become one of the reliable ways to secure information.
The attacker performs illegal reading, analysis, dissection and other measures on the password chip so as to obtain useful information and illegal benefits. Laser Fault Injection attack (LFI for short) discovered by the british scholars Sergei skorocogatov in 2002 becomes a very threatening Fault Injection attack method. An attacker only needs a small amount of error results, and the error results are compared with the correct results for analysis, so that part of or even all of the key bit information can be obtained.
Laser fault implantation, ion beam fault implantation and the like are fault implantation means capable of accurately focusing the target position of the chip, so that an attacker can obtain ideal error output conveniently. Such high-precision (radius of radiation ionization is in deep submicron level) fault injection attack method is not easy to defend, and is an important means for fault injection attackers at present. Currently, the detection specifications of international cryptography security modules, such as FIPS Publication140-3 draft issued by NIST in 2012 by the national standards and technology office, have explicitly written the defense against fault injection attacks into the security requirements of the cryptography security modules. In the detection criteria of the cryptographic circuit published by the national crypto-administration in 2012, a commercial cryptographic chip with high security level is also required to have the defense capability against fault injection attack.
The irradiation research using high-energy ion beams requires expensive special equipment, which generally includes a particle accelerator, a terminal beam machine, an oscilloscope, and the like. Such experiments are currently only performed by a few colleges and research institutions. Scientists have found that pulsed lasers can be used to simulate the single event effects of heavy cosmic ray ions in microelectronics and integrated circuits. J.S.Melinger et al studied the test and the basic mechanism of the laser single event effect in 1994, analyzed the interaction process of the laser and the electronic device material in more detail, and thought that although there is a great difference between the electron-hole pair plasma structure generated by the laser and the electron-hole pair plasma track structure generated by the heavy ion, it can still be used as an important laboratory evaluation means in the single event effect test. And in engineering design application, the laser single event effect test means is more practical than a gravity particle accelerator.
Modern semiconductor manufacturing has adopted 45nm node technology on a large scale, and 22nm node and 16nm node technologies are also referred to as "datchinson". Under such deep submicron process conditions, variations in process parameters in fabrication necessarily result in reduced reliability of the radiation effect. It is necessary to perform accurate and comprehensive fault injection on the chip and perform a test of system analysis on the response.
Research results show that picosecond pulse laser can focus the size of a laser beam spot to a micron-sized size, and single-particle upset sensitivity of a single transistor in an integrated circuit can be inspected in the early development stage of a semiconductor process. However, when the semiconductor manufacturing process is developed from micron to deep submicron, even nano node, the focusing requirement can not be met by using the conventional nano and picosecond laser. The development of high-precision irradiation by using a new laser technology has become an urgent need for studying the single event effect.
Disclosure of Invention
The embodiment of the invention provides a chip safety test method based on fault injection, which aims to solve the technical problem that high-precision fault injection test cannot be carried out when a semiconductor manufacturing process is developed from micron to deep submicron or even nano nodes in the prior art. The method comprises the following steps: sequentially focusing femtosecond laser on different positions on the surface of a chip to be detected through a synchronous control unit, and performing fault injection on the different positions of the chip to be detected, wherein the femtosecond laser generates two-photon absorption in the chip to be detected so as to overturn a logic unit in the chip to be detected; under the condition that the chip to be tested is irradiated at different positions by the femtosecond laser, respectively collecting operation results output by the chip to be tested; and respectively comparing and analyzing the acquired operation results with preset correct operation results of the chip to be detected, and determining whether effective faults occur at the positions of the chip to be detected irradiated by the femtosecond laser, wherein the number of the positions with the effective faults is a basis for judging the safety degree of the chip to be detected, and the effective faults refer to a type of errors of the secret key which can be analyzed by comparing and operating error operation results generated when the effective faults occur with the preset correct operation results.
In one embodiment, the safety degree of the chip to be tested is inversely proportional to the proportion of the number of positions with effective faults in all the positions irradiated by the femtosecond laser.
In one embodiment, the photon energy of the femtosecond laser is larger than the energy gap bandwidth of the chip to be tested.
In one embodiment, the femtosecond laser has a wavelength that meets the penetration depth requirement of the femtosecond laser focused on the logic unit in the chip to be tested and meets the energy requirement of the occurrence of energy level transition.
In one embodiment, a chip to be tested is placed on a stage below a confocal microscope in a femtosecond laser, and the femtosecond laser is sequentially focused on different positions on the surface of the chip to be tested through a synchronous control unit, including: when the chip to be tested sends a starting signal to the synchronous control unit and the chip to be tested starts functional operation, the synchronous control unit controls a femtosecond laser to focus the femtosecond laser on one position of the surface of the chip to be tested, when the chip to be tested sends a pause signal to the synchronous control unit and the chip to be tested completes one functional operation, the synchronous control unit controls the femtosecond laser to stop fault injection on the one position of the surface of the chip to be tested, and an operation result of the chip to be tested for completing the current functional operation is collected; when the chip to be tested sends a next starting signal to the synchronous control unit and the chip to be tested starts functional operation, the synchronous control unit controls the confocal microscope to move the objective table by a preset step length, the femtosecond laser is focused on the next position on the surface of the chip to be tested, when the chip to be tested sends a next pause signal to the synchronous control unit and the chip to be tested completes one-time functional operation, the synchronous control unit controls the femtosecond laser to stop fault injection on the next position on the surface of the chip to be tested, and the operation result of the chip to be tested for completing the current functional operation is collected; and sequentially circulating until the whole surface of the chip to be tested is traversed, and sending a completion signal to the chip to be tested through the synchronous control unit to finish fault injection.
The embodiment of the invention also provides a chip safety test system based on fault injection, which aims to solve the technical problem that the high-precision fault injection test cannot be carried out when the semiconductor manufacturing process is developed from micron to deep submicron or even nano node in the prior art. The system comprises: the femtosecond laser device is used for emitting femtosecond laser, and a chip to be tested is placed on an objective table below the common focusing microscope in the femtosecond laser device; the synchronous control unit is used for controlling the femtosecond laser to focus the femtosecond laser on different positions on the surface of a chip to be tested in sequence and injecting faults into the different positions of the chip to be tested, wherein the femtosecond laser generates two-photon absorption in the chip to be tested so that a logic unit in the chip to be tested is turned over; the data acquisition equipment is used for respectively acquiring operation results output by the chip to be detected under the condition that the chip to be detected is irradiated by the femtosecond laser at different positions; and the data analysis equipment is used for respectively comparing and analyzing the acquired operation results with preset correct operation results of the chip to be detected, and determining whether effective faults occur at the positions of the chip to be detected irradiated by the femtosecond laser, wherein the number of the positions with the effective faults is a basis for judging the safety degree of the chip to be detected, and the effective faults refer to a type of errors of the secret key, which can be analyzed by comparing and operating error operation results generated when the effective faults occur with the preset correct operation results.
In one embodiment, the safety degree of the chip to be tested is inversely proportional to the proportion of the number of positions with effective faults in all the positions irradiated by the femtosecond laser.
In one embodiment, the photon energy of the femtosecond laser is larger than the energy gap bandwidth of the chip to be tested.
In one embodiment, the femtosecond laser has a wavelength that meets the penetration depth requirement of the femtosecond laser focused on the logic unit in the chip to be tested and meets the energy requirement of the occurrence of energy level transition.
In an embodiment, the synchronous control unit is specifically configured to control the femtosecond laser to focus femtosecond laser on a position on the surface of the chip to be tested when a start signal is received from the chip to be tested and the chip to be tested starts functional operation, control the femtosecond laser to stop fault injection on the position on the surface of the chip to be tested when a pause signal is received from the chip to be tested and the chip to be tested completes one functional operation, and control the data acquisition device to acquire an operation result of the chip to be tested completing the functional operation; when a next starting signal sent by the chip to be tested is received and the chip to be tested starts functional operation, controlling the confocal microscope to move the objective table by a preset step length, focusing the femtosecond laser on the next position on the surface of the chip to be tested, controlling the femtosecond laser to stop fault injection on the next position on the surface of the chip to be tested when a next pause signal sent by the chip to be tested is received and the chip to be tested completes one functional operation, and controlling the data acquisition equipment to acquire an operation result of the chip to be tested to complete the current functional operation; and sequentially circulating until the whole surface of the chip to be tested is traversed, and sending a completion signal to the chip to be tested to finish fault injection.
In the embodiment of the invention, femtosecond laser is focused on different positions on the surface of a chip to be detected through a synchronous control unit to perform fault injection on different positions of the chip to be detected, operation results output by the chip to be detected are respectively collected under the condition that the chip to be detected is irradiated by the femtosecond laser at different positions, and finally the collected operation results are compared and analyzed with a preset correct operation result of the chip to be detected, so that whether effective faults occur at the positions of the chip to be detected, which are irradiated by the femtosecond laser, can be determined, and the safety degree of the chip to be detected can be further judged according to the number of the positions with the effective faults. Due to the fact that femtosecond laser generates two-photon absorption in a chip to be tested, error (namely fault) injection attack in a very small range can be achieved, the problem that a fault injection area cannot be accurately controlled due to the fact that the diameter of a conventional laser focusing beam is too large after an integrated circuit manufacturing process enters a nano node in the prior art can be solved, the accuracy of fault injection can be further improved, and the accuracy of a chip safety test result based on fault injection can be improved.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following embodiments and accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
In an embodiment of the present invention, a chip security testing method based on fault injection is provided, as shown in fig. 1, the method includes:
step 101: sequentially focusing femtosecond laser on different positions on the surface of a chip to be detected through a synchronous control unit, and performing fault injection on the different positions of the chip to be detected, wherein the femtosecond laser generates two-photon absorption in the chip to be detected so as to overturn a logic unit in the chip to be detected;
step 102: under the condition that the chip to be tested is irradiated at different positions by the femtosecond laser, respectively collecting operation results output by the chip to be tested;
step 103: and comparing and analyzing the acquired operation result with a preset correct operation result of the chip to be detected, and determining whether effective faults occur at the positions of the chip to be detected irradiated by the femtosecond laser, wherein the number of the positions with the effective faults is a basis for judging the safety degree of the chip to be detected, and the effective faults refer to a type of errors of the secret key which can be analyzed by comparing and operating an error operation result generated when the effective faults occur with the preset correct operation result.
As can be seen from the process shown in fig. 1, in the embodiment of the present invention, the femtosecond laser is focused on different positions on the surface of the chip to be detected by the synchronous control unit to perform fault injection on different positions of the chip to be detected, and under the condition that the chip to be detected is irradiated by the femtosecond laser at different positions, the operation results output by the chip to be detected are respectively collected, and finally, the collected operation results are compared with the preset correct operation results of the chip to be detected, so that whether an effective fault occurs at the position of the chip to be detected irradiated by the femtosecond laser can be determined, and the safety degree of the chip to be detected can be determined according to the number of the positions where the effective fault occurs. Due to the fact that femtosecond laser generates two-photon absorption in a chip to be tested, error (namely fault) injection attack in a very small range can be achieved, the problem that a fault injection area cannot be accurately controlled due to the fact that the diameter of a conventional laser focusing beam is too large after an integrated circuit manufacturing process enters a nano node in the prior art can be solved, the accuracy of fault injection can be further improved, and the accuracy of a chip safety test result based on fault injection can be improved.
In specific implementation, the inventor finds that the laser focusing system used in the current research has a transverse spot size larger than the laser wavelength due to the limitation of diffraction rules. And Femtosecond (10-15s) pulse Laser (femto second Laser) has the characteristics of narrow pulse width and high peak power, and is not a single-photon process but a two-photon or multi-photon process when interacting with a substance. Therefore, the femtosecond laser beam with Gaussian lateral distribution interacts with the substance not in the whole focal spot range but far smaller than the light spot, and the focusing scale can reach 1/20 wavelengths, theoretically tens of nanometers. The femtosecond laser pulse is firstly obtained by utilizing the principle of collision pulse mode locking (CPM) in a dye laser, with the development of the crystal growth technology in 80 years, a series of solid lasers with excellent performance are appeared, represented by titanium-doped Sapphire crystals (Ti: Sapphire) which are published in 1982, and compared with the dye laser, the laser which takes titanium Sapphire as a gain medium has a wider tuning range, which is equivalent to a waveband covered by four-five dye combinations. The micro-nano processing of the femtosecond laser is widely applied to various fields such as ultra-precise laser processing, a multi-photon microscope, nonlinear spectroscopy and the like, and research on micro-nano processing of the femtosecond laser is carried out in Chengyi of Shanghai optical mechanical institute of Chinese academy of sciences in 2014, Qinhua of Jilin university and the like. The mechanism of two-photon absorption in the composite material was studied in 2015 by shanghai optical engine wang jun, etc.
The peak power of femtosecond laser is hundreds kilowatts, and the laser pulse width is less than 100fs (1fs is 10--15s). The focused femtosecond laser has extremely high field intensity, and many media present remarkable nonlinear properties when the intensity of an optical field in the media is comparable to the intensity of an electric field in a molecule. Two-photon absorption is a typical third-order nonlinear optical effect, with the probability of generation proportional to the square of the photon flux density. The two-photon absorption can only occur under strong light intensity, and the two-photon absorption is only limited to the space volume of the focal point of the objective lens, which is about lambda3(λ is the wavelength of the incident light). The incident light can only obtain a higher power density at this point, and multiphoton absorption and ionization occur, thereby turning over the logic unit inside the security chip. In addition, the femtosecond laser is used for realizing two-photon absorption, long-wavelength laser is adopted, the penetrating power is strong, and the laser can directly act inside the material. Namely, the femtosecond laser can be used as a high-precision chip fault injection tool. Therefore, the inventor proposes the chip safety test method based on femtosecond laser fault injection by utilizing the principle that the femtosecond laser generates two-photon absorption in the chip to be tested so as to overturn the logic unit in the chip to be tested, so as to improve the accuracy of fault injection and ensure the precision of the chip safety test result.
In specific implementation, in order to ensure that the femtosecond laser can be focused on the logic unit in the chip to be tested and then two-photon absorption can occur in the chip to be tested so as to turn over the logic unit in the chip to be tested, in this embodiment, the wavelength of the femtosecond laser meets the penetration depth requirement of the femtosecond laser focused on the logic unit in the chip to be tested and the energy requirement of energy level transition.
Specifically, the laser wavelength is selected according to the material of the semiconductor (i.e., the chip to be tested), and the semiconductor generally adopts a silicon material, so that the premise that the femtosecond laser generates two-photon absorption and ionization is that the photon energy of the femtosecond laser exceeds the energy gap bandwidth (>1.1eV) of the semiconductor, and for example, the femtosecond laser wavelength is below 1064 nm. However, the smaller the wavelength is, the shallower the penetration depth is, however, the mechanism of two-photon absorption indicates that when the ultrashort pulse laser propagates in the molecular medium, the processes of generation of higher harmonics, Stimulated Raman Scattering (SRS), spontaneous emission Amplification (ASE), and superfluorescence emission (SF) are always accompanied, so that when selecting the wavelength, we need to expand the wavelength, and try to use a band above 900nm, which has a relatively deeper penetration depth and a smaller ionization rate, that is, the wavelength of the femtosecond laser meets both the penetration depth requirement that the femtosecond laser can focus on a logic unit in a chip to be tested and the energy requirement that energy level transition can occur, for example, the wavelength of the femtosecond laser may be smaller than 1064nm and larger than 900nm, so as to ensure that the femtosecond laser generates two-photon absorption in a semiconductor, and the logic unit is turned over.
During specific implementation, before the femtosecond laser is sequentially focused on different positions on the surface of the chip to be tested through the synchronous control unit, the incident parameter of the femtosecond laser can be determined according to the process and the test requirement of the chip to be tested, and the femtosecond laser emitted by the femtosecond laser under the incident parameter can overturn the logic unit in the chip to be tested. Specifically, the incident parameter may be photon energy and wavelength of the femtosecond laser.
In specific implementation, the output power of the femtosecond laser can be adjusted in an incremental manner by 1% to find appropriate power approximately at about 1.4W according to the existing laser fault injection attack experiment.
During specific implementation, in the fault injection process, in order to realize that the femtosecond laser is focused on different positions of the surface of the chip to be tested through the synchronous control unit in sequence, in this embodiment, the chip to be tested is placed on the objective table below the common focusing microscope in the femtosecond laser, and the femtosecond laser is focused on different positions of the surface of the chip to be tested through the synchronous control unit in sequence, including: when the chip to be tested sends a starting signal to the synchronous control unit and the chip to be tested starts functional operation, the synchronous control unit controls a femtosecond laser to focus the femtosecond laser on one position of the surface of the chip to be tested, when the chip to be tested sends a pause signal to the synchronous control unit and the chip to be tested completes one functional operation, the synchronous control unit controls the femtosecond laser to stop fault injection on the one position of the surface of the chip to be tested, and an operation result of the chip to be tested for completing the current functional operation is collected; when the chip to be tested sends a next starting signal to the synchronous control unit and the chip to be tested starts functional operation, the synchronous control unit controls the confocal microscope to move the objective table by a preset step length, the femtosecond laser is focused on the next position on the surface of the chip to be tested, when the chip to be tested sends a next pause signal to the synchronous control unit and the chip to be tested completes one-time functional operation, the synchronous control unit controls the femtosecond laser to stop fault injection on the next position on the surface of the chip to be tested, and the operation result of the chip to be tested for completing the current functional operation is collected; and sequentially circulating until the whole surface of the chip to be tested is traversed, and sending a completion signal to the chip to be tested through the synchronous control unit to finish fault injection.
Specifically, in the fault injection process, in order to realize that the femtosecond laser is focused on different positions on the surface of the chip to be detected through the synchronous control unit in sequence, secondary development is carried out by utilizing the matching software of the laser and the confocal microscope, the laser light intensity and the synchronous strategy are controlled, the two-dimensional movement of the chip to be detected is realized by utilizing the two-dimensional objective table of the microscope, the femtosecond laser and the confocal microscope are controlled through the synchronous control unit, and the irradiation start/finish of the femtosecond laser is finished under the cooperative control of the synchronous control unit and the chip to be detected. For example, fig. 2 shows a handshake protocol between the synchronization control unit and the chip under test. We define three handshake signals: start signal, Stop signal and Done signal. The Start signal is a Start signal sent to the synchronous control unit after the chip to be tested is ready, the stop signal is a pause signal of femtosecond laser fault injection, and the Done signal is a completion signal.
The specific fault injection process is as follows: when the chip to be tested is ready, a Start signal Start is sent to the synchronous control unit, the chip to be tested starts logic operation, the synchronous control unit controls the femtosecond laser to focus the femtosecond laser on one position on the surface of the chip to be tested, when the chip to be tested finishes one function operation, a pause signal stop of femtosecond laser fault injection is sent to inform the synchronous control unit to pause fault injection, the chip to be tested performs self-reset at the same time, the influence of the last soft error is eliminated, the next round of fault injection test is prepared, the synchronous control unit controls the femtosecond laser to stop injecting faults on the one position of the chip to be tested, and simultaneously the data acquisition equipment is controlled to acquire the operation result of the chip to be tested to finish the function operation, and the operation result of the function operation is used for comparing and analyzing with the preset correct operation result, to determine whether the one position of the chip under test is faulty. After the femtosecond laser is suspended, the synchronous control unit controls the confocal microscope to move the two-dimensional objective table by a pre-designed step length, the femtosecond laser is focused on the next position on the surface of the chip to be detected, when the next Start signal Start sent by the chip to be detected to the synchronous control unit is received, simultaneously, the chip to be tested starts logic operation work, the synchronous control unit controls the femtosecond laser to focus the femtosecond laser on the next position on the surface of the chip to be tested for fault injection, when the chip to be tested finishes one function operation, a pause signal stop of next femtosecond laser fault injection is sent to inform the synchronous control unit to pause the fault injection, and simultaneously, the chip to be tested is subjected to self-reset, the influence of the last soft error is eliminated, the next round of fault injection test is prepared, and the synchronous control unit controls the data acquisition equipment to acquire the operation result of the chip to be tested to complete the function operation. And circulating the fault injection process in sequence, when the synchronous control unit finishes traversing test on the whole surface of the chip to be tested, sending a completion signal Done to the chip to be tested, finishing fault injection, sending all the operation results collected before to an irradiation effect reliability/safety analyzer, in the irradiation effect reliability/safety analyzer, respectively comparing and analyzing all the operation results collected with preset correct operation results of the chip to be tested, determining whether effective faults occur at the positions of the chip to be tested irradiated by the femtosecond laser, and determining the positions where the effective faults occur.
In the radiation effect reliability/safety analyzer, only the sensitive logic circuits may be concerned about the injection error, and the circuits irrelevant to the safety may not be concerned about, so that the fault injection theory of the cryptographic algorithm needs to be combined to judge whether the scanning point is a sensitive point (i.e. the position where the effective fault occurs). Due to the two-photon absorption of the femtosecond laser, fault injection attack in a very small range can be realized, so that whether the output result is matched with the candidate fault types or not can be analyzed according to the attack theory of fewer bit faults of 1 bit or 2 bits, and whether the scanning point is a sensitive point or not can be further determined. And comparing the operation output result of each irradiation with the correct operation result, if the operation output result is different from the correct operation result, indicating that an error is injected into the logic circuit at the irradiation part in the operation process of the chip to be detected, and marking the corresponding position as a fault injection sensitive point. If the operation output result is the same as the correct operation result, the fault is not injected into the logic circuit at the irradiation position. After the operation output results of all irradiation positions are sequentially analyzed, a fault injection sensitive point positioning diagram shown in fig. 3 can be obtained, and black crosses in fig. 3 represent fault injection sensitive points.
Specifically, the safety degree of the chip to be tested is inversely proportional to the proportion of the number of the positions with faults in all the positions irradiated by the femtosecond laser, for example, as shown in fig. 3, the fault injection sensitive point can be regarded as the position with effective fault injection in the chip to be tested, and the more the positions with effective fault injection are, that is, the greater the proportion of the positions with effective fault injection in all the positions irradiated by the femtosecond laser is, the lower the safety degree of the chip to be tested is.
In specific implementation, when a femtosecond laser fault injection test experiment platform is built, a Mai Tai deep femtosecond laser of the American Spectra-Physics company and an A1MP + series confocal microscope of the Nikon company can be adopted. The former can provide irradiation with adjustable power in the range of 680nm-1040nm, and adopts the ultra-stable regeneration mold technology, the wavelength adjustment and excitation configuration is simple and easy to adjust, the light beam pointing is stable, the power fluctuation is small, and the wavelength drift is eliminated. The latter directly embeds a femtosecond laser and carries out light path design, and can focus the femtosecond laser in a space range of 1 μm, and can focus the femtosecond laser beam in a smaller space range if a proper objective lens is adopted. The chip to be tested can adopt an RSA encryption circuit of an FPGA based on ALTERA DE 2-115. The confocal microscope is provided with a two-dimensional electric objective table, so that two-dimensional movement of a sample can be realized, and fault injection attack of the whole surface range of the electronic chip can be realized. Preliminary results can be obtained on an experimental platform, for example, taking an FPGA chip as an example, irradiating a circuit from the front with a wavelength of 900nm, the diameter of a focused light spot being 0.6um, the stage moving with a step size of 0.1um, and generating a stable error when the power is 2.5W. Table 1 shows the decryption parameters of binary RSA based on FPGA, and table 2 shows the decryption results of errors generated in the decryption operation by the RSA cryptographic circuit under the irradiation of femtosecond laser.
TABLE 1
TABLE 2
Based on the same inventive concept, the embodiment of the present invention further provides a chip safety test system based on fault injection, as described in the following embodiments. Because the principle of solving the problems of the chip safety test system based on fault injection is similar to the chip safety test method based on fault injection, the implementation of the chip safety test system based on fault injection can refer to the implementation of the chip safety test method based on fault injection, and repeated parts are not described again. As used hereinafter, the term "unit" or "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 4 is a block diagram of a structure of a chip security testing system based on fault injection according to an embodiment of the present invention, as shown in fig. 4, including:
a femtosecond laser 401 for emitting femtosecond laser, wherein a chip to be tested is placed on an objective table below a common focusing microscope in the femtosecond laser;
the synchronous control unit 402 is used for controlling the femtosecond laser to sequentially focus the femtosecond laser on different positions on the surface of a chip to be tested and perform fault injection on the different positions of the chip to be tested, wherein the femtosecond laser generates two-photon absorption in the chip to be tested so as to overturn a logic unit in the chip to be tested;
the data acquisition equipment 403 is used for respectively acquiring operation results output by the chip to be detected under the condition that the chip to be detected is irradiated by the femtosecond laser at different positions;
and the data analysis device 404 is configured to compare and analyze the acquired operation results with preset correct operation results of the chip to be tested, and determine whether effective faults occur at positions of the chip to be tested irradiated by the femtosecond laser, where the number of the positions where the effective faults occur is a basis for determining a safety degree of the chip to be tested, where the effective faults refer to a type of errors that can be analyzed by performing a comparison operation on an erroneous operation result generated when the effective faults occur and the preset correct operation result.
In one embodiment, the safety degree of the chip to be tested is inversely proportional to the proportion of the number of positions with effective faults in all the positions irradiated by the femtosecond laser.
In one embodiment, the photon energy of the femtosecond laser is larger than the energy gap bandwidth of the chip to be tested.
In one embodiment, the femtosecond laser has a wavelength that meets the penetration depth requirement of the femtosecond laser focused on the logic unit in the chip to be tested and meets the energy requirement of the occurrence of energy level transition.
In an embodiment, the synchronous control unit is specifically configured to control the femtosecond laser to focus femtosecond laser on a position on the surface of the chip to be tested when a start signal is received from the chip to be tested and the chip to be tested starts functional operation, control the femtosecond laser to stop fault injection on the position on the surface of the chip to be tested when a pause signal is received from the chip to be tested and the chip to be tested completes one functional operation, and control the data acquisition device to acquire an operation result of the chip to be tested completing the functional operation; when a next starting signal sent by the chip to be tested is received and the chip to be tested starts functional operation, controlling the confocal microscope to move the objective table by a preset step length, focusing femtosecond laser on the next position on the surface of the chip to be tested, controlling the femtosecond laser to stop fault injection on the next position on the surface of the chip to be tested when a next pause signal sent by the chip to be tested is received and the chip to be tested completes one functional operation, and controlling the data acquisition equipment to acquire an operation result of the chip to be tested to complete the current functional operation; and sequentially circulating until the whole surface of the chip to be tested is traversed, and sending a completion signal to the chip to be tested to finish fault injection.
In the embodiment of the invention, femtosecond laser is focused on different positions on the surface of a chip to be detected through a synchronous control unit to perform fault injection on different positions of the chip to be detected, operation results output by the chip to be detected are respectively collected under the condition that the chip to be detected is irradiated by the femtosecond laser at different positions, and finally the collected operation results are compared and analyzed with a preset correct operation result of the chip to be detected, so that whether effective faults occur at the positions of the chip to be detected, which are irradiated by the femtosecond laser, can be determined, and the safety degree of the chip to be detected can be further judged according to the number of the positions with the effective faults. Due to the fact that femtosecond laser generates two-photon absorption in a chip to be tested, error (namely fault) injection attack in a very small range can be achieved, the problem that a fault injection area cannot be accurately controlled due to the fact that the diameter of a conventional laser focusing beam is too large after an integrated circuit manufacturing process enters a nano node in the prior art can be solved, the accuracy of fault injection can be further improved, and the accuracy of a chip safety test result based on fault injection can be improved.
It will be apparent to those skilled in the art that the modules or steps of the embodiments of the invention described above may be implemented by a general purpose computing device, they may be centralized on a single computing device or distributed across a network of multiple computing devices, and alternatively, they may be implemented by program code executable by a computing device, such that they may be stored in a storage device and executed by a computing device, and in some cases, the steps shown or described may be performed in an order different than that described herein, or they may be separately fabricated into individual integrated circuit modules, or multiple ones of them may be fabricated into a single integrated circuit module. Thus, embodiments of the invention are not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes may be made to the embodiment of the present invention by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.